University of Southern California Dissertations and Theses

From single molecules to bacterial nanowires: functional and dynamic imaging of the extracellular electron transfer network in Shewanella oneidensis MR-1

(USC Thesis Other)

From single molecules to bacterial nanowires: functional and dynamic imaging of the extracellular electron transfer network in Shewanella oneidensis MR-1

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content From Single Molecules to Bacterial Nanowires: Functional and Dynamic Imaging of the Extracellular Electron Transfer Network in Shewanella oneidensis MR-1 by Grace W Chong A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulllment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (MOLECULAR BIOLOGY) August 2021 Copyright 2021 Grace W Chong Acknowledgments First, thanks to my advisor, Moh El-Naggar. He is a great scientist, advisor and boss. In Moh's lab, I also got to meet many great labmates, including Sahand Pirba- dian, Lori Zacharo, Matt Xu, Amruta Karbelkar, Marko Chavez, Karla Abuyen, Christina Cole, Leila Mahrokh, Tingting Yang, Josh Atkinson, Hank Yu, Greyson MacLean, Annie Rowe, Yamini Jangir, Nicole Beedle, Matthieu Kirkland and Hye- suk Byun. Though I overlapped with some of you more than others, it was fun getting to know you all through the sharing of science, space, time, and food. Thanks to my collaborators, including those in my lab, those at USC, and those (many) elsewhere, but especially those who have contributed or were consulted for the research presented in this thesis: • Sahand Pirbadian, my senior labmate who gave advice both in general and especially for Chapter 2, having the unique perspective as one of few in the world to have studied Shewanella nanowires before me, and who performed simulations applying my experimental results in Chapter 4. • Yunke Zhao and Fabien Pinaud, our collaborators in Pinaud Lab at USC, who shared their microscope with me and trained me in microscopy, single- particle tracking, and diusion analyses that form a big part of Chapter 4. ii • Lori Zacharo, my senior labmate and resident bio-expert who had answers to questions when I had trouble with the many molecular biology techniques required for Chapter 3. • Je Gralnick, our collaborator and genetics expert in University of Minnesota who shared various mutant strains and empty plasmid used in Chapters 2, 3, and 4. • Namita Schro and Steve Finkel from the Finkel Lab at USC, who taught me the basics of PCR and primer design and shared assorted lab materials (from test tubes to X-ray lm) for Chapter 3. • Tom Clarke, whose lab in University of East Anglia specializes in structural work, who we consulted on where to place the peptide tag used in Chapters 3-4. • The microscopy centers that enabled work for Chapter 2: USC Core Center for Excellence in Imaging (CNI), home to a 200 kV transmission electron microscope; Chris Buser at the Huntington Medical Research Institutes, who imaged our electron microscopy samples at 80 kV; and the USC Center of Excellence in NanoBiophysics, the basement home to our lab's old but simple atomic force microscope. • Our upstairs neighbors in the James Boedicker Lab at USC, who shared vari- ous equipment such as PCR machine, DNA gel machine, and electroporation machine used in Chapter 3, and • Greyson MacLean, our favorite physics undergrad turned pipette master who helped with assorted tasks like making media, growing cells, and plasmid purication (I think he joined us sometime in the middle of Chapters 3-4). iii Thanks to my dissertation committee members, Steve Finkel, Fabien Pinaud, and Ken Nealson. Thanks to my department(s) at USC, the one I joined ocially for my degree (Molecular and Computational Biology Section in the Department of Biological Sciences) and the one I joined unocially when I became a part of Moh's lab (Department of Physics). Thanks to the funding sources that sponsored these works and me: the Air Force Oce of Scientic Research Presidential Early Career Award for Scientists and Engineers Award (grant FA955014-1-0294, to Moh El-Naggar), the United States Department of Energy (grant DE-FG02-13ER16415), and the National Science Foundation Graduate Research Fellowship Program (NSF GRFP, grant DGE1418060). Finally, thanks to the people who made my life better outside of lab. iv Contents Acknowledgments ii List of Figures viii List of Tables xx Abstract xxii 1 Introduction 1 1.1 Background 1: Dening the system of study . . . . . . . . . . . . . 1 1.1.1 From electron transfer to extracellular respiration and its applications . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1.2 Shewanella oneidensis MR-1: Discovery and basic features . 4 1.1.3 Multiheme cytochromes and extracellular electron transfer (EET) in S. oneidensis . . . . . . . . . . . . . . . . . . . . . 5 1.1.4 Multiheme cytochromes, bacterial nanowires, and long-distance electron transfer (ET) in S. oneidensis . . . . . . . . . . . . 11 1.1.5 Outward, inward, and cell-to-cell EET in diverse microor- ganisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 1.2 Background 2: Methods . . . . . . . . . . . . . . . . . . . . . . . . 14 1.2.1 Perfusion ow imaging platform . . . . . . . . . . . . . . . . 14 1.2.2 Transmission electron microscopy . . . . . . . . . . . . . . . 16 1.2.3 Atomic force microscopy . . . . . . . . . . . . . . . . . . . . 17 1.2.4 Total internal re ection uorescence microscopy . . . . . . . 18 1.2.5 Single-particle tracking and basic types of diusion . . . . . 19 1.3 Overview of chapters . . . . . . . . . . . . . . . . . . . . . . . . . . 21 2 Surface-induced formation and redox-dependent staining of outer membrane extensions in Shewanella oneidensis MR-1 25 2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 2.2 Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . 29 2.2.1 Cell cultivation . . . . . . . . . . . . . . . . . . . . . . . . . 29 v 2.2.2 Fluorescence microscopy . . . . . . . . . . . . . . . . . . . . 29 2.2.3 Heme staining and transmission electron microscopy . . . . . 31 2.3 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . 33 2.3.1 Surface contact is sucient to induce production of outer membrane extensions by Shewanella oneidensis MR-1 . . . . 33 2.3.2 Redox-dependent staining of extracellular laments . . . . . 42 2.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 3 Developing a system for site-specic labeling of cell surface cytochromes in Shewanella oneidensis MR-1 47 3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 3.2 Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . 49 3.2.1 Strains, plasmids, and culture conditions . . . . . . . . . . . 49 3.2.2 Plasmid construction . . . . . . . . . . . . . . . . . . . . . . 50 3.2.3 Transformation and quality checking . . . . . . . . . . . . . 56 3.2.4 DNA sequencing . . . . . . . . . . . . . . . . . . . . . . . . 59 3.2.5 SDS-PAGE and heme staining . . . . . . . . . . . . . . . . . 59 3.2.6 In vivo biotinylation . . . . . . . . . . . . . . . . . . . . . . 60 3.2.7 Western blot . . . . . . . . . . . . . . . . . . . . . . . . . . 61 3.2.8 Fluorescence microscopy . . . . . . . . . . . . . . . . . . . . 63 3.3 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . 64 3.3.1 Successful and specic in vivo labeling of cell surface cytochromes MtrC and OmcA . . . . . . . . . . . . . . . . . . . . . . . . 64 3.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 4 Single molecule tracking of bacterial cell surface cytochromes reveals dynamics that impact long-distance electron transport 71 4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 4.2 Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . 77 4.2.1 Cell cultivation . . . . . . . . . . . . . . . . . . . . . . . . . 77 4.2.2 In vivo biotinylation . . . . . . . . . . . . . . . . . . . . . . 77 4.2.3 Total internal re ection uorescence (TIRF) microscopy . . 78 4.2.4 Single-particle tracking (SPT) and diusion analyses . . . . 81 4.2.5 Kinetic Monte Carlo simulations of long-distance electron transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 4.3 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . 87 4.3.1 Single-particle imaging and tracking reveals mobility of MtrC and OmcA along cell surface and membrane extensions . . . 87 4.3.2 Quantifying the dynamics of MtrC and OmcA along the cell surface and membrane extensions . . . . . . . . . . . . . . . 90 vi 4.3.3 Simulations of overall electron transport along membrane surfaces combine direct electron hopping and physical diu- sion of cytochromes . . . . . . . . . . . . . . . . . . . . . . . 108 4.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 5 Conclusion 115 Bibliography 121 vii List of Figures 1.1 Redox tower listing several common electron donors and acceptors. Chemicals are listed by their standard reduction poten- tial (E o ', at pH 7 and standard state) and arranged in order of higher energy (more negative E o ') to lower energy (more positive E o '). Positive E o ' means it is energetically favorable for a reaction to proceed in the order written, e.g. O 2 reduced to H 2 O. Negative E o ' means it is energetically favorable for the reaction to proceed in the other direction, e.g. H 2 oxidized to 2H + . It is energetically favor- able for electrons to ow from more negative to more positive E o '. Those closer to the bottom of this tower have the greatest anity for electrons and are thus the best electron acceptors. The red arrows classify some organisms by the type of electron donors/acceptors used. Figure courtesy of Moh El-Naggar. . . . . . . . . . . . . . . . 3 1.2 Crystal structures and heme arrangement of two decaheme cytochromes important in extracellular electron transport and metal reduction in S. oneidensis. (A-B) Crystal struc- ture and heme arrangement of extracellular facing outer membrane cytochrome MtrC (PDB 4LM8) obtained from S. oneidensis MR-1, gure from [14]. (C-D) Crystal structure and heme arrangement of outer membrane-associated periplasmic cytochrome MtrA obtained from Shewanella baltica OS185, gure from [15]. . . . . . . . . . . . 7 1.3 Structure of a biological transmembrane wire. MtrABC porin- cytochrome complex crystallized from S. baltica OS185, including decaheme cytochromes MtrC (blue) and MtrA (pink) as well as the porin MtrB (yellow green) that allows them to interact across the membrane. The two hemes at the interface of MtrA and MtrC are separated by a distance of8 A. All other inter-heme distances are <8 A. Total length of transmembrane wire is185 A. Figure from [15]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 viii 1.4 Outward extracellular electron transfer (EET) pathways proposed for S. oneidensis MR-1. Cytochromes on the inner membrane (where the electron transport chain is also located) can transfer electrons to soluble periplasmic cytochromes, which can then relay electrons to the outer membrane via porin-cytochrome conduits to extracellular electron acceptors. Once electrons are exported to the cell surface, terminal electron acceptors (including solid-phase minerals and electrodes) can be reduced by direct con- tact through extracellular multiheme cytochromes (MHCs), MHC- bound redox cofactors, or MHCs found on outer membrane exten- sions, or indirectly through diusion of soluble electron carriers. Black arrows illustrate possible electron ow. Red shape represents inner membrane cytochromes (e.g. CymA), orange shape represents periplasmic soluble cytochromes (e.g. STC, FccA), yellow/gray shapes represent outer membrane porin-cytochrome conduits (e.g. Mtr/Omc pathway), and green circles represent soluble redox-active molecules (e.g. avins). Abbreviations: electron transfer (ET), outer membrane (OM), inner membrane (IM). Figure adapted from [1]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 1.5 Schematic of perfusion ow imaging platform. Cells are injected into the chamber, and attached to a glass cover slip (or EM grid) which is vacuum sealed to the chamber. Cells are sus- tained with a continuous ow of medium, membrane stained with FM 4-64FX, and monitored with uorescence microscopy. Figure from [25]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 ix 1.6 Diagram of single-particle tracking (SPT) process and basic types of diusion. SPT begins with image acquisition by time- lapse microscopy, where the same 2-D area of the sample is imaged repeatedly over time (1 image frame captured per time increment). Image (top left) is a snapshot of S. oneidensis and its membrane extensions (green). Red circles illustrate molecules of interest (e.g. quantum dots, labeling proteins of interest) which can be tracked on the surface of a bacterial cell or its membrane extensions (green). Image (bottom left) shows total internal re ection uorescence (TIRF) microscopy, a method of image acquisition for SPT. (Top right) Next, a particle localization method is used to nd the center of each dot representing the uorescent signal from a molecule of interest. This is repeated for all frames in the time-lapse series. The localized particle positions are then connected frame-by-frame to build tra- jectories. (Bottom right) The mean squared displacement (MSD) of a single particle trajectory plotted as a function of time can be used to classify a trajectory into one of 3 major types of diusion, based on the shape of its MSD curve. Figure images adapted from [24, 64, 66, 67]. Figure courtesy of Moh El-Naggar. . . . . . . . . . 20 2.1 Outer membrane extensions are commonly formed by surface- attached perfusion culture cells. (A) Time-lapse uorescence microscopy snapshot of outer membrane extensions (OMEs, white arrows) produced by S. oneidensis MR-1 at a single timepoint in a 3.5-h perfusion ow imaging experiment. Cells and OMEs are visu- alized with the red membrane stain FM 4-64FX. (B) Statistics of OME production from over 5,400 cells in 4 replicate 3.5-h perfusion culture experiments illustrates that a majority (78%) of cells pro- duce OMEs visible over time. The remaining cells were seen with only outer membrane vesicles (OMVs), or nothing at all. Error bars show mean ± SEM (Scale bar: 10 m). . . . . . . . . . . . . . . . . 33 2.2 Surface attachment is sucient to induce production of outer membrane extensions. (A) Diagram illustrates experi- mental procedure. (B,C) Microscopy images of S. oneidensis MR- 1 cells and membrane extensions (white arrows) labeled with the red membrane stain FM 4-64FX. Time (t = 0 min) indicates estimated time of cells contacting the glass surface. (B) Demonstrates pro- duction of outer membrane extensions (OMEs) by surface-attached cells in the aerated glass-bottomed reactor. Here, 6.5{7.5 ppm = 400{470 mM O 2 . (C) Demonstrates OME production by plank- tonic cells from the reactor which were transferred to a new coverslip surface after events in (B) were conrmed (Scale bars: 5 m). . . . 34 x 2.3 Outer membrane extensions are produced quickly by plank- tonic cells in rich aerobic medium soon after cell-to-surface contact. Diagram illustrates experimental procedure. Microscopy images depict S. oneidensis MR-1 cells and outer membrane exten- sions (OMEs, white arrows) labeled with the red membrane stain FM 4-64FX. Time (t = 0 min) indicates estimated time of cells contacting the glass surface. (Scale bars: 5 m.) . . . . . . . . . . . 36 2.4 Outer membrane extensions are produced in a variety of surface-attached conditions, regardless of medium compo- sition, surface chemistry, agitation, or aeration. S. oneiden- sis MR-1 cells and membrane extensions are visualized by mem- brane stains FM 4-64FX (red), FM 1-43FX (green), or TMA-DPH (blue). Unless otherwise specied, cells were imaged at the sur- face of glass coverslips with ow or agitation of oxygen-limited min- imal medium. Outer membrane extensions are observed in (A) oxygen-limiting perfusion conditions, described previously [24, 25], (B) oxygen-abundant, high cell density conditions, (C) oxygen- abundant, low cell density conditions, (D) in rich (LB) medium, (E) in buer (PBS), (F) on a carbon-coated electron microscopy grid, and (G) without ow or agitation. (Scale bars: 5 m.) . . . . 37 2.5 Cells can produce multiple outer membrane extensions per cell. (A-B) Transmission electron microscopy images of chemically xed, negatively stained S. oneidensis MR-1 cells and membrane extensions (Scale bars: 200 nm.) (C-D) Atomic force microscopy tapping mode phase images of chemically xed cells and membrane extensions. (Scale bars: 5 m.) . . . . . . . . . . . . . . . . . . . . 40 2.6 Outer membrane extensions can reach a length of >100 m, produced at a rate >40 m/h. Image sequence from time- lapse uorescence microscopy of surface-attached perfusion cultured S. oneidensis MR-1 Mtr/mtrB/mtrE cells. Cells and outer membrane extensions (OMEs) were visualized with the red mem- brane stain FM 4-64FX. Time (t = 0) marks time since OME pro- duction (white arrow). Images depict progression of events, such as (A) cell contacting surface, (B) OME rst visible (white arrow), (C) OME elongates and begins to fold, (D) further OME elonga- tion, (E) OME reaches longest point visible during this experiment (Scale bars: 10 m). . . . . . . . . . . . . . . . . . . . . . . . . . . 41 xi 2.7 Redox components are present only on outer membrane extensions, not pili or agella. Histochemical redox-dependent staining with 3,3'-diaminobenzidine (2.5 h staining step) and trans- mission electron microscopy distinguishes between types of extra- cellular laments in S. oneidensis MR-1. Images depict dark pre- cipitate (yellow arrows and lines) labeling only outer membrane extensions, but not adjacent extracellular structures (A) pili (white arrow), and (B) agella (black arrow). Cells are indicated by aster- isk symbols (*) (Scale bars: 200 nm). . . . . . . . . . . . . . . . . . 43 2.8 Presence of multiheme cytochromes important for extra- cellular electron transfer leads to signicantly higher fre- quency and intensity of redox-dependent staining on outer membrane extensions. (A{D) Transmission electron microscopy images depict outer membrane extensions (OMEs, white arrows) stained by 3,3'-diaminobenzidine (DAB; 1 h staining step) in wild type and cytochrome-decient (Mtr/mtrB/mtrE) S. oneiden- sis MR-1 cells. Cells are indicated by asterisk symbols (*). (A) Wild type OMEs are stained by DAB precipitate. (B,C) Mutant OMEs treated by DAB exhibit varying degrees of staining. (D) Wild type OMEs in chemical controls where H 2 O 2 was omitted appear unstained aside from negative stain. (E) Frequency of stain- ing displayed by OMEs in wild type, Mtr/mtrB/mtrE mutant, and wild type chemical control where H 2 O 2 was omitted. 2.4-fold more OMEs were stained in wild type than in the mutant (p < 0.0001). Statistical signicance was determined by p-value from Pearson's chi-squared test. (F) Intensity of staining displayed by OMEs is 3.6-fold higher in wild type than in Mtr/mtrB/mtrE mutant (p < 0.0001). Statistical signicance was determined by two-tailed p-value from Student's t-test, two-sample assuming equal variances. Error bars show mean ± SEM (Scale bars: 200 nm). . . . 45 3.1 Labeling strategy. (A) Structure of MtrC (PDB ID 4LM8) illus- trates location of biotin acceptor peptide (AP) tag, fused to C- terminus of MtrC (or OmcA) near Heme 10. Hemes and por- phyrin rings are colored orange, and AP tag is colored blue. (B) Schematic of labeling strategy. The biotin acceptor peptide (AP: GLNDIFEAQKIEWHE) is fused to MtrC (or OmcA). At the cell surface, biotin ligase BirA biotinylates the AP, and quantum dot (QD)-streptavidin conjugates (or other streptavidin conjugates) bind the biotinylated MtrC-AP (or OmcA-AP). . . . . . . . . . . . . . . 48 xii 3.2 Schematic of design for MtrC-AP construct, dubbed pMtrC- AP. (A) Plasmid design. MtrC-AP gene fusion was inserted between restriction sites for XhoI and XbaI in pBBR1-MCS2 plasmid [98] with kanamycin resistance. (B) Insert design. Here, the C-terminus of MtrC was fused to a 45-bp \AP Tag" encoding the 15-amino acid biotin acceptor peptide (AP: GLNDIFEAQKIEWHE) from E. coli, as described in [99, 102]. Insert was generated from S. oneidensis MR-1 genomic DNA template using three rounds of overhang PCR with primers for MtrC listed in Table 3.2. DNA insert included an XhoI restriction site, 118 bp upstream of mtrC (including native promoter), protein-coding region for MtrC, a short glycine-serine linker, and the 45-bp biotin acceptor peptide (AP) tag sequence just before the stop codon and XbaI restriction site. Total length of DNA inserted into pBBR1-MCS2 plasmid was 2185 bp for a nal construct size of 7276 bp. . . . . . . . . . . . . . . . . . . . . . . . . 51 3.3 Schematic of design for OmcA-AP construct, dubbed pOmcA- AP. (A) Plasmid design. OmcA-AP gene fusion was inserted between restriction sites for XhoI and XbaI in pBBR1-MCS2 plasmid [98] with kanamycin resistance. (B) Insert design. Here, the C-terminus of OmcA was fused to a 45-bp \AP Tag" encoding the 15-amino acid biotin acceptor peptide (AP: GLNDIFEAQKIEWHE) from E. coli, as described in [99, 102]. Insert was generated from S. oneidensis MR-1 genomic DNA template using three rounds of overhang PCR with primers for OmcA listed in Table 3.2. DNA insert included an XhoI restriction site, 114 bp upstream of omcA (including native promoter), protein-coding region for OmcA, a short glycine-serine linker, and the 45-bp biotin acceptor peptide (AP) tag sequence just before the stop codon and XbaI restriction site. Total length of DNA inserted into pBBR1-MCS2 plasmid was 2373 bp for a nal construct size of 7464 bp. . . . . . . . . . . . . . . . . . . . . . . . . 52 xiii 3.4 Heme staining of protein gels via peroxidase activity assay reveals heme in newly tagged cytochromes MtrC-AP and OmcA-AP. Redox-dependent staining using a 3,3'-diaminobenzidine (DAB) and hydrogen peroxide (H 2 O 2 ) peroxidase activity assay. SDS-PAGE and subsequent staining was performed using whole cell lysate from liquid cultures of respective S. oneidensis strains labeled at the top of each lane. Also labeled are relevant bands in protein ladder (80 kDa, 58 kDa), as well as the approximate positions of proteins of interest (MtrC, OmcA) and fumarate reductase (FccA) which is present in all samples. Lanes 4 and 5 in both gels con- tain a dark band conrming the presence of heme associated with protein of interest (AP-tagged MtrC or OmcA). Wild type sam- ple (Lane 1) was used as a positive control; respective gene dele- tion mutants mtrC and omcA (Lanes 2 and 3) are included as negative controls missing protein of interest MtrC or OmcA; and included as a secondary negative control is a cytochrome mutant (Mtr/mtrB/mtrE, Lane 6) missing a total of 8 periplasmic and outer membrane-associated cytochromes (including proteins of interest MtrC and OmcA). . . . . . . . . . . . . . . . . . . . . . . . 66 3.5 Key labeling controls demonstrate successful and specic labeling of MtrC. (A) Western blot labeling control for MtrC where key parts of the labeling process were systematically omit- ted. When using streptavidin (streptavidin-horseradish peroxidase, HRP) to probe for biotinylated proteins, a thick dark band of biotiny- lated MtrC-AP is detected only in Lane 5 when all key compo- nents are present. The faint band slightly below labeled MtrC- AP (approx. 79.6 kDa) and present in all samples is an endoge- nously biotinylated protein (acetyl-CoA carboxylase, approx. 76 kDa). (B) Microscopy labeling control for MtrC where key parts of the labeling process were systematically omitted. Top row are wide- eld (WF) images showing many cells in each sample. Bottom row images show uorescence (Fl) signal from streptavidin-conjugated Alexa Fluor 647 (SA-AF647) that was used to detect biotinylated MtrC-AP; uorescence labeling was detected strongly in the bot- tom right image, and only when all key labeling components were present. All microscopy images are approx. 36.5 m by 36.5 m. . . 68 xiv 3.6 Key labeling controls demonstrate successful and specic labeling of OmcA. (A) Western blot labeling control for OmcA where key parts of the labeling process were systematically omit- ted. When using streptavidin (streptavidin-horseradish peroxidase, HRP) to probe for biotinylated proteins, a thick dark band of biotiny- lated OmcA-AP is detected only in Lane 5 when all key compo- nents are present. The faint band slightly below labeled OmcA-AP (approx. 87 kDa) and present in all samples is an endogenously biotinylated protein (acetyl-CoA carboxylase, approx. 76 kDa). (B) Microscopy labeling control for OmcA where key parts of the labeling process were systematically omitted. Top row are wide- eld (WF) images showing many cells in each sample. Bottom row images show uorescence (Fl) signal from streptavidin-conjugated Alexa Fluor 647 (SA-AF647) that was used to detect biotinylated OmcA-AP; uorescence labeling was detected strongly in the bot- tom right image, and only when all key labeling components were present. All microscopy images are approx. 36.5 m by 36.5 m. . . 69 4.1 Lateral diusion and labeling strategy. (A) Schematic of diusion-assisted electron hopping along the Shewanella oneidensis MR-1 outer membrane. Lateral motion of multiheme cytochromes (diusion coecient D phys ) leads a collision-exchange mechanism of inter-protein electron transport over large distances. Red spots represent hemes in multiheme cytochromes. Labeled proteins are outer membrane cytochromes MtrC and OmcA, outer membrane- associated periplasmic cytochrome MtrA, and outer membrane porin MtrB. (B) Structure of MtrC (PDB ID 4LM8) illustrates location of biotin acceptor peptide (AP) tag, fused to C-terminus of MtrC (or OmcA) near Heme 10 as described in Chapter 3. Hemes and porphyrin rings are colored orange, and AP tag is colored blue. (C) Schematic of labeling strategy, established in our system as described in Chapter 3. The biotin acceptor peptide (AP: GLN- DIFEAQKIEWHE) is fused to MtrC (or OmcA). At the cell sur- face, biotin ligase BirA biotinylates the AP, and QD-streptavidin conjugates bind the biotinylated MtrC-AP (or OmcA-AP). . . . . . 75 xv 4.2 Imaging and single molecule tracking of quantum-dot labeled OmcA using total internal re ection uorescence (TIRF) microscopy. (A) Snapshot of OmcA-AP trajectories (white) in multiple cells (cyan). Trajectories from 1.5 min of time-lapse microscopy (40 ms/frame) were overlaid onto the corresponding mean intensity projection image of cells labeled with lipid membrane dye FM 1-43FX. Trajectories in white dashed boxes are blown up in panels B and C. Scale bar: 2m. (B-C) Some example trajectories from the two cells outlined in panel A. Scale bars: 500 nm. (D-E) Streptavidin-coated QD705 was used to detect exogenously biotiny- lated OmcA-AP (red). Cell membrane and membrane extensions are labeled by FM 1-43FX (cyan). (D) Trajectories from a sin- gle quantum dot labeled OmcA-AP as it moved along the surface of a cell. Here, the quantum dot signal (red) and its trajectories (white) are overlaid with the mean intensity projection image of the cell (cyan). For clarity, only the rst frame of quantum dot signal is shown; trajectories are from the entire video (approx. 86 s, 40 ms/frame). Scale bar: 500 nm. (E) Snapshot of OmcA-AP trajectories overlaid on an outer membrane extension. Trajectories (white) are from6 min (40 ms/frame) of time-lapse microscopy tracing several quantum dot-labeled OmcA-AP (red; mean inten- sity projection image) on a membrane extension that appears to connect two cells (cyan; mean intensity projection image). Scale bar: 500 nm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 4.3 Individual mean squared displacement (MSD) analyses for 3 example trajectories. Example trajectories and corresponding individual MSD curves show a range of diusion by OmcA-AP on cell surface, in order of generally slower and more conned (smaller diusion coecients D, connement radii R) to faster and less con- ned (bigger D, R). The duration of each trajectory is also labeled above each trajectory image. Error bars in plots show ± SEM. Red curve is the conned diusion t (Eqn. 4.2), from which correspond- ing D and R values were calculated. Blue line is the free (Brownian) diusion model (Eqn. 4.1) t to the rst 3t in each trajectory's MSD, from which a corresponding D value was calculated. . . . . . 92 xvi 4.4 Distribution of diusion coecients from individual MSD analyses for MtrC-AP (blue) and OmcA-AP (orange) dif- fusing on the cell surface. To generate these histograms, instan- taneous D values were calculated for all individual trajectories by tting the free diusion model (Eqn. 4.1) to the rst 3t in each trajectory's MSD, as done for the blue t lines in Figure 4.3. For illustration, the trajectories from Figure 4.3A-C correspond to the histogram positions labeled by blue arrows. Both histograms show 1 major Gaussian distribution for both MtrC and OmcA. These his- tograms represent an MtrC-AP cell surface dataset containing 5,110 trajectories and OmcA-AP cell surface dataset containing 6,155 tra- jectories, as observed in 500-1,000 cells each. . . . . . . . . . . . . . 93 4.5 Ensemble mean squared displacement (MSD) analysis shows overall conned diusion behavior by MtrC (blue) and OmcA (orange) on the cell surface. Y-axis shows mean displacement squared (r 2 ) for each time lag (t) on the X-axis. Fitting the plots with a conned diusion model (Eqn. 4.2) yields diusion coe- cients D and connement radii R as labeled. These curves repre- sent an MtrC-AP cell surface dataset containing 5,110 trajectories and OmcA-AP cell surface dataset containing 6,155 trajectories, as observed in 500-1,000 cells each. Error bars show r 2 p N , where N is the number of displacements measured for a given t. . . . . . . 94 4.6 Diusion analyses for MtrC (blue) and OmcA (orange) dif- fusion on cell surface assuming 2 major populations of con- ned diusion behavior. These plots suggest a larger, slower, more conned population and a smaller, faster, less conned popu- lation of diusion by each protein on the cell surface. The ensemble MSD curves were plotted as mean displacement squared r 2 as a function of time lag t for respective major populations of dius- ing behavior. Fitting these curves with a conned diusion model (Eqn. 4.2) yields diusion coecients D and connement radii R as labeled on the respective plots. Percentages indicate the respective fractions belonging to each population determined by PDSD analy- sis, as described in [119, 121]. These curves represent an MtrC-AP cell surface dataset containing 5,110 trajectories and OmcA-AP cell surface dataset containing 6,155 trajectories, as observed in 500- 1,000 cells each. Error bars show SEM p N(%N) , where %N is the fraction of all displacements N that had been allocated to the given sub-population for a given t. . . . . . . . . . . . . . . . . . . . . . 96 xvii 4.7 Distribution of individual diusion coecients for MtrC- AP (blue) and OmcA-AP (orange) diusing on the sur- face of outer membrane extensions (OMEs). Individual dif- fusion coecients were calculated by tting a free diusion model (Eqn. 4.1) to the rst 3t in each trajectory's mean squared dis- placement (MSD). These distributions, unlike those for cell surface datasets, now contain a noticeable right-side tail suggesting more than 1 Gaussian population. Thus, each plot here also shows t lines for 2 Gaussian distributions (blue or orange dashed lines) and their cumulative t (black line). These histograms represent two datasets, MtrC-AP on OMEs (945 trajectories from 5 OMEs) and OmcA-AP on OMEs (4,633 trajectories from 22 OMEs). . . . . . . 99 4.8 Ensemble Mean Squared Displacement (MSD) analyses for MtrC and OmcA on outer membrane extensions (OMEs) demonstrate 2 major populations of conned diusion behav- ior. Left (A-B): MtrC-AP. Right (C-D): OmcA-AP. (A-D) Ensemble MSD curves were plotted as mean displacement squared r 2 as a function of time lag t for respective major populations of diusing behavior: (A) MtrC-AP, slow population; (B) MtrC-AP, fast population; (C) OmcA-AP, slow population; and (D) OmcA- AP, fast population. Fitting these curves with a conned diusion model (Eqn. 4.2) yields diusion coecients D and connement radii R as labeled on the respective plots. Percentages indicate the respective fractions belonging to each population determined by PDSD analysis, as described in [119, 121]. The ensemble MSD plots suggest a larger, slower, more conned population and a smaller, faster, less conned population of diusion by each protein on outer membrane extensions. Error bars show SEM p N(%N) , where %N is the fraction of all displacements N that had been allocated to the given sub-population for a given t. These ensemble MSD curves represent two datasets, MtrC-AP on OMEs (945 trajectories from 5 OMEs) and OmcA-AP on OMEs (4,633 trajectories from 22 OMEs). 100 xviii 4.9 Ensemble mean squared displacement (MSD) analysis shows overall conned diusion behavior by MtrC (blue) and OmcA (orange) on outer membrane extensions (OMEs). Y-axis shows mean displacement squared (r 2 ) for each time lag (t) on the X-axis. Fitting the plots with a conned diusion model (Eqn. 4.2) yields diusion coecients D and connement radii R as labeled. Error bars show r 2 p N , where N is the number of displacements measured for a given t. These curves represents two datasets, MtrC-AP on OMEs (945 trajectories from 5 OMEs) and OmcA-AP on OMEs (4,633 trajectories from 22 OMEs). . . . . . . . . . . . . 102 4.10 Distribution of diusion coecients from individual MSD analyses for MtrC-AP (blue) and OmcA-AP (orange) dif- fusing on the cell surface, when cells were ooded by200- fold excess biotin after addition of streptavidin-coated quan- tum dots. Diusion coecients were calculated by tting a free diusion model (Eqn. 4.1) to the rst 3t in each trajectory's MSD. Both histograms show 1 major Gaussian distribution for both MtrC and OmcA. Diusion from biotin ood experiments are consis- tent with previous datasets of diusion on cell surface (Figure 4.4), suggesting that the quantum dot probes are not binding to mul- tiple protein targets and subsequently hindering diusion. These histograms represent an MtrC-AP dataset containing 69,793 tra- jectories and OmcA-AP dataset containing 58,543 trajectories, as observed in >1,000 cells each. . . . . . . . . . . . . . . . . . . . . . 107 4.11 Simulation results of overall electron transport (ET) along the surface of whole cells or membrane extensions, based on experimentally measured diusion coecients. ET rates on the Y-axis are plotted on a log scale as a function of the fractional loading of redox carriers (X ) on (A) the surface of a whole cell (2 m long and 0.5 m in diameter) or (B) the surface of an outer membrane extension (1 m long and 100 nm in diameter). These results come from simulations using either t e = 10 -5 s (lled shapes) or 10 -6 s (unlled shapes) for a range of experimentally derived dif- fusion coecients D phys = 10 -1 m 2 /s (black squares) or 10 -2 m 2 /s (red circles). Note: All simulations represented in this gure were designed and performed by co-author Sahand Pirbadian. . . . . . . 109 xix List of Tables 1.1 Some of the known electron donors and acceptors used by S. oneidensis MR-1. External solid electron donors and accep- tors are indicated in bold. . . . . . . . . . . . . . . . . . . . . . . . 5 3.1 Strains and plasmids used in this study. . . . . . . . . . . . . 49 3.2 Primers used in this study. . . . . . . . . . . . . . . . . . . . . 54 3.3 Ingredients used for overhang PCR. Volume is for a 25 L reaction. For specic primers used in each step, see Table 3.5. . . . 55 3.4 Thermocycler protocol for overhang PCR. For specic anneal- ing temperatures used for each primer pair, see Table 3.5. . . . . . . 55 3.5 Specic primers and annealing temperatures used in 3 con- secutive rounds of overhang PCR. Annealing temperatures were chosen for template-binding portions of each primer (Anneal- ing 1) or for the full primer set (Annealing 2) according to the polymerase manufacturer (http://tmcalculator.neb.com/#!/main). . 55 3.6 Ingredients used for colony PCR. Volume is for a 50 L reac- tion. For colony PCR, MtrC or OmcA Forward primers were used along with MtrC/OmcA Reverse 1 primer to amplify DNA from putatively transformed E. coli DH5 colonies. . . . . . . . . . . . . 58 3.7 Thermocycler protocol for colony PCR. Annealing temper- atures were chosen for template-binding portions of each primer (Annealing 1) or for the full primer set (Annealing 2) according to the polymerase manufacturer (http://tmcalculator.neb.com/#!/main). 58 4.1 Strains used in this study. . . . . . . . . . . . . . . . . . . . . . 77 4.2 Summary of overall diusion measurements. Diusion coe- cients D and connement radii R listed here were obtained by tting the conned diusion model (Eqn. 4.2) to ensemble mean squared displacement (MSD) curves for the listed datasets, corresponding to Figures 4.5 and 4.9. All diusion coecients are reported as ± standard deviation of the t value. OME: outer membrane extension.104 xx 4.3 Summary of diusion measurements, assuming 2 popula- tions. Percentages indicate the respective fractions belonging to each population determined by probability distribution of square displacement (PDSD) analysis in SLIMfast. Diusion coecients D and connement radii R described here were obtained by tting the conned diusion model (Eqn. 4.2) to ensemble mean squared displacement (MSD) curves for the listed datasets, corresponding to Figures 4.6 and 4.8. All diusion coecients are reported as ± standard deviation of the t value. OME: outer membrane extension.105 xxi Abstract Electron transfer is a fundamental aspect of life, driving key biological energy con- version processes such as respiration and photosynthesis. While the mechanisms of biological electron transfer over nanometer length scales are now well established, discoveries of fast microbial electron transfer across micron- to centimeter-scale distances are challenging our state of knowledge and are of particular interest for the purpose of wiring microbes to electrodes in bioelectrochemical renewable energy technologies. Dissimilatory metal-reducing bacteria, including Shewanella oneidensis MR-1, can gain energy by extracellular electron transport (EET) to external solids, such as minerals or electrodes, which substitute for oxygen or other soluble electron acceptors in respiration. In S. oneidensis, a network of mul- tiheme cytochromes facilitate EET by forming biological electron conduits that bridge the otherwise insulating inner membrane, periplasm, and outer membrane to external redox-active surfaces. These microbes also form outer membrane exten- sions (OMEs), from vesicles to micrometer-scale appendages known as bacterial nanowires that are proposed to aid in their use of external electron acceptors. The cytochromes abundant on the cell outer membrane and important in EET are simi- larly found along its membrane extensions and give these OMEs their conductivity (at least as measured in dry, chemically xed conditions). xxii These proposed bacterial nanowires, which can be several times the cell length, are thereby thought to extend EET to more distant electron acceptors. However, it was still unclear why these extensions form, and to what extent they contribute to respiration in living cells. In Chapter 2, I investigated physical contributors to their formation using in vivo uorescence microscopy. While previous studies focused on the display of S. oneidensis OMEs as a response to oxygen limitation, I found that cell-to-surface contact is sucient to trigger the production of OMEs, including some that reach >100 m in length, irrespective of medium composi- tion, agitation, or aeration. To visualize the extent of heme redox centers along OMEs, and help distinguish these structures from other extracellular laments, I also performed histochemical redox-dependent staining with transmission electron microscopy on wild type and cytochrome-decient strains. I demonstrated that redox-active components are limited to OMEs and not present on other extra- cellular appendages, such as pili and agella. I also observed that the loss of 8 functional periplasmic and outer membrane cytochromes signicantly decreased both the frequency and intensity of redox-dependent staining found widespread on OMEs. These results improved our understanding of the environmental conditions that in uence the formation of S. oneidensis OMEs, as well as the distribution and functionality of EET components along extracellular appendages. While the role of multiheme cytochromes in transporting electrons across the cell wall is well established, these cytochromes were also recently found to facil- itate long-distance (micrometer-scale) redox conduction along outer membranes and across multiple cells bridging electrodes. Recent studies proposed that long- distance conduction arises from the interplay of electron hopping and cytochrome diusion, which allows collisions and electron exchange between cytochromes along membranes. However, the diusive dynamics of the multiheme cytochromes had xxiii never been observed or quantied in vivo, making it dicult to assess their hypoth- esized contribution to the collision-exchange mechanism. In Chapters 3 and 4, I used targeted quantum dot labeling, total internal re ection uorescence microscopy, and single-particle tracking to quantify the lateral diusive dynamics of the outer membrane-associated decaheme cytochromes MtrC and OmcA, two key compo- nents of EET in S. oneidensis. I observed conned diusion behavior for both quantum dot-labeled MtrC and OmcA along cell surfaces (diusion coecients D MtrC = 0.0306± 0.0082m 2 /s, D OmcA = 0.0121± 0.0019m 2 /s) and the mem- brane extensions thought to function as bacterial nanowires. I found that these dynamics can trace a path for electron transport via overlap of cytochrome trajec- tories, consistent with our recently proposed long-distance conduction mechanism. The measured dynamics informed kinetic Monte Carlo simulations that combine direct electron hopping and redox molecule diusion, revealing signicant electron transport rates along cells and membrane nanowires. xxiv Chapter 1 Introduction 1.1 Background 1: Dening the system of study This section takes a simplied approach to dening the system of study and explores some of its unique or otherwise important characteristics. The following research chapters (Chapters 2, 3, and 4) will build o of these introductory topics and take a more aggressive approach to dening the current state of knowledge relevant to the issues addressed in each study. Note: Select portions of this section, particularly subsection 1.1.5, were adapted from [1]: Grace W. Chong, Amruta A. Karbelkar, and Mohamed Y. El-Naggar. Nature's Conductors: What can microbial multi-heme cytochromes teach us about electron transport and biological energy conversion? Current Opinion in Chemical Biology, 47:7{17, 2018. ISSN 18790402. doi: 10.1016/j.cbpa.2018.06.007. 1.1.1 From electron transfer to extracellular respiration and its applications Electron transfer is a fundamental part of how all living things extract energy from their environment. In most respiratory organisms, including humans, this energy- harvesting process works by the transfer of electrons from an electron donor (food) to an electron acceptor (e.g. oxygen), through the cellular electron transport chain (ETC) [2]. The electron transfer reactions (also called redox reactions) between 1 components of the ETC move electrons from higher-energy donors to lower-energy acceptors, translating this energy into a chemical gradient across a biological mem- brane, which in turn drives the production of energy molecules that power the cell. Figure 1.1 ranks several types of common electron donors and acceptors in order of higher energy to lower energy. Substances listed at the top of this tower tend to be better electron donors, while those closer to the bottom tend to be better electron acceptors. Electron ow is energetically favorable if it occurs in the direction of top to bottom on this list, as indicated by the arrows. More energy is gained by respiration involving electron donors and acceptors that are further apart on this table. For organisms capable of aerobic respiration, oxygen is the preferred terminal electron acceptor, since it is the most energetically favorable. For many respiratory organisms, maintaining the electron ow required for life depends on the availability of soluble electron donors (e.g. organic molecules) and acceptors (e.g. O 2 ) that can enter cells to interact with bioenergetic machinery. In the absence of oxygen, many microorganisms have evolved other mechanisms to harvest energy by utilizing other combinations of electron donors and acceptors. A few of these diverse types of respiration are labeled in Figure 1.1. Furthermore, soluble electron donors and acceptors are not always available. To solve this prob- lem, some microbes have equipped extracellular electron transfer (EET) conduits that bridge intracellular reactions to insoluble oxidants or reductants, including electrodes, outside the cells. Microbes evolved this strategy to access abundant redox-active respiratory electron acceptors and donors (e.g. S, Fe, Mn) found in minerals [3]. A fundamental understanding of microbial EET has many promising applica- tions, especially in the development of bioelectrochemical systems (BES) [4, 5]. BES generally rely on microbes catalyzing activity on an anode, cathode, or both, 2 Figure 1.1: Redox tower listing several common electron donors and acceptors. Chemicals are listed by their standard reduction potential (E o ', at pH 7 and standard state) and arranged in order of higher energy (more negative E o ') to lower energy (more positive E o '). Positive E o ' means it is energetically favorable for a reaction to proceed in the order written, e.g. O 2 reduced to H 2 O. Negative E o ' means it is energetically favorable for the reaction to proceed in the other direction, e.g. H 2 oxidized to 2H + . It is energetically favorable for electrons to ow from more negative to more positive E o '. Those closer to the bottom of this tower have the greatest anity for electrons and are thus the best elec- tron acceptors. The red arrows classify some organisms by the type of electron donors/acceptors used. Figure courtesy of Moh El-Naggar. to achieve a desired outcome, such as: (1) electricity generation via cell-to-electrode EET in microbial fuel cells [6]; (2) wastewater treatment powered by microbial oxidation of organic wastes [7]; (3) bioremediation, where toxic contaminants are transformed into less soluble, less toxic forms [8]; and (4) production of biofu- els or other value-added products, as done by electrode-to-cell EET in microbial electrosynthesis cells [9]. It may also shed light on what microbes capable of extracellular respiration are doing in the environment, where metabolically ver- satile microbes can couple the oxidation of various electron donors (e.g. carbon 3 compounds) to the reduction of a variety of electron acceptors (e.g. minerals), contributing to global biogeochemical cycles. Lastly, elucidating the capabilities of mineral-reducing microbes is of potential interest to the eld of astrobiology (understanding how life may exist on other planets, especially where oxygen is lacking but minerals are abundant). 1.1.2 Shewanella oneidensis MR-1: Discovery and basic features In this thesis, I focus on S. oneidensis MR-1, a Gram-negative, facultative anaero- bic heterotrophic bacterium rst reported in 1988 and one of the rst microbes to be found capable of extracellular electron transfer [10]. Since then, it has become an extensively-studied model organism for EET. It gained its strain name \MR- 1" due to its ability to reduce solid manganese compounds (MR = manganese reducer), since its initial discovery in Lake Oneida, New York was prompted by circumstances of unusual changes in the lake's chemical composition that could not be described simply by known inorganic processes [10]. Its characteristic as a facultative anaerobe makes it conveniently grown in laboratory conditions, since it can be grown in both aerobic and anaerobic conditions, and exposure to oxygen is not detrimental to the health of the organism. In addition to soluble oxygen gas and insoluble manganese minerals, S. oneidensis MR-1 is known to harvest energy from its environment by utilizing a vast array of electron donors and acceptors [11], including those brie y listed below in Table 1.1. However, its characteristic as a Gram-negative bacterium means that its elec- tron transport chain is located on the cell's inner membrane, and this electron transport chain is separated from the outside world (and external inorganic elec- tron donors or acceptors) by otherwise electrically insulating layers of periplasmic 4 space and outer membrane. In S. oneidensis, these insulating layers are about 25 nm thick (approx. 20 nm thick periplasm and 5 nm thick outer membrane) [12], which is further than the2 nm distance required for ecient electron tunneling [2]. So, how does S. oneidensis bridge these insulating layers and connect the electron transport chain on the inner membrane with solid electron donors and acceptors outside the cell? Table 1.1: Some of the known electron donors and acceptors used by S. oneidensis MR-1. External solid electron donors and acceptors are indicated in bold. Electron donors Electron acceptors Formate O 2 Lactate NO 3 - , NO 2 - Pyruvate Mn(IV) Amino Acids Mn(III) H 2 Fe(III) Electrodes Fumarate DMSO TMAO S 2 O 3 2- SeO 3 2- TeO 3 2- U(VI) Cr(VI), Tc, As.... Electrodes 1.1.3 Multiheme cytochromes and extracellular electron transfer (EET) in S. oneidensis The aforementioned respiratory versatility is made possible in S. oneidensis by a series of multiheme c-type cytochromes that transport electrons from the electron transport chain on the inner membrane, across the electrically insulated periplas- mic space and outer membrane, to solid materials outside the cell, in a process 5 known as EET [1, 11, 13]. Brie y, cytochromes are proteins that contain heme cofactors, small molecules with an iron center that enables the protein to catalyze redox reactions and participate in electron transfer. Multiheme cytochromes con- tain a number of hemes (in S. oneidensis, up to 10 hemes), usually arranged just a few angstroms apart as if to form a sort of molecular `wire' within the protein itself. To illustrate this characteristic, the heme arrangement of two multiheme cytochromes important for EET in S. oneidensis are shown in Figure 1.2, from [14, 15]. As noted in this gure, the hemes (and their iron centers) within these cytochromes are within the2 nm distance required for ecient electron tunneling [2] and thus enable heme-to-heme electron hopping within individual proteins. Extensive studies have revealed an EET network of multiheme cytochromes that enable extracellular respiration by bridging the cell envelope, including the inner membrane tetraheme menaquinol dehydrogenase CymA, periplasmic cytochromes such as STC (small tetraheme cytochrome, aka CctA) or FccA (fumarate reductase), outer membrane porin-cytochrome complexes such as MtrAB and cell surface cytochromes such as MtrC and OmcA [13, 15, 16]. MtrA is a decaheme cytochrome located on periplasmic side of the outer membrane and connected to the surface by the transmembrane porin MtrB, where the outward- most heme in MtrA can then interact with cell surface cytochromes such as MtrC and OmcA [15, 16], which in turn act as an external interface between the cell and extracellular electron acceptors [13]. The decaheme outer membrane-associated MtrC and OmcA are largely extracellularly exposed [15], attached to the cell sur- face by a lipidated cysteine at the N-terminus [17]. As depicted in Figure 1.2C-D, both MtrC and OmcA also have a staggered cross heme arrangement, which allows electron transfer to redox partners on either side as well as in the direction per- pendicular to the cell surface. The crystal structure of a porin-cytochrome conduit 6 was recently solved for MtrABC in Shewanella baltica OS185, another species of Shewanella (Figure 1.3, from [15]). Figure 1.2: Crystal structures and heme arrangement of two decaheme cytochromes important in extracellular electron transport and metal reduction in S. oneidensis. (A-B) Crystal structure and heme arrangement of extracellular facing outer membrane cytochrome MtrC (PDB 4LM8) obtained from S. oneidensis MR-1, gure from [14]. (C-D) Crystal structure and heme arrangement of outer membrane-associated periplasmic cytochrome MtrA obtained from Shewanella baltica OS185, gure from [15]. 7 While MtrABC (and OmcA) are the most well characterized for their estab- lished roles in electron transport across the outer membrane, the S. oneidensis EET network also contains three other known (MtrDEF, DmsEF) or predicted (SO4359- 60) porin-cytochrome complexes with similar structure and function [11, 13]; two (MtrDEF and SO4359-60) have been shown to have some impact on extracellular respiration in the absence of MtrABC [18, 19], while the third (DmsEF) shuttles electrons across the outer membrane for the reduction of DMSO [20]. Once elec- trons have reached the surface of the cell, they can be transferred to extracellular electron acceptors by direct contact with these cell surface cytochromes, or by indirect contact via soluble redox shuttles, such as avins [1]. An illustration of outward EET pathways proposed for S. oneidensis is depicted in Figure 1.4. 8 Figure 1.3: Structure of a biological transmembrane wire. MtrABC porin-cytochrome complex crystallized from S. baltica OS185, including decaheme cytochromes MtrC (blue) and MtrA (pink) as well as the porin MtrB (yellow green) that allows them to interact across the membrane. The two hemes at the inter- face of MtrA and MtrC are separated by a distance of8 A. All other inter-heme distances are <8 A. Total length of transmembrane wire is185 A. Figure from [15]. 9 Figure 1.4: Outward extracellular electron transfer (EET) pathways pro- posed for S. oneidensis MR-1. Cytochromes on the inner membrane (where the electron transport chain is also located) can transfer electrons to soluble periplasmic cytochromes, which can then relay electrons to the outer membrane via porin-cytochrome conduits to extracellular electron acceptors. Once electrons are exported to the cell surface, terminal electron acceptors (including solid-phase minerals and electrodes) can be reduced by direct contact through extracellular multiheme cytochromes (MHCs), MHC-bound redox cofactors, or MHCs found on outer membrane extensions, or indirectly through diusion of soluble electron carriers. Black arrows illustrate possible electron ow. Red shape represents inner membrane cytochromes (e.g. CymA), orange shape represents periplasmic solu- ble cytochromes (e.g. STC, FccA), yellow/gray shapes represent outer membrane porin-cytochrome conduits (e.g. Mtr/Omc pathway), and green circles represent soluble redox-active molecules (e.g. avins). Abbreviations: electron transfer (ET), outer membrane (OM), inner membrane (IM). Figure adapted from [1]. 10 1.1.4 Multiheme cytochromes, bacterial nanowires, and long-distance electron transfer (ET) inS. oneidensis In addition to EET from inside to outside the cell (or vice versa), EET com- ponents in S. oneidensis can also enable long-distance lateral electron transport along membranes and across multiple cells, as demonstrated recently with elec- trochemical gating measurements of electron conduction through S. oneidensis cellular monolayers connecting interdigitated electrodes [21]. However, though this micrometer-scale conduction was dependent on multiheme cytochromes in the Mtr/Omc EET pathway [21], the exact mechanism of long-distance electron trans- port along membrane surfaces is yet unknown. It turns out that S. oneidensis also produces micrometer-scale \bacterial nanowires", which were rst discovered in 2006, where they were rst thought to form in response to oxygen-limiting conditions [22]. Since their discovery, they have been proposed to act as bacterial nanowires that contribute to respiration of distant electron acceptors (a proposed EET pathway depicted in Figure 1.4). After- wards, atomic force microscopy revealed that these appendages are conductive, at least in the dry, xed conditions, and this conductivity was dependent on the same multi-heme cytochromes important in EET [23]. It was later demonstrated by u- orescence microscopy that these appendages are cytochrome-containing extensions of the cell's outer membrane and periplasmic space [24]. These structures most often appear to have a vesicle chain morphology [24, 25]. They have recently been examined in near-native-state conditions by electron cryotomography, where the distribution of electron densities matching the size and shape of putative outer membrane-associated cytochromes MtrA and MtrC were mapped, giving an esti- mate of cytochrome distribution on the surface [25]. However, among other ques- tions (such as the reasons behind their formation), it is not yet known if these 11 structures can actually full their proposed function as bacterial nanowires in liv- ing cells, and if so, by what mechanism this electron transport occurs. 1.1.5 Outward, inward, and cell-to-cell EET in diverse microorganisms While much of our knowledge of multiheme cytochrome (MHC)-enabled EET stems from Shewanella and Geobacter, particularly S. oneidensis and G. sulfurre- ducens [3, 13, 26], genomic and metagenomic reports are continuously emerging, documenting a diverse assortment of bacteria [27{29] and archaea [30{32] with putative MHCs potentially involved in EET|some containing as many as 50 heme- binding motifs in a single gene [32]! Meanwhile, new isolates present opportunities for expanding our knowledge, as seen in Ardenticatena maritima. This bacterium forms bundled structures of multicellular laments where cells connected end-to- end are coated with putative MHCs and iron minerals, together proposed to facil- itate both outward and inward EET [33]. In addition to outward EET, it turns out that S. oneidensis and G. sulfurre- ducens are both capable of performing inward EET, taking in electrons from electrodes [34, 35]. Other organisms thought to be capable of bidirectional EET include the environmental isolates Alcaligenes faecalis and Ardenticatena maritima [33, 36, 37]. Though the mechanisms of cathodic microbial electron uptake are poorly understood compared to anodic systems, current models include: (1) medi- ated electron transfer via small molecules, such as H 2 [38]; (2) through electrode- bound mediator-generating enzymes, such as hydrogenases [39]; or (3) through direct contact via redox proteins, including MHCs, that relay electrons into the cell [39]. In recent years, more organisms are being found capable of inward EET, including (i) methanogens and acetogens [40, 41], (ii) iron-oxidizing bacteria such 12 as Acidithiobacillus ferrooxidans [42], Mariprofundus ferrooxydans [43], Rhodopseu- domonas palustris [44] and Sideroxydans lithotrophicus [45], and (iii) sulfate reduc- ing bacteria such as Desulfopila corrodens [46] and Desulfovibrio ferrophilus [47]. The phenomenon of direct interspecies electron transfer (DIET) was discov- ered in laboratory-evolved aggregates of Geobacter metallireducens and G. sul- furreducens, where a MHC was found essential to syntrophic electron transfer that coupled ethanol oxidation in one bacterium to fumarate reduction in the other [48]. DIET has generated widespread interest since its discovery, leading to increasing observations of how microbial conductive components, such as conductive pili and MHCs, facilitate EET from one syntrophic partner to another to achieve metabolic processes that may otherwise be thermodynamically unfavorable [49, 50]. Since then, conductive pili and/or MHCs have been implicated in direct electron transfer performed by other dual-species partnerships, such as G. metallireducens with the methanogens Methanosaeta harundinacea [51] and Methanosarcina barkeri [52], as well as G. sulfurreducens with the green sulfur bacterium Prosthecochloris aestau- rii in a process dubbed syntrophic anaerobic photosynthesis [53]. Going forward, microbes capable of DIET may emerge as important players for a wide range of bio- electrochemical systems, where complex communities catalyze anodic or cathodic processes for renewable energy recovery. One of the most important metabolic interactions in nature occur within syntrophic consortia of anaerobic methanotrophic archaea (ANME) and sulfate- reducing bacteria (SRB), various lineages of which are thought to use DIET to perform the environmentally signicant anaerobic oxidation of methane (AOM) in ocean sediments [54, 55]. Use of heme-reactive staining has painted a picture of a heme-rich intercellular region that bridges adjacent cells in both mesophilic and thermophilic AOM consortia [56, 57]. MHCs (especially those with 4{12 hemes) 13 are overexpressed in both ANME and SRB partners during syntrophic activity in diverse AOM consortia, with some partners also producing numerous pili [56, 58]. Meanwhile, metagenomic and genomic evidence points to the abundance of MHCs in microbes potentially capable of DIET, including archaeal [59] and bacterial [60, 61] members of AOM consortia. A mechanism for DIET by ANME-SRB syn- trophs has been proposed based on several of these recent works [57, 60], but the exact mechanisms of electron transport between these syntrophic organisms are still being investigated. A basic understanding of DIET in this system is criti- cal, given the environmental signicance of AOM for mitigating greenhouse gas emission from the ocean to the atmosphere. 1.2 Background 2: Methods This section gives a basic description of several types of microscopy-related meth- ods used later in this thesis. Topics were chosen and described based on relevance to this work, familiarity to others, and the level of detail included in the following chapters. 1.2.1 Perfusion ow imaging platform For some experiments presented in Chapter 2, I used a perfusion ow imaging platform for in vivo uorescence microscopy. This imaging setup was rst used in the El-Naggar Lab to observe bacterial nanowires in S. oneidensis in 2014, as reported and described in detail in [24]. Afterwards, a close-up picture and schematic of the same setup was included in [25]. For reference, the schematic from [25] is shown in Figure 1.5. 14 Figure 1.5: Schematic of perfusion ow imaging platform. Cells are injected into the chamber, and attached to a glass cover slip (or EM grid) which is vacuum sealed to the chamber. Cells are sustained with a continuous ow of medium, membrane stained with FM 4-64FX, and monitored with uorescence microscopy. Figure from [25]. This platform provides a continuous laminar ow of medium, which basically means the liquid medium moves in \layers" above the cells on the coverslip. Any membrane extensions produced by the cells are subsequently kept in the same focal plane as the layer the cells are in, pushed in the same direction the medium is owing. This makes it easy to image both cells and membrane appendages in focus over time and allows for precise quantication of their production during an experiment. Incidentally, this laminar ow also keeps the layer of medium closest to the cells on the coverslip in a general state of oxygen limitation; we previously 15 calculated that even if an aerobically prepared medium is supplied, cells in the chamber quickly use up the majority of oxygen that is accessible to the cells in the surrounding layer(s) of medium [24], though it is unclear to what extent this oxygen limitation plays a role in nanowire formation. In Chapter 2, I use this imaging platform and other non- ow imaging setups in combination with assorted lipid membrane dyes to visualize cells and membrane extensions. Combined with other features of the Nikon Ti-E inverted uorescence microscope, I can (1) keep objects of interest in focus even if the sample \drifts" very slightly over time (drift correction feature), (2) image the same objects of interest repeatedly in order to monitor their behavior over time (time-lapse capability), and (3) do this for multiple areas (i.e., elds of view) of the sample/coverslip to increase the number of cells monitored in a single experiment (multi-point feature). Altogether, our perfusion ow imaging setup allows us to monitor cells and record their production of membrane extensions over time. 1.2.2 Transmission electron microscopy Brie y, transmission electron microscopy (TEM) works by pointing an electron beam towards a sample in its path, and collecting what passes through the sample. In brighteld TEM, where there is more matter, more electrons will scatter, and the image collected will appear darker; where there is less matter, more electrons will pass through the sample, and the image collected will appear brighter. The higher kV the instrument, the stronger the electron beam; a 200 kV instrument may yield sharper images, but be more destructive towards the sample, which is why a 80 kV instrument is sometimes preferable for fragile biological samples. Samples are placed on a small metal mesh grid, often supporting a very thin lm (e.g. thin carbon lm) which in turn supports the sample; here, the metal grid 16 provides support (dense matter that appears dark in nal image) and the sample is visible where it is suspended only by the thin carbon lm (less matter, appears lighter on nal image). Since samples must be dry, TEM doesn't allow for live cell imaging; in Chapter 2, I chemically xed samples (cells and membrane extensions adhered on a grid, cultivated immediately prior by perfusion ow imaging), stained, and air-dried the sample prior to imaging. Prior to drying the sample, samples are often treated with negative stains containing heavy metals (e.g. uranyl acetate, osmium tetroxide) to enhance contrast of samples of interest, so my sample prep in Chapter 2 included these contrasting agents as well as a redox-reactive staining procedure relevant to this study. While it's possible to embed samples in resin and slice them very thinly (80 micron thickness) e.g. to observe cross-sections of cells, in Chapter 2 I focused on whole cell TEM, since the objects of interest (bacterial appendages e.g. membrane extensions, pili, and agella) were all thin enough (4 to200 nm in diameter) to be clearly visible by TEM without sectioning. 1.2.3 Atomic force microscopy In Chapter 2, I used atomic force microscopy (AFM), a type of scanning probe microscopy. Tapping-mode AFM is one of the most basic types of AFM, where a very small oscillating probe called a cantilever (basically, a mechanical pointer) slowly traces the topography of a dried at sample (e.g. cells chemically xed to a surface and air-dried), tapping the surface point by point and line by line until a topographic image is obtained for the area that was traced. The cantilever oscillates near its resonance frequency (specic to the type of probe) with a drive amplitude (i.e., a preset/default setting) of up to 100 nm [62]. When it reaches an object of dierent height, its actual amplitude of oscillation 17 changes, and this is recorded. Bacterial appendages (e.g. outer membane exten- sions, agella, pili) are usually resolved clearly, since they are relatively thin (4 to200 nm in diameter) and similar in height relative to the at surface the sam- ple is on top of (e.g., a glass coverslip for microscopy). The probe sometimes has diculty transitioning between tracing the relatively at surface to tracing the surface of500 nm diameter bacterial cells (a relatively dramatic shift in sample height). The scan speed can be optimized to allow for more clear images, where a slower scan speed gives the probe more time to adjust for (and clearly resolve) samples of variable height. In Chapter 2, I showed some phase contrast images, which are a type of image obtained by tapping mode AFM; to produce these images, the instrument software compares the phase of the drive amplitude oscillation (i.e., the default oscillation of the probe) and the feedback/recorded oscillation (i.e., the actual oscillation of the probe which may change as it encounters objects on the surface), allowing the user to clearly distinguish objects on a surface. 1.2.4 Total internal re ection uorescence microscopy Total internal re ection uorescence (TIRF) microscopy is a type of uorescence microscopy that selectively illuminates a certain focal plane (100 nm), usually near the surface of a microscopy coverslip, while minimizing background signal from other planes outside the focal plane of interest. Basically, it directs light toward the sample at a certain angle such that most light is re ected or bent away from the sample rather than illuminating many layers of sample. It takes advantage of the dierence in refractive indices of dierent materials (e.g. glass in coverslip vs water in liquid sample), since the light will bend when it encounters a material with dierent refractive index. Further details about how TIRF works can 18 be found in [63]. A basic illustration of TIRF used in the context of single-particle tracking (described below), is depicted in Figure 1.6. 1.2.5 Single-particle tracking and basic types of diusion Single-particle tracking (SPT) traces the motion of a particle of interest over time. SPT is often combined with single-molecule labeling and image acquisition by TIRF microscopy, described brie y above. Fluorescent labels such as quantum dots are ideal for SPT due to their high signal-to-noise ratio and photostability [64, 65]. Time-lapse imaging is rst performed with sucient resolution and low enough concentration of uorescent label such that individual molecules can be detected (single molecule imaging). Afterwards, the center of each dot of signal (e.g. from a uorescent molecule) is found for each dot in each frame in the time-lapse series. Then, the localized positions of these molecules are connected frame-by-frame to build a trajectory. Mean squared displacement (MSD) analysis can be performed on single molecule trajectories to classify the trajectory based on the shape of the MSD curve as one of several major types of diusion, such as active, conned, or free (Brownian) diusion. The process of SPT, from image acquisition to basic diusion analyses, is illus- trated below in Figure 1.6. I used this work ow in Chapter 4 to extract the diusive dynamics of multiheme cytochromes on the surface of S. oneidensis cells and membrane extensions. Further detail about diusion analyses, and how they can be used to describe and quantify a molecule's mobility (i.e., diusion), will be described in Chapter 4. 19 Figure 1.6: Diagram of single-particle tracking (SPT) process and basic types of diusion. SPT begins with image acquisition by time-lapse microscopy, where the same 2-D area of the sample is imaged repeatedly over time (1 image frame captured per time increment). Image (top left) is a snapshot of S. oneidensis and its membrane extensions (green). Red circles illustrate molecules of interest (e.g. quantum dots, labeling proteins of interest) which can be tracked on the surface of a bacterial cell or its membrane extensions (green). Image (bottom left) shows total internal re ection uorescence (TIRF) microscopy, a method of image acquisition for SPT. (Top right) Next, a particle localization method is used to nd the center of each dot representing the uorescent signal from a molecule of interest. This is repeated for all frames in the time-lapse series. The localized particle positions are then connected frame-by-frame to build trajectories. (Bottom right) The mean squared displacement (MSD) of a single particle trajectory plotted as a function of time can be used to classify a trajectory into one of 3 major types of diusion, based on the shape of its MSD curve. Figure images adapted from [24, 64, 66, 67]. Figure courtesy of Moh El-Naggar. 20 1.3 Overview of chapters The main goal of this thesis is to investigate how biological electron transport works in Shewanella oneidensis MR-1, a bacterium that can breathe rocks, conduct electricity, and form bacterial nanowires. The work presented in this thesis is divided into several chapters: This chapter contained a brief introduction to various topics relevant to the system of study, as well as a basic description of several techniques used in Chapters 2, 3, and 4. Each of the research chapters themselves will also include a more focused and detailed introduction that is relevant to the topics presented within that chapter. In Chapter 2, I used several microscopy methods to further characterize the outer membrane extensions (OMEs) also known as bacterial nanowires in S. onei- densis MR-1. In the rst half of Chapter 2, I set out to determine the conditions underlying their formation. While it was previously thought that these structures form in response to oxygen limitation, it was unclear to what extent oxygen limita- tion plays a role in their formation. Similarly, it was not yet known whether other physical or chemical conditions contribute to their formation. To investigate these, I designed in vivo uorescence microscopy experiments allowing me to examine the specic role of oxygen limitation and other physical conditions which might in u- ence the production of membrane extensions in S. oneidensis MR-1. I found that cell-to-surface contact is sucient to trigger the formation of S. oneidensis OMEs under a wide range of conditions. As a bonus, I also quantied the frequency of cells producing OMEs in perfusion ow culture conditions; I also observed the longest OME recorded in S. oneidensis to date, and estimated its elongation rate. In the second half of chapter 2, I assessed the extent of cytochrome-dependent redox activity in S. oneidensis membrane extensions. I combined heme-dependent 21 staining with transmission electron microscopy to compare OMEs in wild type and cytochrome-decient strains. In doing so, I also probed 3 types of extracellular laments (OMEs, agella, and pili) for these EET components. I found that periplasmic and outer membrane cytochromes are responsible for most of the redox activity detected using this assay, and demonstrated that these components are limited to OMEs and do not associate with agella or pili. As a bonus, the methods used in this section also allowed me to dene a maximum center-to-center distance between redox partners on the surface of membrane extensions. Altogether, the works presented in Chapter 2 provide a greater understanding of how bacterial nanowires work, what they look like, and why they form. Chapters 3 and 4 are two parts of a combined project. The overarching goal of these chapters was to investigate how long-distance (micrometer-scale) electron transport works along membrane surfaces in S. oneidensis. Based on previous studies that will be elaborated in the introduction of Chapter 4, we had recently hypothesized that long-distance conduction arises from a combination of electron hopping and cytochrome diusion, which allows collisions and electron exchange between cytochromes along membranes. However, the diusive dynamics of multi- heme cytochromes had never been observed or quantied in vivo, making it dicult to assess their hypothesized contribution to the collision-exchange mechanism. So, in Chapter 3, I established a biotin-streptavidin labeling scheme in our system that would ultimately allow the visualization and tracking of individual membrane cytochromes. In this labeling scheme, proteins tagged with a biotin acceptor pep- tide (AP) can be biotinylated in vivo with biotin ligase, and subsequently detected by streptavidin conjugates (e.g. uorophore or quantum dots in microscopy, or horseradish peroxidase for western blot). First, I used an assortment of molecular biology techniques to fuse the AP tag to the C-termini of cell surface cytochromes 22 MtrC and OmcA. I placed plasmids encoding tagged MtrC or tagged OmcA into mutant strains lacking MtrC or OmcA, respectively. In between each major step in the process of engineering these strains, I also performed various quality control experiments to check that everything works before proceeding. Then, perhaps most importantly, I tested the labeling scheme itself. I performed systematic labeling controls in western blot and microscopy where the key components of the labeling scheme are systematically omitted; with these, I demonstrated the success and specicity of this labeling scheme. Chapter 3 describes supporting work that was necessary to establish the label- ing scheme in our system before it could be used in Chapter 4. In Chapter 4, I progress to our main goal of investigating how long-distance electron transport works in S. oneidensis, addressing several related questions, such as (1) can I suc- cessfully label and visualize individual bacterial cell surface cytochromes in living cells? (2) Are they mobile? (3) If so, can I measure this mobility? And (4) How do these dynamics impact overall electron transport? Using a combination of tar- geted quantum dot labeling and single molecule imaging, I visualized individual cytochromes MtrC and OmcA (tagged in Chapter 3), two key components of the S. oneidensis EET network, and found that they are indeed mobile on the surface of living cells and membrane extensions. I then used single particle tracking and diusion analyses to describe and quantify this mobility. I found that these dynam- ics can trace a path for electron transport via overlap of cytochrome trajectories along cells and membrane extensions, consistent with our proposed long-distance conduction mechanism. We also used my experimental measurements to inform kinetic Monte Carlo simulations of electron transport. Overall, in this chapter, I described the rst dynamics measurements of bacterial cell surface cytochromes, 23 and found that this mobility supports our hypothesis of how long-distance electron transport works along membrane surfaces in S. oneidensis. 24 Chapter 2 Surface-induced formation and redox-dependent staining of outer membrane extensions in Shewanella oneidensis MR-1 This chapter has been adapted from [68]: Grace W. Chong, Sahand Pirbadian, and Mohamed Y. El-Naggar. Surface- Induced Formation and Redox-Dependent Staining of Outer Membrane Exten- sions in Shewanella oneidensis MR-1. Frontiers in Energy Research, 7:1{9, Aug 2019. ISSN 2296-598X. doi: 10.3389/fenrg.2019.00087. URL https://www.frontiersin.org/article/10.3389/fenrg.2019.00087/full. 25 2.1 Introduction Shewanella oneidensis MR-1 is a Gram-negative, facultative anaerobic het- erotrophic bacterium with versatile respiratory capabilities: in its quest for energy, it can utilize an array of soluble and insoluble electron acceptors, from oxygen to extracellular solid surfaces such as minerals and electrodes. This ability to couple intracellular reactions to the respiration of external surfaces, known as extracel- lular electron transfer (EET), allows microbial catalytic activity to be harnessed on the electrodes of bioelectrochemical technologies ranging from microbial fuel cells to microbial electrosynthesis [4, 5]. As an extensively studied model organ- ism for EET, studies of S. oneidensis revealed the critical role of periplasmic and outer membrane multiheme cytochromes in forming extracellular electron conduits that bridge the cell envelope [1, 11, 12, 16, 69, 70]. Specically, the periplasmic decaheme cytochrome MtrA connects through the MtrB porin to the outer mem- brane decaheme cytochrome MtrC that, along with another decaheme cytochrome OmcA, function as the terminal reductases of external electron acceptors or soluble electron shuttles [16]. In addition to this well-established role in directing electron transfer across the cell envelope, the Mtr/Omc components have been recently shown to facilitate long-distance electron transport across the membranes of mul- tiple cells via a redox conduction mechanism thought to arise from a combination of multistep hopping along cytochrome heme chains and cytochrome-cytochrome interactions [21]. S. oneidensis also forms extensions of the outer membrane and periplasm that include the Mtr/Omc multiheme cytochromes responsible for EET [23{25]. These outer membrane extensions (OMEs) are proposed to function as bacterial nanowires that also facilitate long-distance EET through redox conduction. How- ever, in contrast to electrode-spanning cells measured by electrochemical gating 26 [21], the cytochrome-dependent conductivity of these proposed bacterial nanowires has only been directly assessed under dry, chemically xed conditions [23, 71]. A full understanding of the role of S. oneidensis OMEs will therefore require chal- lenging in vivo measurements of their specic impact on extracellular respiration and observations of the membrane protein dynamics that allow inter-cytochrome electron exchange and redox conduction [72]. Beyond the detailed mechanism of electron transport along these structures, additional questions remained regarding the physical and environmental condi- tions that trigger their formation. The S. oneidensis OMEs can extend to several times the cell length, and have been observed with a range of morphologies from chains of interconnected outer membrane vesicles to membrane tubes [24]. Since early reports suggested that they form in response to electron acceptor limita- tion, particularly oxygen limitation [22], subsequent studies involving these OMEs have been performed in oxygen limiting conditions [23{25, 73]. However, while the increased expression and production of multiheme cytochromes under oxygen limiting and anaerobic conditions is well established [24, 73, 74], it was not clear if oxygen limitation is the sole contributor to the membrane extension phenotype in S. oneidensis. In fact, a recent gene expression study hinted at independent reg- ulatory mechanisms for extending the membrane and localizing the EET proteins [73]. Furthermore, membrane extensions have been reported in multiple organisms under a variety of growth conditions [75{81], including those in the form of vesicle chains [25, 82{85], as is the case for S. oneidensis. It was previously shown that S. oneidensis membrane vesicles, which form the basis of OMEs, are redox-active, and that this activity likely stems from the cytochromes present on the puried vesicles [86]. The native-state characterization 27 of cytochromes on the OMEs themselves was so far limited to microscopic obser- vations ranging from immuno uorescence [24] to electron cryotomography [25], rather than mapping the activity of the redox centers. The possible association of redox-active components with other extracellular laments in Shewanella, beyond OMEs, also remained largely unexplored. Recent studies in both bacteria and archaea, however, have demonstrated that a combination of histochemical heme- reactive staining and electron microscopy can be used to visualize redox-dependent activity of cytochromes that enable functions ranging from mineral oxidation to interspecies electron transfer within methanotrophic consortia [57, 87]. This study set out to address some of these outstanding questions regarding S. oneidensis OMEs. To determine the conditions underlying OME formation, I designed in vivo uorescence microscopy experiments allowing me to examine the specic role of oxygen limitation and other physical conditions which might in u- ence OME production in S. oneidensis MR-1. I found that cell-to-surface contact is sucient to trigger the formation of S. oneidensis OMEs under a wide range of conditions. To assess the extent of cytochrome-dependent redox activity in these structures, I implemented heme-dependent staining with transmission elec- tron microscopy to compare OMEs in wild type and cytochrome-decient strains. In doing so, I also probed 3 types of extracellular laments (OMEs, agella, and pili) for these EET components. I found that periplasmic and outer membrane cytochromes are responsible for most of the redox activity detected using this assay, and that these components are limited to OMEs and do not associate with agella or pili. 28 2.2 Materials and Methods 2.2.1 Cell cultivation For experiments probing the conditions of OME formation with uorescence microscopy, Shewanella oneidensis MR-1 cells were grown aerobically from frozen (-80°C) stock in 50 mL LB broth overnight at 30°C and 150 rpm up to late log- arithmic phase (OD600 2.4{2.8). From this overnight culture, 5 mL of cells were collected by centrifugation at 4,226 × g for 5 min and washed twice in sterile dened medium [24]. Cells were then introduced into a perfusion ow imaging platform described previously [24] or the coverslip-bottom glass reactor described below after appropriate dilution to achieve a desirable cell density on the surface for uorescence time-lapse imaging. Heme staining and transmission electron microscopy were performed on anaer- obic cultures of S. oneidensis MR-1 and JG1486 (Mtr/mtrB/mtrE) [88]. First, 5 mL of an aerobic overnight LB pre-culture was pelleted by centrifugation, washed in dened medium [24], and used to inoculate 100 mL of anoxic dened medium in sealed serum bottles with 30 mM fumarate as the sole electron accep- tor. After 24 h at 30°C and 150 rpm, at OD600 0.28, this anaerobic culture was harvested by centrifugation at 7,142 × g for 10 min, washed by centrifugation (4,226× g for 5 min), and resuspended in dened medium for a total volume of 10 mL. Cells were then injected into the perfusion ow imaging platform containing an electron microscopy grid, as described previously [25]. 2.2.2 Fluorescence microscopy Lipid membrane stains FM 4-64FX (Life Technologies; 0.25 g/mL), FM 1-43FX (Life Technologies; 0.25 g/mL), or TMA-DPH (Cayman Chemical Company; 10 29 M) were used to visualize cells and OMEs on an inverted uorescence microscope (Nikon Eclipse Ti-E) using the TRITC, FITC, or DAPI channels (Nikon lter sets G-2E/C, B-2E/C, and UV-2E/C) with 500, 500, and 100 ms exposure times, respectively. FM 4-64FX was generally used as the membrane stain, except in experiments with no ow or agitation, as this concentration of dye faded more quickly over time in unmixed solutions. Two experimental platforms were used for uorescence imaging experiments: a perfusion ow setup used previously (Figure 1.5) [24, 25] or a coverslip-bottom glass reactor constructed to allow gas injection and measurement of dissolved oxygen levels while visualizing cells (Figure 2.2A). The reactor consisted of a clean glass tube (thickness 1.5 mm, interior diameter 24.7 mm, and length 50 mm) glued on to a clean 43 × 50 mm no. 1 thickness glass coverslip (Thermo Scientic) using waterproof silicone glue (General Electric). The autoclaved reactor was placed on the inverted microscope, and a peristaltic pump (Cole-Parmer Master ex L/S Easy-Load II) was used to control injection of ltered air at a rate of 3.6 mL/min into the reactor. The air inlet (22G 3" sterile needle) was placed 1{2 mm from the coverslip bottom of the reactor so as to ensure oxygen availability and good mixing near the focal plane. Time-lapse imaging was started immediately following introduction of 10 mL of the cell-media mixture into the reactor and continued for 2 h with images acquired in 5 min increments. Oxygen levels in the reactor were measured by a dissolved oxygen probe (Milwaukee Instruments MW600) at various levels (e.g., 1 mm from bottom, middle, and 1 mm from top) over time after cells were added. To check whether the planktonic cells also displayed OMEs, imaging was stopped after the surface-attached cells produced OMEs, and 400 L of the planktonic mixture (obtained within 1{2 mm from the top solution-air interface) 30 was gently pipetted to a new clean coverslip, and immediately imaged for another 2 h. 2.2.3 Heme staining and transmission electron microscopy All heme staining experiments were performed on cells attached to electron microscopy grids recovered from the perfusion ow imaging platform after conr- mation of OME production using uorescence microscopy, as described previously [25]. To accomplish this, an X-thick holey carbon-coated, R2/2, 200 mesh Au NH2 London nder Quantifoil EM grid was glued to the glass coverslip, with the carbon lm-coated side facing away from the glass, before sealing the perfusion chamber. The chamber was lled with ow medium, then 400{600 L of washed cells were injected for a surface density of 50{150 cells visible per 74 × 74 m square in the 200 mesh grid. Cells were allowed to settle for 5{15 min on the grid before resuming perfusion ow at a volumetric ow rate of 6.1 ± 0.5L/s. Imaging con- tinued for about 3.5 h in 5 min increments before medium ow was stopped and the chamber opened under sterile medium. The EM grid was then removed, chem- ically xed, and prepared for electron microscopy visualization of heme iron, using a staining protocol adapted from [57]. First, the sample was xed for 30 min in 2.5% glutaraldehyde (dissolved in 25 mM HEPES, pH 7.4, 17.5 g/L NaCl), washed 5 times by soaking 1 min each in buer (50 mM HEPES, pH 7.4, 35 g/L NaCl), then incubated for 1 h or 2.5 h with the heme-reactive stain 3,3'-diaminobenzidine (DAB; 0.0015 g/mL, dissolved in 50 mM Tris HCl, pH 8) with or without 0.02% hydrogen peroxide (H 2 O 2 ). After 5 washes (100 mM HEPES, pH 7.8), the sam- ple was stained for 1 h in 1% osmium tetroxide, and washed again 5 times. The sample was negative stained in 1% uranyl acetate or 1% phosphotungstic acid for 2 min and air dried overnight. Dried samples were stored in a desiccator before 31 transmission electron microscopy (TEM) imaging. TEM images were acquired on a JEOL JEM-2100F instrument operated at 200 kV, a FEI Morgagni 268 instrument operated at 80 kV, or a FEI Talos F200C instrument operated at 200 kV. To determine and quantify the extent of cytochrome-reactive staining after treatment with DAB, ImageJ was used to measure the mean pixel intensity (arbi- trary gray value units re ecting electron transmission) across an area in the inte- rior of an extension (A), or an area in the background (B). For each image, a background threshold value (C ) was generated by taking the mean background intensity (B) and subtracting its standard deviation (D); thus, C = B{D. If the mean intensity of an extension (A) was lower than this threshold (C ), then it was categorized as stained. For each condition (wild type, mutant, and chemical control), the percentage of stained OMEs (E) was calculated. To calculate the staining intensity of a single OME (F), the mean pixel intensity of the extension (A) was subtracted from that of the background (B), giving F = B{A. A value of F was calculated for each of the OMEs assessed in each replicate experiment for each condition (wild type, mutant, and chemical control). For each condition, the mean of all F values was calculated, giving G WT , G mutant , and G control . Then, G WT and G mutant were corrected by subtracting G control , where G WT {G control = H WT , and G mutant {G control = H mutant . These values H WT and H mutant represent mean staining intensities of all the OMEs in each strain, corrected for the contri- bution of negative staining (G control ). To calculate the fold dierence in staining frequency between wild type and the mutant, the percentage of OMEs stained in the wild type (E WT ) was divided by that of the mutant (E mutant ). To calculate the fold dierence in staining intensity between wild type and the mutant, the mean staining intensity of the wild type (H WT ) was divided by that of the mutant (H mutant ). 32 2.3 Results and Discussion 2.3.1 Surface contact is sucient to induce production of outer membrane extensions by Shewanella oneiden- sis MR-1 Production of OMEs by a majority of S. oneidensis cells was observed in the oxygen limiting perfusion ow platform, as previously described [24, 25] (Figures 1.5 and 2.1), but also in near-saturating oxygen conditions (6.5{7.5 ppm = 400{470 mM O 2 ) provided by a glass-bottomed reactor that allowed air injection during in vivo microscopy (Figure 2.2A). Figure 2.1: Outer membrane extensions are commonly formed by surface- attached perfusion culture cells. (A) Time-lapse uorescence microscopy snapshot of outer membrane extensions (OMEs, white arrows) produced by S. oneidensis MR-1 at a single timepoint in a 3.5-h perfusion ow imaging experi- ment. Cells and OMEs are visualized with the red membrane stain FM 4-64FX. (B) Statistics of OME production from over 5,400 cells in 4 replicate 3.5-h perfu- sion culture experiments illustrates that a majority (78%) of cells produce OMEs visible over time. The remaining cells were seen with only outer membrane vesicles (OMVs), or nothing at all. Error bars show mean ± SEM (Scale bar: 10 m). 33 Though it can take up to several hours for a majority of surface-attached cells to produce OMEs, I observed production of OMEs as early as 10 min after cells contacted the surface of a glass coverslip (Figures 2.2 and 2.3). To further exam- ine the role of surface contact, planktonic cells from the bulk oxygenated reactor were sampled 2 h after the reactor was inoculated (1.5 h after OMEs started being produced by surface-attached cells) and transferred to clean coverslips for observation. These previously planktonic cells showed no evidence of OMEs at the Figure 2.2: Surface attachment is sucient to induce production of outer membrane extensions. (A) Diagram illustrates experimental procedure. (B,C) Microscopy images of S. oneidensis MR-1 cells and membrane extensions (white arrows) labeled with the red membrane stain FM 4-64FX. Time (t = 0 min) indicates estimated time of cells contacting the glass surface. (B) Demon- strates production of outer membrane extensions (OMEs) by surface-attached cells in the aerated glass-bottomed reactor. Here, 6.5{7.5 ppm = 400{470 mM O 2 . (C) Demonstrates OME production by planktonic cells from the reactor which were transferred to a new coverslip surface after events in (B) were conrmed (Scale bars: 5 m). 34 time of sampling, but then also went on to begin to display OMEs within 35 min after contacting the surface (Figure 2.2). These observations were not limited to the dened minimal medium used, a particular surface chemistry, or mixing con- ditions; post-attachment OME production was also observed in rich (LB) medium or in buer (PBS), on dierent surfaces (glass coverslips and carbon-coated elec- tron microscopy grids), and regardless of liquid ow or agitation (Figures 2.3 and 2.4). To ensure that the used cell density did not result in O 2 -limiting conditions selectively at the surface, I also experimented with sparse coverage, down to 5{20 cells per eld of view (112× 112m) in a well-mixed and oxygenated reactor, and conrmed that these cells also produced OMEs (Figure 2.4C). Taken collectively, these observations of OME production by surface-attached cells, but not by planktonic cells until subsequent attachment, and regardless of medium composition, surface type, and oxygen availability, point to surface con- tact as the primary determinant of OME production by S. oneidensis. Previous studies on the role of cytochrome-functionalized OMEs as bacterial nanowires pri- marily focused on the formation of these structures under O 2 -limited conditions [22{25]. Previous works indicated that O 2 limitation is necessary for enhanced production of the multiheme cytochromes required for EET [73, 74, 89], and that oxygen availability also impacts S. oneidensis performance and current production in bioelectrochemical systems [89, 90]. However, my observations suggest that the membrane extension phenotype is predominantly controlled by surface attachment. My ndings are consistent with a previous proposal based on transcriptome and mutant analyses [73] that independent pathways are responsible for producing the EET components and extending the outer membrane, while implicating surface contact in controlling the latter pathway. 35 Figure 2.3: Outer membrane extensions are produced quickly by plank- tonic cells in rich aerobic medium soon after cell-to-surface contact. Diagram illustrates experimental procedure. Microscopy images depict S. onei- densis MR-1 cells and outer membrane extensions (OMEs, white arrows) labeled with the red membrane stain FM 4-64FX. Time (t = 0 min) indicates estimated time of cells contacting the glass surface. (Scale bars: 5 m.) While my observations show that surface attachment is sucient to induce OMEs, it is important to note that they do not rule out the in uence of O 2 limi- tation on the frequency of OME production. In perfusion ow imaging, I was able to precisely dene the percentage of OME-producing cells: observation of 5,400 cells over four replicate experiments revealed that 78% of surface-attached cells produced OMEs during 3.5 h of perfusion culture (Figure 2.1). This precise quan- tication is possible in perfusion ow imaging because the laminar ow helps to restrict the structures to the focal plane near the surface. However, this laminar ow establishes O 2 limitation as a result of cellular O 2 consumption and the no-slip 36 Figure 2.4: Outer membrane extensions are produced in a variety of surface-attached conditions, regardless of medium composition, surface chemistry, agitation, or aeration. S. oneidensis MR-1 cells and membrane extensions are visualized by membrane stains FM 4-64FX (red), FM 1-43FX (green), or TMA-DPH (blue). Unless otherwise specied, cells were imaged at the surface of glass coverslips with ow or agitation of oxygen-limited minimal medium. Outer membrane extensions are observed in (A) oxygen-limiting per- fusion conditions, described previously [24, 25], (B) oxygen-abundant, high cell density conditions, (C) oxygen-abundant, low cell density conditions, (D) in rich (LB) medium, (E) in buer (PBS), (F) on a carbon-coated electron microscopy grid, and (G) without ow or agitation. (Scale bars: 5 m.) condition at the surface-solution interface [24]. Thus, I could precisely determine the frequency of OME production only in O 2 -limiting perfusion conditions, but not in oxygenated well-mixed reactors where the structures could uctuate in and out of the focal plane. It was previously shown that phage-induced cell lysis (particularly the release of extracellular DNA) promotes biolm development in S. oneidensis [91], and that explosive cell lysis in actively growing Pseudomonas aeruginosa biolms is accompanied by the release of extracellular DNA and the production of vesicles containing cytoplasmic content [92]. However, lysis does not appear to play a role in the large fraction (78%) of S. oneidensis cells seen producing OMEs under my experimental conditions. My in vivo uorescence observations reported here 37 (Figures 2.1, 2.2, 2.6, and Movie S1 in [68]), as well as our previous correlative uorescence and cryo-electron microscopy observations [25], capture the exten- sion process from originating intact cells with no evidence for loss of cell integrity and eliminate the possibility that the OMEs are lysis products that subsequently attach to nearby cells. Moreover, we previously reported GFP-labeling obser- vations where only periplasmic GFP, but not cytoplasmic GFP, localized to the extensions, demonstrating an intact cytoplasmic membrane during the OM exten- sion process [24]. However, while these observations demonstrate that OMEs are not simply a consequence of cell lysis, they do not rule out that OM extension may be partially associated with an instability or change in membrane integrity. Membrane extensions, including those formed as chains of membrane vesicles (MVs), are not limited to S. oneidensis [25, 75{85]. The nding that surface con- tact plays an important role is consistent with prior observations of vesicle chains and OMEs produced by surface-attached cells of other bacteria, including She- wanella vesiculosa [82], Bacillus subtilis [85], and biolms of Myxococcus xanthus [83]. In addition, another M. xanthus study noted that static, rather than shaken, conditions promote more OME production [84]. However, while MVs are ubiqui- tous features of bacteria, the mechanisms behind MV and MV chain formation are still widely unknown [93, 94], but they do not appear to simply result from loss of membrane integrity [95]. The importance of surface-attached, biolm, or static conditions may point to a generalized mechanism where MVs are successively pro- duced and merged into long extensions rather than shed away under more dynamic (e.g., free-swimming or shaken culture) conditions. Once formed, these extensions may then enable a variety of functions ranging from facilitating cell-cell interac- tions [76, 83, 85] to the long-distance EET role proposed for S. oneidensis OMEs [22, 23]. 38 It was also previously proposed that MVs and OMEs can increase the likelihood of encountering neighboring cells and external redox-active surfaces by virtue of the signicant change in surface area-to-volume ratio that these structures present [24]. Consistent with this proposal, I occasionally captured multiple extensions from single cells (Figure 2.5) as well as in vivo uorescent observations of remarkably long OMEs, likely the longest observed to date. Figure 2.6 and Movie S1 in [68] capture a cell producing a >100 m OME at a rate over 40 m/h, at the same time that the cell surface area appeared to shrink by an amount consistent with the newly displayed OME. 39 Figure 2.5: Cells can produce multiple outer membrane extensions per cell. (A-B) Transmission electron microscopy images of chemically xed, nega- tively stained S. oneidensis MR-1 cells and membrane extensions (Scale bars: 200 nm.) (C-D) Atomic force microscopy tapping mode phase images of chemically xed cells and membrane extensions. (Scale bars: 5 m.) 40 Figure 2.6: Outer membrane extensions can reach a length of >100 m, produced at a rate >40 m/h. Image sequence from time-lapse uo- rescence microscopy of surface-attached perfusion cultured S. oneidensis MR-1 Mtr/mtrB/mtrE cells. Cells and outer membrane extensions (OMEs) were visualized with the red membrane stain FM 4-64FX. Time (t = 0) marks time since OME production (white arrow). Images depict progression of events, such as (A) cell contacting surface, (B) OME rst visible (white arrow), (C) OME elongates and begins to fold, (D) further OME elongation, (E) OME reaches longest point visible during this experiment (Scale bars: 10 m). 41 2.3.2 Redox-dependent staining of extracellular laments The localization of the multiheme cytochromes responsible for EET to OMEs has been previously demonstrated by immuno uorescence observations of MtrC and OmcA [24], as well as electron cryotomography observations of outer membrane and periplasmic electron densities consistent with cytochrome dimensions [25]. To examine the distribution and activity of the heme iron redox centers along the OMEs, I applied the heme-reactive 3,3'-diaminobenzidine (DAB)-H 2 O 2 staining procedure [57]. Here, the iron centers in heme will catalyze the oxidation of DAB coupled to the reduction of H 2 O 2 . The oxidized DAB forms a localized dark precip- itate that can be observed with the resolution of transmission electron microscopy (TEM). In my TEM experiments, I observed only 3 types of extracellular laments: OMEs, pili, and agella (Figures 2.7, 2.8, and 2.5). As expected, the OMEs clearly stained for heme; with prolonged exposure to DAB (2.5 h DAB staining step) a noticeable <50 nm band of dark precipitate lined the vesicles that compose the entire structure (Figure 2.7). Staining was clearly limited to the OMEs and was absent from the other extracellular laments observed, demonstrating that the cytochromes do not associate with pili and agella (Figure 2.7). The absence of staining in these structures, even when observed in contact with the OMEs (Figure 2.7), also demonstrates that DAB precipitate does not diuse to non-specically stain nearby structures. Meanwhile, the <50 nm thickness of precipitate lining OMEs (i.e., precipitate expansion in the direction perpendicular to the surface of the OME) suggests <50 nm lateral distribution of heme redox centers on OMEs, consistent with the surface distribution of putative cytochromes on OMEs visual- ized by electron cryotomography [25]. However, in general this technique cannot be used to localize specic cytochromes or to dene their exact distribution on the membrane. For this reason, I also used image analysis (Figure 2.8) to compare 42 overall staining displayed by many OMEs rather than focusing on discrete sites on a single structure. Figure 2.7: Redox components are present only on outer membrane extensions, not pili or agella. Histochemical redox-dependent staining with 3,3'-diaminobenzidine (2.5 h staining step) and transmission electron microscopy distinguishes between types of extracellular laments in S. oneidensis MR-1. Images depict dark precipitate (yellow arrows and lines) labeling only outer mem- brane extensions, but not adjacent extracellular structures (A) pili (white arrow), and (B) agella (black arrow). Cells are indicated by asterisk symbols (*) (Scale bars: 200 nm). In addition to chemical controls for staining (e.g., wild type experiments in which H 2 O 2 was omitted), I systematically compared redox-dependent staining in OMEs from wild type S. oneidensis and a mutant lacking genes encoding eight functional periplasmic and outer membrane cytochromes (Mtr/mtrB/mtrE), including the entire Mtr/Omc pathway of decaheme cytochromes [88]. This mutant is unable to perform EET [21, 88, 96] or support long-distance redox conduction across electrodes [21]. I performed two replicate experiments for each of three con- ditions: wild type (Figure 2.8A), mutant (Figures 2.8B,C), and wild type chemical 43 control with no H 2 O 2 (Figure 2.8D), with a total of 45{60 OMEs analyzed per con- dition. Using image processing to compare OME staining to background intensities (see Section 2.2.3 in Materials and Methods), I found that the majority (92%) of wild type OMEs stained for heme, but none stained in the chemical control where H 2 O 2 was omitted (Figure 2.8E). In contrast, a fraction (39%) of OMEs in the mutant strain exhibited heme staining, 2.4-fold less than in wild type (p < 0.0001, Pearson's chi-squared test; Figure 2.8E). While lacking all cytochromes necessary for EET, staining in the mutant OMEs was likely due to the additional periplas- mic cytochromes, including the avocytochrome FccA present that functioned as the terminal fumarate reductase to support respiration of fumarate in my anaer- obic cultures. Consistent with this interpretation, staining intensity was 3.6-fold stronger in the wild type than in the mutant (p < 0.0001, Student's t-test, two- sample assuming equal variances; Figure 2.8F). Relative to the mutant control, the observed wild type increase in both staining frequency and intensity indicates that the periplasmic and outer membrane cytochromes necessary for EET contribute much of the redox capacity of the OMEs. Given its ability to discriminate between cytochrome-containing and cytochrome-free extracellular laments, and to examine the eect of specic muta- tions, this heme visualization strategy may hold promise for understanding the presence of redox centers in a variety of microbial systems. However, a detailed understanding of the extent to which these redox centers enable long-distance electron transport along OMEs requires: (i) applying electrochemical techniques, recently used to measure redox conduction in biolms [21, 97], specically to OMEs or their MV constituents; and (ii) measurements of the diusive dynamics of redox molecules along membranes, to test the hypothesis that these dynamics facilitate a collision-exchange mechanism of inter-protein electron transport over micrometer 44 Figure 2.8: Presence of multiheme cytochromes important for extra- cellular electron transfer leads to signicantly higher frequency and intensity of redox-dependent staining on outer membrane extensions. (A{D) Transmission electron microscopy images depict outer membrane exten- sions (OMEs, white arrows) stained by 3,3'-diaminobenzidine (DAB; 1 h staining step) in wild type and cytochrome-decient (Mtr/mtrB/mtrE) S. oneidensis MR-1 cells. Cells are indicated by asterisk symbols (*). (A) Wild type OMEs are stained by DAB precipitate. (B,C) Mutant OMEs treated by DAB exhibit vary- ing degrees of staining. (D) Wild type OMEs in chemical controls where H 2 O 2 was omitted appear unstained aside from negative stain. (E) Frequency of staining displayed by OMEs in wild type, Mtr/mtrB/mtrE mutant, and wild type chemical control where H 2 O 2 was omitted. 2.4-fold more OMEs were stained in wild type than in the mutant (p < 0.0001). Statistical signicance was determined by p-value from Pearson's chi-squared test. (F) Intensity of staining displayed by OMEs is 3.6-fold higher in wild type than in Mtr/mtrB/mtrE mutant (p < 0.0001). Statistical signicance was determined by two-tailed p-value from Student's t-test, two-sample assuming equal variances. Error bars show mean ± SEM (Scale bars: 200 nm). 45 length scales [25]. The El-Naggar Lab is actively pursuing these electrochemical measurements, and I describe these dynamics measurements in Chapters 3 and 4. 2.4 Conclusion In this chapter, I investigated physical contributors to the production of OMEs by Shewanella oneidensis MR-1 and applied heme-reactive staining to examine the extent of the redox centers along the extensions. While previous studies focused on the role of oxygen limitation in triggering the formation of these structures, I demonstrated that surface contact is sucient to trigger production of OMEs under a variety of medium, agitation, and aeration conditions. In addition, I showed that the multiheme cytochromes necessary for EET contribute much of the redox- dependent staining widespread on OMEs, and that these EET components do not associate with other extracellular laments. In addition to describing some repro- ducible microscopic and histochemical techniques to observe redox-functionalized membrane extensions, these observations motivate additional studies to under- stand the extent to which Shewanella oneidensis OMEs can contribute to EET and long-distance redox conduction. 46 Chapter 3 Developing a system for site-specic labeling of cell surface cytochromes in Shewanella oneidensis MR-1 This chapter has been adapted from: Grace W. Chong, Sahand Pirbadian, Yunke Zhao, Lori A. Zacharo, Fabien Pinaud, and Mohamed Y. El-Naggar. \Single molecule tracking of bacterial cell surface cytochromes reveals dynamics that impact long-distance electron trans- port." (2021, Article in preparation) 47 3.1 Introduction The goal of the work presented in this chapter was to establish the labeling scheme described in Figure 3.1 for proteins of interest in our system. Specically, I labeled cell surface cytochromes MtrC and OmcA, two key proteins in the Shewanella oneidensis MR-1 electron transport network. The preparatory work performed in this chapter enabled targeted quantum dot labeling and single-particle tracking of individual cell surface cytochromes, which are described in detail in Chapter 4. The combined goal of Chapters 3-4 was to observe and measure the diusion of these electron transport proteins, and to see how their dynamics impact long-distance electron transport along membrane surfaces in living bacterial cells. Figure 3.1: Labeling strategy. (A) Structure of MtrC (PDB ID 4LM8) illus- trates location of biotin acceptor peptide (AP) tag, fused to C-terminus of MtrC (or OmcA) near Heme 10. Hemes and porphyrin rings are colored orange, and AP tag is colored blue. (B) Schematic of labeling strategy. The biotin acceptor peptide (AP: GLNDIFEAQKIEWHE) is fused to MtrC (or OmcA). At the cell sur- face, biotin ligase BirA biotinylates the AP, and quantum dot (QD)-streptavidin conjugates (or other streptavidin conjugates) bind the biotinylated MtrC-AP (or OmcA-AP). 48 3.2 Materials and Methods 3.2.1 Strains, plasmids, and culture conditions Bacterial strains and plasmids used or generated in this study are listed in Table 3.1. Generally, all Luria-Bertani (LB) agar plate cultures were grown overnight at 30 C for S. oneidensis or 37 C for Escherichia coli, or they were grown for up to 3 days at room temperature. All aerobic cultures were grown overnight in LB Table 3.1: Strains and plasmids used in this study. Strain Description or relevant genotype Source S. oneidensis MR-1 Wild type [10] MR-1 mtrC [18] MR-1 omcA [18] MR-1 mtrC pMtrC-AP, Km R This study MR-1 omcA pOmcA-AP, Km R This study MR-1 Mtr/mtrB/mtrE [88] E. coli DH5 Host for cloning Lab collection DH5 pMtrC-AP, Km R This study DH5 pOmcA-AP, Km R This study Plasmids pBBR1-MCS2 Broad range cloning vector, Km R [98] pMtrC-AP mtrC and 118 bp upstream sequence, and biotin acceptor peptide (AP) tag in pBBR1-MCS2, Km R This study pOmcA-AP omcA and 114 bp upstream sequence, and biotin acceptor peptide (AP) tag in pBBR1-MCS2, Km R This study 49 broth at 200 rpm and 30 C for S. oneidensis or 37 C for E. coli. All anaerobic S. oneidensis cultures were prepared by pelleting 5 mL of aerobic overnight LB pre- culture, washing in dened medium [24], and using it to inoculate 100 mL of anoxic dened medium in sealed serum bottles with 30 mM fumarate as the sole electron acceptor. These anaerobic cultures were then allowed to grow for approximately 24 hours at 30 C and 200 rpm where it reached late logarithmic phase (approx. 0.24-0.28 OD600). Frozen stocks of bacterial strains were stored in 30% glycerol at -80 C. Antibiotics (Kanamycin, 50 g/mL) were added to media for bacterial cultures as needed; specically, they were used for all Kanamycin-resistant strains in order to maintain selection for plasmid containing Kanamycin-resistant gene (also noted as KanR or Km R ) at all times. 3.2.2 Plasmid construction Genetic constructs were generated to add a 15-amino acid biotin acceptor pep- tide (AP) tag sequence [99{102] to the C-termini of outer membrane-associated cytochromes MtrC and OmcA in S. oneidensis MR-1. Plasmid design is illus- trated in Figures 3.2 and 3.3 and their construction is described here. To make DNA inserts encoding the AP-tagged genes, S. oneidensis MR-1 genomic DNA template was rst obtained from stationary phase LB cultures using a Mo Bio UltraClean Microbial DNA Isolation Kit. Primers listed in Table 3.2 were used in three consecutive rounds of overhang extension PCR to add the AP tag (DNA sequence CTCGTGCCACTCGATCTTCTGGGCCTCGAA- GATATCGTTCAGGCC; amino acid sequence GLNDIFEAQKIEWHE) connected by a serine-glycine linker (DNA sequence TGAACC, with serine closer to the AP tag), to the end of each gene immediately before the stop codon. Forward primers were also designed to amplify the native promoter region of each gene (118 bp 50 Figure 3.2: Schematic of design for MtrC-AP construct, dubbed pMtrC- AP. (A) Plasmid design. MtrC-AP gene fusion was inserted between restriction sites for XhoI and XbaI in pBBR1-MCS2 plasmid [98] with kanamycin resistance. (B) Insert design. Here, the C-terminus of MtrC was fused to a 45-bp \AP Tag" encoding the 15-amino acid biotin acceptor peptide (AP: GLNDIFEAQKIEWHE) from E. coli, as described in [99, 102]. Insert was generated from S. oneidensis MR-1 genomic DNA template using three rounds of overhang PCR with primers for MtrC listed in Table 3.2. DNA insert included an XhoI restriction site, 118 bp upstream of mtrC (including native promoter), protein-coding region for MtrC, a short glycine-serine linker, and the 45-bp biotin acceptor peptide (AP) tag sequence just before the stop codon and XbaI restriction site. Total length of DNA inserted into pBBR1-MCS2 plasmid was 2185 bp for a nal construct size of 7276 bp. 51 Figure 3.3: Schematic of design for OmcA-AP construct, dubbed pOmcA- AP. (A) Plasmid design. OmcA-AP gene fusion was inserted between restriction sites for XhoI and XbaI in pBBR1-MCS2 plasmid [98] with kanamycin resistance. (B) Insert design. Here, the C-terminus of OmcA was fused to a 45-bp \AP Tag" encoding the 15-amino acid biotin acceptor peptide (AP: GLNDIFEAQKIEWHE) from E. coli, as described in [99, 102]. Insert was generated from S. oneidensis MR-1 genomic DNA template using three rounds of overhang PCR with primers for OmcA listed in Table 3.2. DNA insert included an XhoI restriction site, 114 bp upstream of omcA (including native promoter), protein-coding region for OmcA, a short glycine-serine linker, and the 45-bp biotin acceptor peptide (AP) tag sequence just before the stop codon and XbaI restriction site. Total length of DNA inserted into pBBR1-MCS2 plasmid was 2373 bp for a nal construct size of 7464 bp. 52 upstream of mtrC or 114 bp upstream of omcA, respectively). Primers were also designed to include restriction sites for XhoI and XbaI, plus 5-bp of protection bases at outside ends of each DNA insert. All primers were synthesized by Inte- grated DNA Technologies and obtained as desalted oligonucleotides. All PCR was performed using a Bio-Rad C1000 Touch™ Thermal Cycler, and protocols for each round of overhang PCR are listed in Tables 3.3-3.5 and are based on manufacturer's instructions for Phusion High Fidelity DNA polymerase (New England Biolabs). Annealing temperatures were calculated by the New England Biolabs annealing temperature (Tm) calculator (http://tmcalculator.neb.com/#!/main). For each 35-cycle PCR run, the rst 5 cycles used the annealing temperature calculated for only the template-binding portions of each primer pair, and the last 30 cycles used the annealing temperature calculated for the full length primer pairs. After each PCR experiment, a portion of the sample was separated on a DNA gel (1% agarose in TBE buer, run for60 minutes at 100 V) to check if the PCR worked (i.e. amplied DNA of appropriate size). All DNA gels were run on a Bio-Rad Mini-Sub Cell GT electrophoresis system followed by visualization on a Bio-Rad Gel Doc EZ™ Imager. If PCR was successful, then the PCR product was puried using a Purelink ® PCR Purication Kit for use as template for the next round of extensions via overhang PCR. Finally, after generating the full-length DNA inserts by PCR, they were puried using QIAquick Gel Extraction Kit (Qiagen). All PCR reactions were 25 L in volume, except the nal extension was repeated with 100 L in reaction volume to make more DNA insert for gel purication. Generally, all genomic DNA was stored at 4 C and all primers and PCR products were stored at -20 C in between steps. All DNA concentrations were measured using a Thermo Scientic Nanodrop 2000c spectrophotometer. 53 Table 3.2: Primers used in this study. Name Sequence (5' to 3') Description MtrC Forward GGAACCTCGAGG GGAATTCTATTT CCAGCATCC 5-bp protection bases, XhoI cut site, binds to region upstream of MtrC MtrC Reverse 1 GATATCGTTCAG GCCTGAACCCAT TTTCACTTTAGT GTGATC Binds region of MtrC gene just before stop codon, adds rst part of linker and AP tag sequence OmcA Forward GGAACCTCGAGG ACTTAGTTAGCC TTACAGGTG 5-bp protection bases, XhoI cut site, binds to region upstream of OmcA OmcA Reverse 1 GATATCGTTCAG GCCTGAACCGTT ACCGTGTGCTTCC ATCA Binds region of OmcA gene just before stop codon, adds rst part of linker and AP tag sequence MtrC/OmcA Reverse 2 ACTCGATCTTCT GGGCCTCGAAGA TATCGTTCAGGC CTGAAC Binds to last 20-bp added by Reverse 1 primers in Round 1 of PCR extension; adds more of the AP tag sequence MtrC/OmcA Reverse 3 GGAACTCTAGAT TACTCGTGCCAC TCGATCTTCTGG GCCTCG Binds to last 20-bp added by Reverse 2 primers in Round 2 of PCR extension; adds the rest of the AP tag sequence, stop codon, XbaI cut site, and 5-bp protection bases M13 Forward (-20) GTAAAACGACGG CCAGT Used to check success of cloning and transformation by PCR gel and DNA sequencing M13 Reverse (-27) CAGGAAACAGCT ATGAC Used to check success of cloning and transformation by PCR gel and DNA sequencing 54 Table 3.3: Ingredients used for overhang PCR. Volume is for a 25L reaction. For specic primers used in each step, see Table 3.5. PCR Ingredients Vol (L) Final Conc RNAse Free Water 15.75 5X Phusion HF Buer 5 1X 10 mM dNTPs 0.5 200 M F primer, 10 M stock 1.25 0.5 M R primer, 10 M stock 1.25 0.5 M Template DNA 1 45 ng Phusion DNA Polymerase 0.25 1 unit/50 L rxn Table 3.4: Thermocycler protocol for overhang PCR. For specic annealing temperatures used for each primer pair, see Table 3.5. Temp ( C) Time Cycles Initial Denaturation 98 30 s Denaturation 98 10 s x 5 Annealing 1 * 30 s Extension 72 1 min 15 s Denaturation 98 10 s x 30 Annealing 2 * 30 s Extension 72 1 min 15 s Final Extension 72 7 min Hold 4 innity Table 3.5: Specic primers and annealing temperatures used in 3 consecutive rounds of overhang PCR. Annealing temperatures were chosen for template-binding portions of each primer (Annealing 1) or for the full primer set (Annealing 2) according to the polymerase manufacturer (http://tmcalculator.neb.com/#!/main). Forward Primer Reverse Primer Annealing 1 ( C) Annealing 2 ( C) MtrC Round 1 MtrC Forward MtrC Reverse 1 57 72 MtrC Round 2 MtrC Forward MtrC/OmcA Reverse 2 61 72 MtrC Round 3 MtrC Forward MtrC/OmcA Reverse 3 64 72 OmcA Round 1 OmcA Forward OmcA Reverse 1 56 72 OmcA Round 2 OmcA Forward MtrC/OmcA Reverse 2 56 72 OmcA Round 3 OmcA Forward MtrC/OmcA Reverse 3 56 72 55 The DNA inserts were cloned via restriction digest into the pBBR1-MCS2 broad host cloning vector [98], which is Kanamycin resistant and can also be used for blue-white screening (aka -galactosidase assay) in E. coli. Plasmid samples were puried from overnight LB cultures of its E. coli host using Purelink Quick Plasmid Miniprep Kit (Invitrogen). Restriction digestion reactions using XhoI and XbaI (New England Biolabs) were performed according to manufacturer's guide- lines. Then, the digested plasmid and insert were immediately combined in the recommended 1:3 vector:insert ratio for ligase reactions, following manufacturer's protocol (New England Biolabs Quick Ligation Kit #M2200). Ligation reaction volumes were scaled up to however much digested insert/plasmid DNA was avail- able. Finally, the freshly ligated constructs were plasmid puried and stored at 4 C before transformation via electroporation. 3.2.3 Transformation and quality checking First, electrocompetent E. coli DH5 cells were freshly prepared on ice. For each sample, 1 mL of an overnight LB culture was centrifuged at 7,900 × g for 1 min, washed 3x by gentle pipetting in chilled (4 C) 10% glycerol at 7,900 × g for 2 min. The last 50-70 L of liquid from the 3rd wash was left in each sample for nal resuspension. Then, the electrocompetent cells were incubated with varying amounts (1-10L) of puried, ligated DNA (approx. 5-60 ng in total) for 2 min on ice. Samples were then electroporated in an Eppendorf Eporator ® machine at 1.7- 2 kV with a time constant of5 ms. Then, samples were quickly resuspended in 0.5-1 mL of fresh LB broth and allowed to recover in a 37 C shaking incubator for 90 min. Afterwards, various amounts (50-200 L) of sample were spread onto LB agar supplemented with Kanamycin for antibiotic selection and X-gal (5-bromo-4- chloro-3-indolyl--D-galactopyranoside, 20 mg/mL in dimethylformamide, 60 L 56 spread on agar plates) for blue-white screening (aka -galactosidase assay). If colonies grew and appeared white on LB + Kan + X-gal plates, that suggested they had received plasmid containing the DNA insert, so up to 20 apparently white colonies were re-streaked onto new LB + Kan + X-gal plates 2-3 times to make sure colonies consistently appeared white rather than pale or dark blue. As a control, a no-DNA sample was also included in each electroporation experiment and plated on both LB agar + Kan (no growth expected) and LB agar (growth expected). Colonies that consistently appeared white during blue-white screening (i.e. pos- sibly successful transformants) were then used for PCR and DNA gel electrophore- sis. Here, I used colony PCR (using appropriate Forward and Reverse 1 Primers from Table 3.2) to see if I could amplify respective DNA inserts from the white transformant colonies. To do colony PCR, a sterile pipet tip was used to pick up a portion of a colony and vigorously resuspend it in 10 L of RNAse-free water. Then, 1 L of this mixture was used as template for PCR. For positive control samples, 1 L of puried DNA insert (previously generated by overhang exten- sion PCR and stored in -20 C from previous experiments) was used as template for PCR and used as a comparison for desired DNA size in DNA gels. Detailed protocols for colony PCR are listed in Tables 3.6 and 3.7 and are based on man- ufacturer's instructions for OneTaq Quick-Load DNA Polymerase (New England Biolabs). For each 30-cycle PCR run, annealing temperatures for the rst 5 cycles were calculated for only the template-binding portion of each primer, while the last 25 cycles were calculated for the full length PCR primers. Based on DNA gels of the PCR products, samples that clearly amplied the desired DNA insert were cultured overnight in 5 mL LB broth and stored in 20% glycerol at -80 C to be used in further experiments. 57 Table 3.6: Ingredients used for colony PCR. Volume is for a 50 L reac- tion. For colony PCR, MtrC or OmcA Forward primers were used along with MtrC/OmcA Reverse 1 primer to amplify DNA from putatively transformed E. coli DH5 colonies. Vol (L) Final Conc RNAse Free Water 34.75 5X OneTaq Reaction Buer 10 1X 10 mM dNTPs 1 200 M F primer, 10 M stock 1 0.5 M R primer, 10 M stock 1 0.5 M Template DNA 2 variable OneTaq Quick-Load DNA Polymerase 0.25 1.25 unit/50 L rxn Table 3.7: Thermocycler protocol for colony PCR. Annealing tempera- tures were chosen for template-binding portions of each primer (Annealing 1) or for the full primer set (Annealing 2) according to the polymerase manufacturer (http://tmcalculator.neb.com/#!/main). Temp ( C) Time Cycles Initial Denaturation 94 30 s Denaturation 94 30 s x 5 Annealing 1 44 (MtrC) or 50 (OmcA) 60 s Extension 68 2 min 30 s Denaturation 94 30 s x 25 Annealing 2 62 60 s Extension 68 2 min 30 s Final Extension 68 5 min Hold 4 innity Plasmids (encoding putative AP-tagged MtrC or OmcA) puried from these E. coli transformant cultures were then used for transformation into S. oneidensis MR-1 in their respective gene deletion backgrounds mtrC or omcA [18] using a recently developed electroporation protocol [103]. Control samples with no DNA were also included in each experiment and plated on both LB agar + Kan (no growth expected) and LB agar (growth expected). Several transformants which grew on LB agar + Kan were then grown overnight in LB Broth + Kan. Then, similarly to quality checking the E. coli transformants, the S. oneidensis transfo- mants were checked by PCR (using appropriate Forward and Reverse 1 Primers 58 from Table 2) and DNA gel electrophoresis to see if I could amplify respective DNA inserts from the transformants. As before, samples using the desired DNA insert were used as a positive control. Finally, transformants veried by PCR were frozen at -20% glycerol and stored at -80 C to be further veried by DNA sequencing and SDS-PAGE heme staining. 3.2.4 DNA sequencing Sanger sequencing was performed to verify if successful transformants contained the correct sequence for the desired AP-tagged genes. To prepare samples for sequencing, I amplied the region containing the DNA inserts from puried plas- mid DNA from overnight LB cultures. Here, I used standard M13 Forward and Reverse primers (Integrated DNA Technologies, also listed in Table 3.2) and per- formed PCR according to manufacturer's instructions as described for Phusion High Fidelity DNA Polymerase (New England Biolabs). Puried PCR products were then mixed with M13 Forward primer and sent to GeneWiz for sequencing. Sequencing results were subsequently viewed by SnapGene Viewer software. 3.2.5 SDS-PAGE and heme staining To prepare lysed cell samples for SDS-PAGE gel electrophoresis, LB cultures were aerobically grown overnight, centrifuged as 1 mL samples at 7,900 × g for 2 min, supernatant discarded, and resuspended in 100L of sample buer with reductant and 25L of 4x loading dye. To make stock sample buer, I mixed 3.55 mL distilled water, 1.25 mL 0.5 M Tris HCl, pH 6.8, 2.5 mL glycerol, 2.0 mL 10% (w/v) SDS, and 0.2 mL 0.5% (w/v) bromophenol blue. Sample buer with reductant was made fresh before each experiment adding 50 L of -mercaptoethanol to 950 L of stock sample buer. Then the samples were boiled in a hot water bath 59 for about 5-10 min at 95-100 C before storage at -20 C or immediate use in gel electrophoresis. Samples were loaded onto SDS-PAGE gels (previously prepared using Bio-Rad TGX FastCast Acrylamide Kit) and run at 200 V for1 h on a Bio-Rad Mini-PROTEAN Tetra Cell System. The electrophoresis running buer was made from 3.03 g/L Tris base, 14.4 g/L glycine, and 1 g/L SDS, pH 8.3. Samples were often run on replicate gels. One gel was reserved for staining with Instant Blue™ (Expedion) total protein dye. Another gel was then visualized by a peroxidase activity-based heme staining technique. Heme staining via peroxidase activity assay with 3,3'-diaminobenzidine (DAB) was adapted from a previously described procedure [104]. After electrophoresis, gels were immediately pre-incubated for 30 min at room temperature in a solution containing freshly made 25 mg DAB in 50 mL of 0.5 M Tris Buer, pH 7 (sonicated as needed up to 1 h). Then, this solution was discarded, and the gels were incubated 30 min to overnight at 4 C in a second solution containing peroxide (50 mg DAB in 50 mL of 50 mM Tris HCl, pH 8, sonicated as needed up to 1 h, with 1.5 mL of 30% hydrogen peroxide added right before use). Staining was generally visible within 30 min after adding peroxide. 3.2.6 In vivo biotinylation Anaerobically pre-grown S. oneidensis cells were harvested by centrifugation for 10 min at 7,142 × g, washed in PBS buer supplemented with 5 mM MgCl 2 (PBS-Mg) for 5 min at 4,226 × g, and resuspended in PBS-Mg, and collected in 1.5-mL tubes with 0.5 mL of cells diluted to 0.8 OD600 per sample. These samples were washed once again in PBS-Mg for 2 min at 7,900 × g, and their supernatant was removed, leaving the cell pellet. The samples were then biotinylated in vivo using a BirA biotin-protein ligase standard reaction kit (Avidity). Following the 60 kit instructions, each cell pellet was quickly resuspended in a 50-L biotin ligase reaction mixture and left at room temperature for 1 h with vigorous shaking on an orbital shaker. Each 50-L reaction mixture contained 50 mM bicine buer (pH 8.3), 10 mM ATP, 10 mM MgCl 2 , 50M biotin, and 0.3M of BirA biotin ligase, dissolved in RNAse-free water. If necessary to prepare bigger samples, sample and reaction sizes were scaled up proportionately. 3.2.7 Western blot Western blot experiments were also performed to test that in vivo exogenous biotinylation worked and that this labeling was specic to tagged proteins of inter- est. Freshly biotinylated cell samples were washed 2x in PBS buer and prepared for SDS-PAGE gel electrophoresis as described above. While SDS-PAGE gels were running, ber pads, lter paper, and nitrocellulose membrane were cut to appro- priate gel size and soaked for>20 min in chilled transfer buer containing 25 mM Tris, pH 8.3, 192 mM glycine, and 20% v/v methanol. After SDS-PAGE, gels were also soaked in chilled transfer buer for>20 min. Then, proteins were transferred from gel to membrane in a Bio-Rad Mini Trans Blot Transfer Cell system run at 100 V and 350 mA for 1 h, in chilled and stirred transfer buer. Afterwards, the membrane was removed and incubated in 10 mL of blocking solution made from 3% BSA in TBST (20 mM Tris-HCl, 140 mM NaCl, pH 7.5, 0.1% v/v Tween- 20) and left for 1 h at room temperature or overnight at 4 C. After blocking, the membrane was washed 3x in TBST for 5 min with gentle shaking. Then, it was incubated with diluted polyclonal rabbit-raised MtrC or OmcA-specic pri- mary antibody [105] (rMtrC: 0.66 g/mL, rOmcA: 0.49 g/mL, both in 1% BSA/TBST) at room temperature for 30 min with gentle shaking. After 3 more 5- min washes in TBST, the membrane was then incubated in horseradish peroxidase 61 (HRP)-conjugated goat anti-rabbit secondary antibody (Abclonal) diluted 1:2000 in 1% BSA/TBST for 30 min at room temperature with gentle shaking. Since the HRP is light sensitive, this incubation and future steps were done in a covered con- tainer. Then the membrane was washed 3x in TBST, 3x in TBS, and 1x in water for 5 min each with gentle shaking. Finally, the membrane was visualized using SuperSignal™ West Pico PLUS Chemiluminescent Substrate (varying incubation time and exposure time as appropriate) and developed on X-ray lm. Then, I prepared the blots for restaining by using a mild stripping procedure adapted from Abcam: blots were incubated 2x for 5-10 min in mild stripping buer (For 1 L: 15 g glycine, 1 g SDS, water, adjusted to pH 2.2) at room temperature over gen- tle shaking, then washed 2x for 10 min in PBS, and 2x for 5 min in TBST. The stripped membranes were then blocked for 1 h in 3% BSA/TBST and washed 3x for 5 min in TBST. Then, they were incubated in HRP-conjugated streptavidin (Thermo Fisher Scientic) diluted to 0.25g/mL in 1% BSA/TBST, gently shak- ing at room temperature for 30 min. Finally, the membranes were washed 3x in TBST, 3x in TBS, and 1x in water for 5 min each and visualized with chemilumi- nescent substrate and developed on X-ray lm as described above. The rst set of lms were used to detect proteins of interest using MtrC and OmcA-specic antibodies (though there is some nonspecic binding, where MtrC antibody also binds to OmcA, and vice versa). Since the streptavidin binding is not reversible, this was probed last. The second set of lms, which captured signal from streptavidin, revealed which proteins are biotinylated, and could be compared to the rst set to verify overlap of signal, and thus specicity of biotinylation to proteins of interest. 62 3.2.8 Fluorescence microscopy Cells were prepared for in vivo microscopy by exogenous biotinylation as described above. Generally, biotinylated samples were washed 6 times in PBS at 12,000 rpm for 3 min each. Then, they were incubated with streptavidin-conjugated label. For the microscopy labeling control experiments described in this thesis chapter (Figures 3.5B and 3.6B), 20 nM streptavidin-conjugated Alexa Fluor 647 (SA-AF647, Thermo Fisher Scientic) was used as the uorescent label instead of quantum dots, due to cost and volume of reagent available. Biotinylated cell samples were then incubated in 20 nM SA-AF647 in 6% BSA in PBS buer for 1 h at room temperature with vigorous shaking. Reaction size was proportionate to initial cell sample (50 L reaction size per 0.5 mL of cells initially diluted to 0.8 OD600). Then, samples were washed 3 more times in PBS and resuspended in a small amount (e.g. 20 L) of PBS. Samples were then ready for imaging. Samples were mounted on high precision microscope glass coverslips (Marien- feld, #1.5, 25 mm) at the bottom of an open-air liquid imaging chamber which was custom-made from stainless steel and could hold up to 1 mL of liquid sample. The steel liquid holders were cleaned by soaking in ethanol (70-100%) and dried before each use. Coverslips were cleaned before use by wiping with ethanol (70- 100%), then rinsed with distilled water, and dried, before loading into the liquid holder. Generally, a small volume of cells (e.g. 5-10L) were dropped in the center of the coverslip. Then 1 mL of PBS was gently pipetted into the chamber. This order was important to promote better cell attachment to the coverslip. Finally, about 5-10 min prior to imaging, green membrane dye FM 1-43FX (Life Technolo- gies; 0.0625-0.125g/mL) was added to the sample in the liquid holder and gently pipetted to mix. 63 Imaging was performed on an inverted Nikon Eclipse Ti-E microscope equipped with total internal re ection optics, a 100× 1.49 NA objective (Nikon), an X-Cite 120XL uorescence illumination system, two iXon Ultra EMCCD cameras (Andor Technology), a dual camera light path splitter (Andor Technology), and laser lines for excitation at 488 and 647 nm (Agilent). For detection of the split green/red signal, a multiband pass ZET405/488/561/647x excitation lter (Chroma), a quad- band ZT 405/488/561/647 dichroic mirror (Chroma), and an emission splitting FF640-FDi01 dichroic mirror (Semrock) were used in combination with appropriate emission lters: ET525/50 (Chroma) for FM 1-43FX, and ET700/75 (Chroma) for AF647. Channels were aligned prior to imaging using 40 nM TransFluoSphere streptavidin-labeled beads (488/645 nm, Life Technologies) as ducial markers. 3.3 Results and Discussion 3.3.1 Successful and specicin vivo labeling of cell surface cytochromes MtrC and OmcA I used the labeling scheme described in [100{102] to label cell surface cytochromes MtrC and OmcA in S. oneidenis MR-1. Brie y, as pictured in Figure 3.1, a 15- amino acid biotin acceptor peptide (AP) tag from E. coli [99] was fused to the C-termini of MtrC and OmcA. Once assembled in the periplasm and exported to the outer membrane, cytochromes expressing the AP tag can then be biotinylated externally by the addition of biotin ligase (BirA, puried from E. coli). Finally, the biotinylated cytochromes can then be detected by streptavidin-conjugated probes, which would allow the labeled cytochromes to be imaged in real-time by microscopy. Another benet of this labeling scheme, which combines a small peptide tag with extracellular labeling (Figure 3.1B), is to minimize interference 64 to the localization of MtrC and OmcA on the outer surface of the cell, where peptides produced in the cytoplasm are transported to the periplasm for protein folding and heme assembly before being exported to the extracellular side of the outer membrane [106]. DNA inserts for AP-tagged MtrC and OmcA (Figures 3.2 and 3.3) were con- structed by overhang PCR and cloned into the pBBR1-MCS2 plasmid [98]. Plas- mid constructs were eventually transformed into respective S. oneidensis MR-1 mtrC or omcA deletion backgrounds from [18]. All strains, plasmids, and primers used in this study are listed in Tables 3.1 and 3.2. Cytochrome c heme content was then detected by staining SDS-PAGE gels with a peroxidase activity assay using 3,3'-diaminobenzidine (DAB) and hydrogen peroxide (H 2 O 2 ) (Figure 3.4), which conrmed that the AP-tagged strains produced heme-containing pro- teins of expected size, compared to positive and negative controls (wild type and gene deletion mutant). Sanger sequencing of plasmids puried from nal host strains was also performed to verify sequence integrity of the AP tag. 65 Figure 3.4: Heme staining of protein gels via peroxidase activity assay reveals heme in newly tagged cytochromes MtrC-AP and OmcA-AP. Redox-dependent staining using a 3,3'-diaminobenzidine (DAB) and hydrogen per- oxide (H 2 O 2 ) peroxidase activity assay. SDS-PAGE and subsequent staining was performed using whole cell lysate from liquid cultures of respective S. oneidensis strains labeled at the top of each lane. Also labeled are relevant bands in pro- tein ladder (80 kDa, 58 kDa), as well as the approximate positions of proteins of interest (MtrC, OmcA) and fumarate reductase (FccA) which is present in all samples. Lanes 4 and 5 in both gels contain a dark band conrming the presence of heme associated with protein of interest (AP-tagged MtrC or OmcA). Wild type sample (Lane 1) was used as a positive control; respective gene deletion mutants mtrC and omcA (Lanes 2 and 3) are included as negative controls missing pro- tein of interest MtrC or OmcA; and included as a secondary negative control is a cytochrome mutant (Mtr/mtrB/mtrE, Lane 6) missing a total of 8 periplas- mic and outer membrane-associated cytochromes (including proteins of interest MtrC and OmcA). 66 Next, I performed Western blot and microscopy controls where I systemati- cally omitted key components in the labeling process, to conrm that the labeling scheme works in our system (Figures 3.5 and 3.6). In Western blots probing for biotinylated proteins using streptavidin-horseradish peroxidase (HRP), MtrC-AP (or OmcA-AP) were detected only when all key components of the labeling pro- cess were provided (Figure 3.5A and 3.6A), indicating that the tagged protein was successfully and specically biotinylated and detected by the streptavidin probe. Similarly, microscopy labeling controls were performed, where biotinylated pro- teins were visualized by streptavidin-conjugated Alexa Fluor 647 (Figures 3.5B and 3.6B). Though cells were visible by standard wideeld imaging in all samples, strong uorescent signal visualizing biotinylated proteins are only detected in the condition where all key labeling components are present. Taken collectively, these Western blot, uorescence microscopy, and associ- ated controls demonstrate successful and specic labeling of MtrC and OmcA. Furthermore, my ability to perform extracellular in vivo labeling and subsequent microscopic detection of MtrC and OmcA via a C-terminal AP tag is consistent with the recently published orientation of MtrC relative to the MtrAB transmem- brane complex (Figure 1.3) [15], where Heme 10 (C-terminal side) is extracellularly exposed and Heme 5 (N-terminal side) is facing the cell surface. 67 Figure 3.5: Key labeling controls demonstrate successful and specic labeling of MtrC. (A) Western blot labeling control for MtrC where key parts of the labeling process were systematically omitted. When using streptavidin (streptavidin-horseradish peroxidase, HRP) to probe for biotinylated proteins, a thick dark band of biotinylated MtrC-AP is detected only in Lane 5 when all key components are present. The faint band slightly below labeled MtrC-AP (approx. 79.6 kDa) and present in all samples is an endogenously biotinylated protein (acetyl-CoA carboxylase, approx. 76 kDa). (B) Microscopy labeling con- trol for MtrC where key parts of the labeling process were systematically omitted. Top row are wideeld (WF) images showing many cells in each sample. Bottom row images show uorescence (Fl) signal from streptavidin-conjugated Alexa Fluor 647 (SA-AF647) that was used to detect biotinylated MtrC-AP; uorescence label- ing was detected strongly in the bottom right image, and only when all key labeling components were present. All microscopy images are approx. 36.5m by 36.5m. 68 Figure 3.6: Key labeling controls demonstrate successful and specic labeling of OmcA. (A) Western blot labeling control for OmcA where key parts of the labeling process were systematically omitted. When using streptavidin (streptavidin-horseradish peroxidase, HRP) to probe for biotinylated proteins, a thick dark band of biotinylated OmcA-AP is detected only in Lane 5 when all key components are present. The faint band slightly below labeled OmcA-AP (approx. 87 kDa) and present in all samples is an endogenously biotinylated pro- tein (acetyl-CoA carboxylase, approx. 76 kDa). (B) Microscopy labeling control for OmcA where key parts of the labeling process were systematically omitted. Top row are wideeld (WF) images showing many cells in each sample. Bottom row images show uorescence (Fl) signal from streptavidin-conjugated Alexa Fluor 647 (SA-AF647) that was used to detect biotinylated OmcA-AP; uorescence labeling was detected strongly in the bottom right image, and only when all key labeling components were present. All microscopy images are approx. 36.5m by 36.5m. 69 3.4 Conclusion In this chapter, I labeled bacterial cell surface cytochromes for the purpose of ultimately measuring their dynamics in vivo. I performed a series of molecular biology techniques to fuse a biotin acceptor peptide (AP) tag to the C-termini of bacterial cell surface cytochromes MtrC and OmcA, two key components of the S. oneidensis EET network. This labeling scheme combines in vivo biotinylation of the AP tag, followed by detection with streptavidin conjugates. Using a series of systematic labeling controls, I demonstrated successful and specic labeling of two proteins of interest in living cells. This work paves the way for the work presented in the next chapter, where I used this labeling scheme to observe and measure the mobility of individual cytochromes on the surface of living cells. 70 Chapter 4 Single molecule tracking of bacterial cell surface cytochromes reveals dynamics that impact long-distance electron transport This chapter has been adapted from: Grace W. Chong, Sahand Pirbadian, Yunke Zhao, Lori A. Zacharo, Fabien Pinaud, and Mohamed Y. El-Naggar. \Single molecule tracking of bacterial cell surface cytochromes reveals dynamics that impact long-distance electron trans- port." (2021, Article in preparation) Note: The dynamics measurements described in this chapter nally made it possible to perform simulations of overall electron transport along membrane sur- faces in S. oneidensis that take into account both electron hopping and cytochrome diusion. These simulations, described in Figure 4.11, Methods Section 4.2.5, and Results Section 4.3.3, were designed and performed by Sahand Pirbadian. 71 4.1 Introduction Redox reactions are a fundamental part of how many organisms extract energy for life; this process involves the transfer of electrons from an electron donor to an electron acceptor, through the cellular electron transport chain [2]. Shewanella oneidensis MR-1 is a Gram-negative, facultative anaerobic bacterium that can gain energy by utilizing a diverse array of electron donors and acceptors, from sol- uble substances like oxygen to insoluble objects outside the cell surface, including minerals and electrodes [11]. This respiratory versatility is made possible by a series of multiheme c-type cytochromes that transport electrons from the electron transport chain on the inner membrane, across the otherwise electrically insulat- ing periplasmic space and outer membrane, to solid materials outside the cell, in a process known as extracellular electron transfer (EET) [1, 11, 13]. The capabil- ity to perform EET makes S. oneidensis and other electroactive microorganisms particularly interesting for applications in bioelectrochemical technologies, such as microbial fuel cells and microbial electrosynthesis [1, 107, 108], as well as emerg- ing concepts for living electronics [72]. Since the discovery of S. oneidensis [10], extensive studies have revealed the EET network of cytochromes that bridge the cell envelope, including inner membrane tetraheme cytochrome CymA, periplasmic cytochromes such as STC, outer membrane porin-cytochrome complexes such as MtrAB and cell surface cytochromes such as MtrC and OmcA [13, 15, 16]. MtrA is a decaheme cytochrome located on periplasmic side of the outer membrane and connected to the surface by the transmembrane porin MtrB, where the outward- most heme in MtrA can then interact with cell surface cytochromes such as MtrC and OmcA [15, 16], which in turn act as an external interface between the cell and extracellular electron acceptors [13]. The decaheme outer membrane-associated 72 MtrC and OmcA are largely extracellularly exposed [15], attached to the cell sur- face by a lipidated cysteine at the N-terminus [17]. Once electrons have reached the surface of the cell, they can be transferred to extracellular electron acceptors by direct contact with these cell surface cytochromes or by indirect contact via soluble redox shuttles, such as avins [1]. In addition to bridging the gap between the electron transport chain on the inner membrane and electron acceptors outside the cell, EET components can also enable long-distance (micrometer scale) lateral electron transport along mem- branes and across multiple cells, as demonstrated recently with electrochemical gating measurements of electron conduction through S. oneidensis cellular mono- layers connecting interdigitated electrodes [21]. This multi-cell redox conduction process is dependent on the presence of the Mtr/Omc EET pathway cytochromes, and exhibits a thermal activation energy consistent with the activation barrier for transport through the decaheme chain of MtrC [21, 109]. These observations, and other remarkable demonstrations of cytochrome-mediated redox conduction in electroactive biolms [97], motivated a better understanding of the cytochrome density and physical electron transport mechanism that can give rise to long- distance conduction. Previous measurements estimated high densities of MtrC and OmcA on the S. oneidensis cell surface (up to 30,000 proteins/m 2 ) [110], but this surface coverage is not sucient to provide a crystalline-like packing that allows direct inter-protein electron hopping along the full micrometer-scale conduc- tion path. Additional knowledge about the cytochrome distribution was recently provided by electron cryotomography (ECT) of the S. oneidensis outer membrane extensions [25]. These extensions, proposed to function as nanowires for elec- tron transport, contain the EET components, are known to form after cell-surface attachment, and have been observed up to 100 m in length at an elongation rate 73 of 40 m/h [24, 68]. The ECT observations revealed a heterogeneous distribution of outer membrane-associated cytochromes, with inter-protein spacings ranging from immediately-adjacent to being separated by tens of nanometers [25]. In light of these ndings, we previously proposed a collision-exchange model (Figure 4.1A), where the lateral diusion of the multiheme cytochromes leads to collisions and inter-protein electron exchange along the membrane [25]. This mechanism, which accounts for both direct electron hopping between redox centers and their phys- ical diusion, played a critical role in understanding conduction through redox polymers [111], but remains underexplored in the context of EET. Recent studies have indeed hinted at the importance of cytochrome mobility as a possible con- tributor to long-distance conduction [21, 25, 112, 113], but this contribution has not been veried by experimental measurements of the diusive dynamics of the membrane-associated EET cytochromes in vivo. While it is expected that membrane components are capable of diusion, it turns out that such dynamics have rarely been measured for outer membrane proteins in Gram-negative bacteria. In fact, few unique outer membrane proteins have actually been studied, and most existing studies generally involve -barrel porin or channel-type integral membrane proteins in Escherichia coli [66, 114, 115]. Out of several commonly used methods for studying diusion dynamics, single-particle tracking (SPT) combined with total internal uorescence (TIRF) microscopy provides several advantages, such as high spatial resolution that enables precise localization and tracking of individually labeled proteins [116]. The absence of experimental data on the mobility of EET components motivated the application of single-molecule techniques to assess their dynamics on the outer membrane of S. oneidensis. To my knowledge, this work presents the rst measurements of the diusive dynamics of bacterial extracellular electron conduits, and it also adds to 74 the relatively short list of cell surface proteins whose diusion was measured in prokaryotes. Figure 4.1: Lateral diusion and labeling strategy. (A) Schematic of diusion-assisted electron hopping along the Shewanella oneidensis MR-1 outer membrane. Lateral motion of multiheme cytochromes (diusion coecient D phys ) leads a collision-exchange mechanism of inter-protein electron transport over large distances. Red spots represent hemes in multiheme cytochromes. Labeled proteins are outer membrane cytochromes MtrC and OmcA, outer membrane-associated periplasmic cytochrome MtrA, and outer membrane porin MtrB. (B) Structure of MtrC (PDB ID 4LM8) illustrates location of biotin acceptor peptide (AP) tag, fused to C-terminus of MtrC (or OmcA) near Heme 10 as described in Chapter 3. Hemes and porphyrin rings are colored orange, and AP tag is colored blue. (C) Schematic of labeling strategy, established in our system as described in Chapter 3. The biotin acceptor peptide (AP: GLNDIFEAQKIEWHE) is fused to MtrC (or OmcA). At the cell surface, biotin ligase BirA biotinylates the AP, and QD- streptavidin conjugates bind the biotinylated MtrC-AP (or OmcA-AP). 75 The combined goals of Chapters 3 and 4 were to (1) visualize individual cytochromes on living cells, (2) assess their mobility along membrane surfaces, (3) quantify their diusive dynamics, and (4) investigate how this diusion impacts overall electron transport, in the context of the collision-exchange mechanism. Chapter 3 described necessary preparatory work required to observe the dynam- ics of individual cytochromes. I tagged the S. oneidensis cell surface cytochromes MtrC and OmcA with biotin acceptor peptides, and established that the biotin- streptavidin labeling scheme works in our system (Figure 4.1B-C). In Chapter 4, I now directly address the 4 overall goals outlined above. Following the established labeling scheme, I performed site-specic targeting of biotinylated proteins with streptavidin-conjugated quantum dots. Then, I measured the diusive dynamics of these electron conduits with in vivo TIRF microscopy and single-particle tracking [64, 100, 101, 117]. I quantied the diusion coecients of both MtrC and OmcA along the surface of the outer membrane and membrane nanowires. We also cal- culated the contribution of these dynamics to long-distance electron conduction through kinetic Monte Carlo simulations. Altogether, this study suggests that the dynamics of EET components play an important role in overall electron transport over micrometer length scales. 76 4.2 Materials and Methods 4.2.1 Cell cultivation In this chapter, I focus on the strains of S. oneidensis MR-1 that were engineered in Chapter 3 to contain AP-tagged MtrC or OmcA in their respective gene deletion backgrounds (Table 4.1). First, aerobic cultures were grown overnight in LB broth at 200 rpm and 30 C. Then, anaerobic cultures were prepared by pelleting 5 mL of aerobic overnight LB pre-culture, washing in dened medium [24], and using it to inoculate 100 mL of anoxic dened medium in sealed serum bottles with 30 mM fumarate as the sole electron acceptor. These anaerobic cultures were then allowed to grow for approximately 24 hours at 30 C and 200 rpm where it reached late logarithmic phase (approx. 0.24-0.28 OD600). Frozen stocks of bacterial strains were stored in 30% glycerol at -80 C. Antibiotics (Kanamycin, 50 g/mL) were added to media for bacterial cultures where appropriate; specically, they were used for all Kanamycin-resistant strains in order to maintain selection for plasmid containing Kanamycin-resistant gene (also noted as KanR or Km R ) at all times. Table 4.1: Strains used in this study. Strain Description or relevant genotype Source S. oneidensis MR-1 mtrC pMtrC-AP, Km R This study MR-1 omcA pOmcA-AP, Km R This study 4.2.2 In vivo biotinylation Anaerobically pre-grown S. oneidensis cells were harvested by centrifugation for 10 min at 7,142 × g, washed in PBS buer supplemented with 5 mM MgCl 2 77 (PBS-Mg) for 5 min at 4,226 × g, and resuspended in PBS-Mg, and collected in 1.5-mL tubes with 0.5 mL of cells diluted to 0.8 OD600 per sample. These samples were washed once again in PBS-Mg for 2 min at 7,900 × g, and their supernatant was removed, leaving the cell pellet. The samples were then biotinylated in vivo using a BirA biotin-protein ligase standard reaction kit (Avidity). Following the kit instructions, each cell pellet was quickly resuspended in a 50-L biotin ligase reaction mixture and left at room temperature for 1 h with vigorous shaking on an orbital shaker. Each 50-L reaction mixture contained 50 mM bicine buer (pH 8.3), 10 mM ATP, 10 mM MgCl 2 , 50M biotin, and 0.3M of BirA biotin ligase, dissolved in RNAse-free water. If necessary to prepare bigger samples, sample and reaction sizes were scaled up proportionately. 4.2.3 Total internal re ection uorescence (TIRF) microscopy Cells were prepared for in vivo microscopy by exogenous biotinylation as described above. Generally, biotinylated samples were washed 6 times in PBS at 12,000 rpm for 3 min each. Then, they were incubated with streptavidin-conjugated label Qdot™ 705 Streptavidin Conjugate (SA-QD705, Thermo Fisher Scientic). Quantum dots (QDs) were prepared as a range of concentrations (e.g. 0.1-10 nM) in 6% BSA in PBS buer, and samples were incubated in QD solution for 1 h at room temperature with vigorous shaking. Reaction size was proportionate to initial cell sample (50 L reaction size per 0.5 mL of cells initially diluted to 0.8 OD600). Then, samples were washed 3 more times in PBS and resuspended in a small amount (e.g. 20 L) of PBS. Samples were then ready for imaging. Samples were mounted on high precision microscope glass coverslips (Marien- feld, #1.5, 25 mm) at the bottom of an open-air liquid imaging chamber which 78 was custom-made from stainless steel and could hold up to 1 mL of liquid sample. The steel liquid holders were cleaned by soaking in ethanol (70-100%) and dried before each use. Coverslips were cleaned before use by wiping with ethanol (70- 100%), then rinsed with distilled water, and dried, before loading into the liquid holder. Generally, a small volume of cells (e.g. 5-10L) were dropped in the center of the coverslip. Then 1 mL of PBS was gently pipetted into the chamber. This order was important to promote better cell attachment to the coverslip. Finally, about 5-10 min prior to imaging, green membrane dye FM 1-43FX (Life Technolo- gies; 0.0625-0.125g/mL) was added to the sample in the liquid holder and gently pipetted to mix. Imaging was performed on an inverted Nikon Eclipse Ti-E microscope equipped with total internal re ection optics, a 100× 1.49 NA objective (Nikon), an X-Cite 120XL uorescence illumination system, two iXon Ultra EMCCD cameras (Andor Technology), a dual camera light path splitter (Andor Technology), and laser lines for excitation at 488 and 647 nm (Agilent). For detection of the split green/red signal, a multiband pass ZET405/488/561/647x excitation lter (Chroma), a quad- band ZT 405/488/561/647 dichroic mirror (Chroma), and an emission splitting FF640-FDi01 dichroic mirror (Semrock) were used in combination with appropriate emission lters: ET525/50 (Chroma) for FM 1-43FX, and ET700/75 (Chroma) for AF647. Channels were aligned prior to imaging using 40 nM TransFluoSphere streptavidin-labeled beads (488/645 nm, Life Technologies) as ducial markers. Time-lapse microscopy was then performed in dual colors at an image acquisition rate of 40 ms/frame in each channel. During each imaging experiment, the signal of interest (SA-QD705) was gener- ally notable due to brightness and shape/size of signal, as well as blinking behavior 79 characteristic of single molecules. Generally, samples with 1-2 quantum dot parti- cles visible per cell were imaged for tracking and diusion analyses, since it makes it easier to localize the signal of interest and build trajectories over time. Since the eciency of quantum dot labeling apparently varied over time between exper- iments (perhaps due to degradation of label over time in storage), samples treated with a gradient of label concentration were prepared prior to each imaging exper- iment so that we could image at least one sample with a desirable concentration of label particles per cell. Biotin ood microscopy experiment For the biotin ood experiment (Figure 4.10), the quantum dot labeling step was modied to \ ood" the sample with excess biotin shortly after addition of SA- QD705. The purpose of this step was to plug excess protein binding sites on the QD probes and prevent individual probes from binding to multiple protein targets. Brie y, following 5 min after addition of SA-QD705, >200-fold excess biotin dissolved in PBS buer was added. The required amount of excess biotin was calculated by the following estimations: assuming 5-10 wild type streptavidin molecules per QD based on manufacturer estimates, 10 streptavidin molecules would have 40 biotin sites, requiring at least 40-fold excess biotin to plug excess possible streptavidin-biotin binding sites. To obtain >200-fold excess biotin, I added 200 L of 1 M biotin to a 50-L sample containing 20 nM QD. Samples were incubated at room temperature with shaking for 15 min prior to nal wash steps and imaging. 80 4.2.4 Single-particle tracking (SPT) and diusion analyses Single-particle localization and tracking was performed using SLIMfast, a program written for MATLAB that uses multiple-target tracing algorithms [118] and can accommodate for the blinking behavior of single molecules. The process of using SLIMfast for SPT has recently been described in detail for a study in Caenorhabdi- tis elegans by [119]. First, Fiji (ImageJ) software was rst used to convert the time- lapse microscopy data to TIFF image sequence les that could then be opened by SLIMfast in MATLAB. Then, SLIMfast was used to localize single molecules based on 2D Gaussian tting of the point spread function from each streptavidin-bound label (e.g. SA-QD705 particles) in each frame. SLIMfast localization settings include: error probability (10 -6 ), evaluation box (9 px), de ation loops (100), PSF model (xed), PSF radius (1.75 px, the theoretical value calculated by SLIMfast according to numerical aperture 1.49, emission wavelength 705 nm, correlation fac- tor 1.2). Trajectories were then built by connecting the localized position of each particle over time from frame to frame, taking into account blinking statistics. Trajectories with at least 4 steps (i.e., localized for at least 5 consecutive frames) were used for diusion analyses based on mean square displacement (MSD) for individual trajectories or probability of square displacement (PDSD) and ensem- ble MSD analysis for all trajectories, as described in [119, 120]. Other SLIMfast tracking settings include: local density window (10 frames), frame-to-frame linking search limit (0.5 to 10 px), track assembly search limit (1 to 10 px), max gap clo- sure (10 frames). For tracking on the cell surface, about 5,000-10,000 trajectories from 800-2,000 cells were analyzed for each condition. For added quality checking, several elds in each dataset were manually inspected frame-by-frame in ImageJ to verify overlay of particle localizations (specically those localizations used to gen- erate trajectories, as some localizations are discarded during the tracking process) 81 with raw signal from red channel (e.g. quantum dot signal). Similarly, those elds were also inspected to ensure that those localizations used for trajectories were found within cells, as labeled by raw signal from green channel (i.e. membrane dye). Also, if needed to remove stray trajectories e.g. from clearly moving cells, or to limit the region of interest to an outer membrane extension, regions of interest (ROI) were dened in SLIMfast, and localization and tracking were repeated with those ROI to obtain nal trajectories for analysis. First, MSD curves for each individual trajectory were generated in SLIMfast. Individual diusion coecients were calculated by tting each MSD curve for the rst 3 time lags of the trajectory with the following model for free diusion, accounting for position error: r 2 = 4Dt + 4 2 (4.1) where r 2 is the square displacement (or MSD), D is the diusion coecient, t is the time lag or time interval, and is the position error. Diusion coecients for each trajectory were then plotted as a histogram in Origin 2019b software. This histogram distribution of free diusion coecients from all trajectories was then tted with Gaussian distributions to yield overall estimates of free D values for apparent populations. This type of analysis helps to illustrate overall diu- sion (assuming free diusion) and can help distinguish if there are very obvious faster/slower diusing populations. Next, probability distribution of square displacement (PDSD) analysis was per- formed in SLIMfast on all trajectories to determine the number of major popu- lations of diusing behavior for each dataset. Brie y, cumulative probability dis- tribution functions of square displacements (r 2 ) for selected t were generated in 82 SLIMfast and tted with a model for 1, 2, or 3 (or more) populations. This model is described in detail by Sch utz et al. under their equations 3 (model for 1 popula- tion) and 5 (model expanded for 2 populations) and can be similarly expanded for 3 or more populations [121]. Further description can also be found in [119, 120]. This is an intermediate step in SLIMfast that sorts the raw displacement data into populations that can then be plotted for ensemble MSD curve analysis, described below. Ensemble MSD curves were then generated for each diusing population in each dataset (populations determined by PDSD analysis). Ensemble MSD curves are also called time-averaged MSD curves, bulk MSD curves, or r 2 (t) curves (mean squared displacement r 2 as a function of time lag t). The shape of this curve indicates the overall diusing behavior (e.g. free, conned) of that population, and subsequently the curve can be tted with appropriate diusion model (e.g. free, conned) which yields information of interest such as diusion coecient D and connement radius R. SLIMfast was used to calculate ensemble MSDs from raw displacement data and the MSD curves were then plotted in Origin 2019b software. Error bars at each time point of the ensemble MSD curve show Y-error = r 2 p N , where N is the number of data points (or displacements) that were averaged to give the mean squared displacement r 2 value for each time interval t. For 2-population MSD analyses, I plotted MSD curves with SEM-based error bars that also take into account the variation in displacement for a given t, with the formula Y-error = SEM p N(%N) , where %N is the fraction of all displacements N that had been allocated to the given sub-population for a given t. Diusion 83 coecients (D) based on conned diusion were determined by tting ensemble MSD curves with a circularly conned diusion model: r 2 =R 2 (1A 1 e 4A 2 Dt R 2 ) + 4 2 (4.2) where R is the corral/connement radius, is the position error, A 1 = 0.99 and A 2 = 0.85 (constants dened by corral geometry, in this case circular connement), as described in [116, 120]. All diusion coecients are reported in micrometer squared per second± stan- dard deviation (SD) of the t value. 4.2.5 Kinetic Monte Carlo simulations of long-distance electron transport Simulations of the overall electron transport (ET) rate along the cell surface or membrane extensions are dependent on the relative rates of direct electron hop- ping vs. redox carrier diusion, characterized by the ratio t e t p , where t e and t p are the time constants of electron hopping and physical motion of redox carriers, respectively [111]. These time constants are related to the diusion coecients for electron hopping between redox carriers (D e ), and the physical motion of the redox carriers (D phys ) according to the following equations: D e = 2 4t e ;D phys = 2 4t p (4.3) where is the center-to-center distance of closest approach between redox carriers and is estimated as the average size of one redox carrier. 84 For our system, t e is estimated by the electron residence time in outer mem- brane decaheme cytochromes to be approximately 10 -5 to 10 -6 s from molecular simulations of the electron ux [109, 122, 123], and t p can be estimated from my diusion measurements of MtrC and OmcA in this study using Eqn. 4.3; taking = 6.33 nm for MtrC [14], D phys values in the 10 -2 to 10 -1 m 2 /s translate to a t p of 10 -3 to 10 -4 s. Since t e < t p in our system, we ruled out the mean-eld approach discussed in [25, 111], and instead performed Monte Carlo simulations in MATLAB that randomly simulate direct electron hopping and redox carrier diusion, following the Blauch{Saveant approach [111] for modeling electron transport in an assembly of redox carriers in two dimensions. In these simulations, the cylindrical surface of a cell or membrane extension with length L and diameter d was represented by a 2D lattice array, where one dimension of the lattice (y-axis) was the length (L) of the surface (cell or membrane extension) and the other (x-axis) was its circumference (d). To account for a representative cell surface we used L = 2 m, d = 0.5 μm. To account for a representative outer membrane extension, we used L = 1m, d = 100 nm. The size of a single lattice site (i.e., a possible redox carrier position) was estimated as 2 , i.e. a square based on the average diameter of a redox carrier. The total number of sites in the lattice was chosen by dividing its surface area (Ld) by the surface area of a single lattice site ( 2 ). At the beginning of each simulation, this 2D array was populated by redox car- riers that were randomly distributed on the surface at a given fractional loading X. The reduced/oxidized state of the redox carriers was initially set up to give a linear gradient of reduced redox carriers along the main axis (y-axis) to decrease the simulation time needed to reach steady state. In each simulation time step, the redox carriers were free to diuse randomly in any direction (left, right, up, or 85 down), from a current occupied position to an adjacent unoccupied position. Sim- ilarly, electrons could hop in any direction (left, right, up, down), from a reduced redox carrier to an adjacent oxidized redox carrier. The left and right side edges of the lattice were connected through periodic boundary conditions to allow unre- stricted movement along the x-axis (i.e., circumference). In each time step, redox carriers were reduced at one end of the main axis (y-axis) and oxidized at the other end. The number of oxidation events (i.e., number of electrons transferred) at the end of the lattice was counted over time. This number (i.e. the net number of electrons that crossed the length of the lattice per unit time) is dened as the overall electron transport rate (I ) and is dependent on the width of the given 2D lattice. Using Eqn. 4.4, we can then normalize I by the width of the 2D lattice (i.e., the circumference of the cell or membrane extension) to nd the electron ux (J): J = I d (4.4) If desired, J can then be used to nd D ap using Fick's rst law of diusion: J =D ap @C @x D ap C x=0 C x=L L =D ap C 0 L (4.5) where C is the concentration of reduced redox carriers, L is the length of the lattice along the main axis (y-axis), and C 0 is the concentration of all redox carriers [25, 111]. To achieve a high enough temporal resolution for accurate simulation of hop- ping and diusion events, the simulation time step must be much smaller than the smallest time constant (t e or t p ). Since in our case t e was the smaller time constant, the time step (t) was chosen to be t = 0.1 × t e in all simulations. 86 Each simulation was run long enough to achieve steady state, leading to a con- stant electron transport rate (I ) (in electrons per second) which was used as the simulation output. The number of steps used in our simulations ranged from 10 6 to 4 × 10 7 . The most import parameter aecting the number of required steps was t p (and thus D phys ), with larger t p values (smaller D phys ) requiring a longer simulation. 4.3 Results and Discussion 4.3.1 Single-particle imaging and tracking reveals mobility of MtrC and OmcA along cell surface and membrane extensions Once the labeling scheme (Figure 4.1B-C) was established in our system (see Chap- ter 3), I proceeded to study cell surface protein dynamics using targeted quantum dot (QD) labeling and single-particle tracking (SPT) [64{66]. QDs were chosen as the uorescent label due to their high signal-to-noise ratio and photostability, which makes them useful for SPT [64, 65]. In addition, this labeling scheme takes advan- tage of the very strong (femtomolar scale) binding anity and very low dissocia- tion rate between biotin and streptavidin, which makes biotin-streptavidin labeling schemes useful for single molecule labeling and other applications [124, 125]. To test the hypothesis that MtrC and OmcA are mobile along the cell surface and to quantify their diusion behavior, I labeled MtrC-AP and OmcA-AP by using in vivo biotinylation and streptavidin-conjugated QDs (Figure 4.1B-C) and imaged their dynamics on the surface of living cells by dual-color time-lapse total internal re ection uorescence (TIRF) microscopy. To visualize the cell outer 87 membrane and membrane extensions, I used FM 1-43FX, a lipid membrane dye. To visualize individual cytochromes, I titrated the concentration of streptavidin- conjugated QDs until it was possible to distinguish individual particles (e.g. 1-5 QDs/cell). I observed that MtrC and OmcA are indeed mobile along the cell surface and outer membrane extensions, and I traced their mobility with SPT (Figure 4.2). Brie y, SPT detects the position of each QD molecule in each frame, and connects these detected positions frame-by-frame (where 1 frame = 40 ms) to build tra- jectories over time (Figure 1.6). In my experiments, the typical QD localization precision was15 nm. Figure 4.2 highlights the work ow of single quantum dot detection and tracking as applied to labeling of OmcA on the S. oneidensis cell surface. Starting with a larger eld of view (Figure 4.2A), the uorescence of local- ized QDs was detected on multiple cells in a population simultaneously to generate trajectories over time scales up to multiple seconds each (40 ms frame acquisition rate). Individual QDs were tracked for much longer (1-6 min, e.g. Figure 4.2D-E), with multiple trajectories punctuated by gaps resulting from the expected blink- ing behavior of single QD molecules [126]. Zooming in on individual cells (Figure 4.2B-C, corresponding to dashed areas in Figure 4.2A) highlights a heterogeneity of trajectory shapes, each of which can be analyzed to classify and quantify the diusive dynamics (see next section). When viewed in the context of the cell surface (Figure 4.2D) and membrane extensions (Figure 4.2E), I observed that QD-labeled MtrC and OmcA can tra- verse a signicant fraction of the underlying membrane surface. In addition, I observed an overlap in trajectories of multiple particles, as illustrated by a mem- brane extension linking two cells in Figure 4.2E. These observations support our 88 Figure 4.2: Imaging and single molecule tracking of quantum-dot labeled OmcA using total internal re ection uorescence (TIRF) microscopy. (A) Snapshot of OmcA-AP trajectories (white) in multiple cells (cyan). Trajecto- ries from1.5 min of time-lapse microscopy (40 ms/frame) were overlaid onto the corresponding mean intensity projection image of cells labeled with lipid membrane dye FM 1-43FX. Trajectories in white dashed boxes are blown up in panels B and C. Scale bar: 2m. (B-C) Some example trajectories from the two cells outlined in panel A. Scale bars: 500 nm. (D-E) Streptavidin-coated QD705 was used to detect exogenously biotinylated OmcA-AP (red). Cell membrane and membrane extensions are labeled by FM 1-43FX (cyan). (D) Trajectories from a single quantum dot labeled OmcA-AP as it moved along the surface of a cell. Here, the quantum dot signal (red) and its trajectories (white) are overlaid with the mean intensity projection image of the cell (cyan). For clarity, only the rst frame of quantum dot signal is shown; trajectories are from the entire video (approx. 86 s, 40 ms/frame). Scale bar: 500 nm. (E) Snapshot of OmcA-AP trajectories over- laid on an outer membrane extension. Trajectories (white) are from6 min (40 ms/frame) of time-lapse microscopy tracing several quantum dot-labeled OmcA- AP (red; mean intensity projection image) on a membrane extension that appears to connect two cells (cyan; mean intensity projection image). Scale bar: 500 nm. 89 proposed collision-exchange mechanism of long-distance electron conduction (Fig- ure 4.1A) [25], where diusive dynamics can bridge gaps between cytochromes and, combined with direct electron hopping, lead to a continuous path for electron transport along the membrane and even across neighboring cells. While this long- distance multi-cell conduction was recently observed by electrochemical gating, and cytochrome diusion was proposed to play a role [21], my measurements provide the rst direct look at these dynamics. Next, I sought to quantitatively analyze the diusion characteristics in order to assess their contribution to biological electron transport over micrometer length scales. 4.3.2 Quantifying the dynamics of MtrC and OmcA along the cell surface and membrane extensions To quantify the observed cytochrome diusion, I performed mean squared displace- ment (MSD) analyses on the trajectories obtained from each microscopy dataset (4 main datasets, corresponding to MtrC or OmcA, on the cell surface or on mem- brane extensions). Methods of tracking and diusion analyses have been described many times for SPT, including previously in [64, 116, 119, 121, 127]. To illustrate this analysis on individual trajectories, Figure 4.3 shows MSD as a function of time for three QD-labelled OmcA trajectories on the cell surface. The shape of a trajectory's MSD curve can be used to classify its mode of diusion, e.g. free (Brownian) diusion (curve has straight upward slope) or conned diusion (curve starts to plateau over time) (Figure 1.6). These single trajectory MSDs gen- erally exhibited conned diusion to various degrees, likely re ecting the crowded nature of the bacterial outer membrane and consistent with some observations of connement for outer membrane proteins in E. coli [114, 115]. From a qualitative perspective, among the heterogeneity of trajectory shapes obtained for diusion 90 91 along the cell surface, a majority of trajectories appeared to be more similar to Figure 4.3A (i.e., particle roughly conned to a certain 2D area), while some tra- jectories appeared to be similar to Figure 4.3C (i.e., particle diusing more freely across larger 2D area), and some trajectories appeared to transition between two diusion modes as in Figure 4.3B. For illustration purposes, these individual MSD curves (Figure 4.3) were tted with a model for conned diusion to extract a diusion coecient (D) and con- nement radius (R); the three pictured examples display a range of slow to fast diusion coecients with small to large connement radii. For a simple assessment of dynamics across many trajectories, individual trajectories can also be dened by an instantaneous diusion coecient, which is calculated by tting a free dif- fusion model to the rst 3 time lags (i.e., the rst 3 points) of its MSD curve, where all curves are fairly linear. An individual MSD curve was built for each of the thousands of trajectories in each dataset. Their instantaneous diusion coef- cients were then assembled into histograms (e.g. Figure 4.4) to look for distinct populations (i.e. clearly faster or slower groups) of trajectories that might exist within each dataset, which might appear as multiple peaks in the histogram. For both MtrC and OmcA diusion on the cell surface, the distribution of instanta- neous diusion coecients from the thousands of individual trajectories appeared Figure 4.3 (preceding page): Individual mean squared displacement (MSD) analyses for 3 example trajectories. Example trajectories and corresponding individual MSD curves show a range of diusion by OmcA-AP on cell surface, in order of generally slower and more conned (smaller diusion coecients D, connement radii R) to faster and less conned (bigger D, R). The duration of each trajectory is also labeled above each trajectory image. Error bars in plots show ± SEM. Red curve is the conned diusion t (Eqn. 4.2), from which corresponding D and R values were calculated. Blue line is the free (Brownian) diusion model (Eqn. 4.1) t to the rst 3t in each trajectory's MSD, from which a corresponding D value was calculated. 92 to be reasonably well represented by one Gaussian curve (Figure 4.4); this suggests that both MtrC and OmcA demonstrate one major diusing population on the cell surface, with a variance including the range of observed diusion coecients from individual trajectories (e.g. Figure 4.3). Figure 4.4: Distribution of diusion coecients from individual MSD analyses for MtrC-AP (blue) and OmcA-AP (orange) diusing on the cell surface. To generate these histograms, instantaneous D values were calcu- lated for all individual trajectories by tting the free diusion model (Eqn. 4.1) to the rst 3t in each trajectory's MSD, as done for the blue t lines in Figure 4.3. For illustration, the trajectories from Figure 4.3A-C correspond to the histogram positions labeled by blue arrows. Both histograms show 1 major Gaussian distri- bution for both MtrC and OmcA. These histograms represent an MtrC-AP cell surface dataset containing 5,110 trajectories and OmcA-AP cell surface dataset containing 6,155 trajectories, as observed in 500-1,000 cells each. 93 This motivated us to evaluate the overall diusing behavior of MtrC or OmcA on the cell surface, by constructing ensemble MSD curves that pool data from all trajectories in each dataset (Figure 4.5). I found that overall, both MtrC and OmcA exhibit conned diusion behavior, as the ensemble MSD curves start to plateau within 10 time lags (0.4 s timescale). Fitting these curves with a conned Figure 4.5: Ensemble mean squared displacement (MSD) analysis shows overall conned diusion behavior by MtrC (blue) and OmcA (orange) on the cell surface. Y-axis shows mean displacement squared (r 2 ) for each time lag (t) on the X-axis. Fitting the plots with a conned diusion model (Eqn. 4.2) yields diusion coecients D and connement radii R as labeled. These curves represent an MtrC-AP cell surface dataset containing 5,110 trajectories and OmcA- AP cell surface dataset containing 6,155 trajectories, as observed in 500-1,000 cells each. Error bars show r 2 p N , where N is the number of displacements measured for a given t. 94 diusion model (Methods section 4.2.4, Eqn. 4.2) yields the overall diusion coef- cient and connement radius for QD-labeled MtrC (D = 0.0306± 0.0082m 2 /s, R = 76 nm) and OmcA (D = 0.0121 ± 0.0019 m 2 /s, R = 56 nm) on the cell surface. While quantitative information regarding the diusion of prokaryotic cell surface proteins is limited, my measurements of MtrC and OmcA are consistent with in magnitude with the few observations of other bacterial outer membrane proteins, which are on the scale of D = 0.006{0.15 m 2 /s and R = 30{600 nm [66, 114, 115, 128]. Furthermore, the MtrC and OmcA connement radii are consistent with our previous measurements of center-to-center distances between putative cell surface cytochromes on the surface of S. oneidensis [25]. Note that the connement radius denes a general region where a single cytochrome does not typically travel out of. However, multiple cytochromes (and other proteins) can still exist, diuse, and collide within the same region; occasionally, they can also escape an area of connement and diuse more freely across a larger distance of the cell surface over time, before encountering other obstacles. I also examined the possibility that a minor population of trajectories may be masked by the wide distribution of instantaneous diusion coecients for MtrC and OmcA on the cell surface (Figures 4.3 and 4.4). To investigate this possibil- ity, I performed probability distribution of square displacements (PDSD) analysis to sort trajectories into dierent populations as described by [119{121] and build respective ensemble MSD curves for each population. In this case, the PDSD was better t by a 2-component rather than 1-component model, suggesting two dius- ing sub-populations for both MtrC and OmcA on the cell surface (Figure 4.6). As expected by their overall diusion behavior (Figure 4.5), all sub-populations exhib- ited conned diusion (Figure 4.6). For MtrC, I found that a majority (90%) of trajectories exhibited slower, more conned diusion (D 1 = 0.00166 ± 0.00182 95 Figure 4.6: Diusion analyses for MtrC (blue) and OmcA (orange) diu- sion on cell surface assuming 2 major populations of conned diusion behavior. These plots suggest a larger, slower, more conned population and a smaller, faster, less conned population of diusion by each protein on the cell surface. The ensemble MSD curves were plotted as mean displacement squared r 2 as a function of time lag t for respective major populations of diusing behavior. Fitting these curves with a conned diusion model (Eqn. 4.2) yields diusion coecients D and connement radii R as labeled on the respective plots. Per- centages indicate the respective fractions belonging to each population determined by PDSD analysis, as described in [119, 121]. These curves represent an MtrC- AP cell surface dataset containing 5,110 trajectories and OmcA-AP cell surface dataset containing 6,155 trajectories, as observed in 500-1,000 cells each. Error bars show SEM p N(%N) , where %N is the fraction of all displacements N that had been allocated to the given sub-population for a given t. 96 m 2 /s, R 1 = 17 nm) and a minority (10%) exhibited faster, less conned dif- fusion (D 2 = 0.141 ± 0.018 m 2 /s, R 2 = 231 nm). Similarly, OmcA exhibited similar diusive sub-populations: slow (90%, D 1 = 0.00197 ± 0.00149 m 2 /s, R 1 = 19 nm) and fast (10%, D 2 = 0.0962 ± 0.0108 m 2 /s, R 2 = 210 nm). The faster, less conned minority of trajectories detected by this analysis may repre- sent events where generally conned redox proteins occasionally escape a more crowded area and are then able to diuse more freely across a larger area. This is also consistent with the heterogeneity in distribution of proteins along the cell surface previously observed by electron cryotomography [25], which is a common feature among membrane proteins in bacteria [114, 129]. Together, these results provide a closer look at the extent of diusion that is generally possible by MtrC and OmcA on the cell surface. Next, I examined the diusion of MtrC and OmcA on the S. oneidensis outer membrane extensions (OMEs). Compared to cell surface measurements, imaging of QD-labeled cytochromes on OMEs presented signicant technical challenges. Our previous perfusion ow imaging platform [25] allowed robust epi uorescence observations of OME production over time by restricting OMEs to the focal plane using laminar media ow, but this was not possible in my TIRF imaging setup, which consists of a simple glass-bottomed cylindrical container. Note that even perfusion ow imaging does not restrict movement in X or Y directions within the focal plane, which is just above and parallel to the glass surface of the microscopy coverslip; the restriction of vertical movement along the Z axis, perpendicular to focal plane, simply helps to keep OMEs visible throughout the imaging period. In our TIRF imaging setup, OMEs can frequently move in all X, Y, and Z directions. Tracking QDs becomes much more dicult if they are on on a structure that is also moving, so I was also limited to analyzing OMEs that did not move during 97 the imaging process. Furthermore, I only imaged OMEs that were labeled by an optimal number of QDs for tracking (i.e., with at least 1 or up to only a few distinct QDs), were clearly connected to a cell (i.e., a \real" OME, not a structure that broke o or other unknown uorescent object), and were not crowded by nearby uorescent objects (i.e. other cells, which is problematic both from a membrane uorescence standpoint if you cannot resolve the OME clearly, as well as QD tracking standpoint if other QDs on nearby cells are too close to the path of a QD on an OME of interest). Compared to the large numbers of whole cells that I was able to use for tracking analyses, I was limited in the number of QD-labeled OMEs that were optimal for tracking, so instead I recorded longer videos of optimally labeled OMEs (6-min video per OME, instead of1.5-min video per eld of whole cells) which allowed us to obtain a reasonable number of trajectories for tracking analyses. Interestingly, the distribution of instantaneous diusion coecients from all individual trajectories obtained from OMEs (Figure 4.7) reveals a noticeable right- side tail, re ecting two sub-populations and their corresponding Gaussian peaks for both MtrC and OmcA. Incidentally, for both proteins, the rst peak (Figure 4.7) occurs at roughly the same position as the main (only) peak visible for cell surface diusion (Figure 4.4), while the second peak is right-shifted (i.e., faster) in com- parison. In light of the two detected peaks, I then performed PDSD and ensemble MSD analyses to investigate diusion behavior assuming 2 major populations for MtrC (Figure 4.8A-B) and likewise for OmcA (Figure 4.8C-D). Here, the curves for OmcA (Figure 4.8C-D) appear smoother than those for MtrC (Figure 4.8A-B) most likely due to the greater statistical power (more OMEs imaged, more tra- jectories obtained). PDSD analysis also indicates that the faster regimes (Figure 98 Figure 4.7: Distribution of individual diusion coecients for MtrC-AP (blue) and OmcA-AP (orange) diusing on the surface of outer mem- brane extensions (OMEs). Individual diusion coecients were calculated by tting a free diusion model (Eqn. 4.1) to the rst 3t in each trajectory's mean squared displacement (MSD). These distributions, unlike those for cell surface datasets, now contain a noticeable right-side tail suggesting more than 1 Gaussian population. Thus, each plot here also shows t lines for 2 Gaussian distribu- tions (blue or orange dashed lines) and their cumulative t (black line). These histograms represent two datasets, MtrC-AP on OMEs (945 trajectories from 5 OMEs) and OmcA-AP on OMEs (4,633 trajectories from 22 OMEs). 4.8B,D) now comprise a somewhat larger fraction (20-30% instead of10%) of tra- jectories, which is consistent with the more noticeable second (faster) peak visible in the distributions of diusion coecients (Figure 4.7). Fitting these ensemble MSD curves with a conned diusion model reveals estimates of diusion coef- cient and connement radius for slow (70%) and fast (30%) populations of trajectories for MtrC (D 1 = 0.00175± 0.00012m 2 /s, R 1 = 51 nm; D 2 = 0.03274 ± 0.00202 m 2 /s, R 2 = 194 nm) and similarly for slow (80%) and fast (20%) populations of OmcA (D 1 = 0.00149± 0.00004m 2 /s, R 1 = 47 nm; D 2 = 0.02744 ± 0.00088 m 2 /s, R 2 = 254 nm). Taken together, these analyses (Figure 4.7 and 4.8) suggest that on OMEs, the cytochromes display 2 populations of independently diusing behavior. However, 99 Figure 4.8: Ensemble Mean Squared Displacement (MSD) analyses for MtrC and OmcA on outer membrane extensions (OMEs) demonstrate 2 major populations of conned diusion behavior. Left (A-B): MtrC-AP. Right (C-D): OmcA-AP. (A-D) Ensemble MSD curves were plotted as mean dis- placement squared r 2 as a function of time lag t for respective major populations of diusing behavior: (A) MtrC-AP, slow population; (B) MtrC-AP, fast pop- ulation; (C) OmcA-AP, slow population; and (D) OmcA-AP, fast population. Fitting these curves with a conned diusion model (Eqn. 4.2) yields diusion coecients D and connement radii R as labeled on the respective plots. Per- centages indicate the respective fractions belonging to each population determined by PDSD analysis, as described in [119, 121]. The ensemble MSD plots suggest a larger, slower, more conned population and a smaller, faster, less conned pop- ulation of diusion by each protein on outer membrane extensions. Error bars show SEM p N(%N) , where %N is the fraction of all displacements N that had been allocated to the given sub-population for a given t. These ensemble MSD curves represent two datasets, MtrC-AP on OMEs (945 trajectories from 5 OMEs) and OmcA-AP on OMEs (4,633 trajectories from 22 OMEs). 100 it is also possible that there is 1 major population of particles (obtained by the individual MSD analyses, represented by histograms as in Figure 4.7) that transi- tions between two diusing modes, thus displaying 2 dierent diusing behaviors (obtained by PDSD and ensemble MSD analyses as in Figure 4.8). To account for the blinking behavior of individual QD molecules [126], my tracking parameters allowed for up to 10-frame gaps in trajectories (i.e., particle disappearance for up to 10 frames, or 0.4 s) before \chopping" the trajectory and waiting for the particle to reappear before starting to record a new trajectory. Keeping this gap allowance to a few frames is helpful especially when multiple particles might be detected nearby, in order to minimize false connections between particle positions when building trajectories. As a side eect of having a relatively short gap allowance, the motion of a particle over a longer period of time, including particle diusion which sometimes switches between two diusing modes (e.g. Figure 4.3B), may be \chopped" into multiple trajectories where some show conned diusion behavior (e.g. Figure 4.3A) while others behave more like free diusion (e.g. Figure 4.3C). This would aect the number of peaks detected in the histogram distribution of diusion coecients (Figure 4.7). Further analyses (e.g. by changing tracking parameters to allow longer gaps of particle disappearance) may reveal if this is the case in our system. I also analyzed ensemble MSD curves for MtrC and OmcA assuming 1 major population on OMEs (Figure 4.9), in order to make a general comparison of the overall diusion behavior on OMEs relative to the cell surface (Figure 4.5). On OMEs, I determined an overall diusion coecient and connement radius for MtrC (D = 0.00895 ± 0.00029 m 2 /s, R = 122 nm) and OmcA (D = 0.00678 ± 0.00019m 2 /s, R = 113 nm) (Figure 4.9). Overall, ensemble MSD curves for diu- sion on OMEs plateau over a greater timescale (<2 s) than for cell surface diusion 101 Figure 4.9: Ensemble mean squared displacement (MSD) analysis shows overall conned diusion behavior by MtrC (blue) and OmcA (orange) on outer membrane extensions (OMEs). Y-axis shows mean displacement squared (r 2 ) for each time lag (t) on the X-axis. Fitting the plots with a conned diusion model (Eqn. 4.2) yields diusion coecients D and connement radii R as labeled. Error bars show r 2 p N , where N is the number of displacements measured for a given t. These curves represents two datasets, MtrC-AP on OMEs (945 trajectories from 5 OMEs) and OmcA-AP on OMEs (4,633 trajectories from 22 OMEs). (<0.4 s), which leads to overall larger connement radii, suggesting that MtrC and OmcA exhibit less conned diusion on OMEs than on cell surface. At the same time, it appears that MtrC and OmcA are moderately slower on OMEs than on the cell surface. The underlying reasons for these dierences remain unclear. However, it is important to note that the OME and cell surface environments may dier in degree of molecular crowding, which may translate to less connement on OMEs. 102 The slower cytochrome dynamics on OMEs may be related to their structure, as the 3D curvature is greater in OMEs compared to the cell surface. Diusion may be slower on OMEs due to curvature along the length of a membrane extension, since rather than consisting of smooth membrane tubes, OMEs often appear more like a vesicle chain with possible junction densities that might limit membrane uidity at each junction. These junction densities were observed in our previous cryotomography work [25] and are likely prokaryotic membrane curvature compo- nents known in eukaryotes as BAR domain proteins [130]. In addition, the motion of cytochromes on curved membranes (both cell surfaces and OMEs) is inherently in 3D, while the imaging and MSD analyses are planar, i.e. they capture the projection of the 3D motion onto a 2D plane. This leads to an underestimate of the diusion coecient since the analysis does not detect the small displacements normal to the image plane, but the diusion coecient can be corrected through a 3D simulation that maps the planar MSD to the true 3D MSD [131]. Renner et al. performed these simulations systematically for lateral diusion on tubular membranes, revealing that the measured 2D diusion coecient was 25% to 50% smaller than the true 3D MSD, with smaller tube diameters (i.e., greater curva- ture) leading to greater underestimations of diusion [132]. Such an underestimate therefore contributes to the moderately smaller observed D values observed on S. oneidensis OMEs relative to the cell surface. In addition to the 2D vs. 3D geometric correction described above (which would scale the diusion coecient up by two fold) [131, 132], other factors that may contribute to making my estimates of MtrC and OmcA diusion coecients lower bounds of their true diusion. For example, the large size of QDs (2030 nm in diameter) may slow the intrinsic dynamics of the targeted biomolecules [65]. Estimates of this eect have been con icting, ranging from no signicant eect 103 [133] to a study demonstrating that antibodies traced by a biotin-QD probe had 35-fold slower diusion compared to when they were conjugated to a small organic uorophore Cy3 [134]. To account for these possible factors, our simulations of the contribution of cytochrome diusion to long distance electron transport (described below) explored both the measured in vivo diusion coecients (10 -2 m 2 /s, from Figures 4.5 and 4.9) and one order of magnitude higher (10 -1 m 2 /s). The latter value is also supported by ex vivo measurements of the MtrCAB transmembrane conduit on lipid bilayers, measured via uorescence recovery after photobleaching (FRAP) to be approximately D = 10 -1 m 2 /s [135]. A summary of overall diusion measurements from this study is presented in Tables 4.2 and 4.3. These tables list the measurements reported in Figures 4.5, 4.6, 4.8, and 4.9. Table 4.2: Summary of overall diusion measurements. Diusion coe- cients D and connement radii R listed here were obtained by tting the conned diusion model (Eqn. 4.2) to ensemble mean squared displacement (MSD) curves for the listed datasets, corresponding to Figures 4.5 and 4.9. All diusion coe- cients are reported as± standard deviation of the t value. OME: outer membrane extension. Surface D (m 2 /s) R (nm) MtrC-AP Cell 0.0306 ± 0.0082 76 OME 0.00895 ± 0.00029 122 OmcA-AP Cell 0.0121 ± 0.0019 56 OME 0.00678 ± 0.00019 113 104 Table 4.3: Summary of diusion measurements, assuming 2 popula- tions. Percentages indicate the respective fractions belonging to each population determined by probability distribution of square displacement (PDSD) analysis in SLIMfast. Diusion coecients D and connement radii R described here were obtained by tting the conned diusion model (Eqn. 4.2) to ensemble mean squared displacement (MSD) curves for the listed datasets, corresponding to Fig- ures 4.6 and 4.8. All diusion coecients are reported as ± standard deviation of the t value. OME: outer membrane extension. Surface Population D (m 2 /s) R (nm) MtrC-AP Cell 90% 0.00166 ± 0.00182 17 10% 0.141 ± 0.018 231 OME 70% 0.00175 ± 0.00012 51 30% 0.03274 ± 0.00202 194 OmcA-AP Cell 90% 0.00197 ± 0.00149 19 10% 0.0962 ± 0.0108 210 OME 80% 0.00149 ± 0.00004 47 20% 0.02744 ± 0.00088 254 Altogether, MtrC and OmcA also appear to diuse in a similar fashion as each other, whether on the cell surface (Figures 4.4, 4.5, and 4.6) or on membrane extensions (Figures 4.7, 4.8 and 4.9). This is not surprising, since overall they are structurally and functionally homologous [136]. Generally, MtrC also diuses slightly faster than OmcA, within the limits of my observation. This could be, at least in part, due to the small size dierence between the two proteins: their overall dimensions are approximately 9 nm × 6 nm × 4 nm for MtrC, and 9.5 nm × 6 nm × 5 nm for OmcA [12, 137]. I also tested if individual QD probes might be binding to multiple cytochromes and possibly limiting my measurements of diusion, since a single streptavidin- conjugated QD molecule has multiple biotin-binding sites. This is a common concern when using QDs for SPT [64]. A recent study compared probes with single vs. multiple biotin binding sites and demonstrated that probes binding to multiple targets appear as left-shifted (slower) peaks in a distribution of diusion coecients 105 [138]. To see if this was occurring in my experiments, I performed a biotin ood experiment where samples were ooded with excess biotin soon after addition of QDs, to ll excess biotin-binding sites before imaging. However, a histogram distribution of diusion coecients from this experiment (Figure 4.10) reveals only one major peak in exactly the same position as in my previous experiments of diusion on the cell surface (Figure 4.4), suggesting that this was not an issue in my experiments. 106 Figure 4.10: Distribution of diusion coecients from individual MSD analyses for MtrC-AP (blue) and OmcA-AP (orange) diusing on the cell surface, when cells were ooded by 200-fold excess biotin after addition of streptavidin-coated quantum dots. Diusion coecients were calculated by tting a free diusion model (Eqn. 4.1) to the rst 3t in each trajectory's MSD. Both histograms show 1 major Gaussian distribution for both MtrC and OmcA. Diusion from biotin ood experiments are consistent with pre- vious datasets of diusion on cell surface (Figure 4.4), suggesting that the quantum dot probes are not binding to multiple protein targets and subsequently hinder- ing diusion. These histograms represent an MtrC-AP dataset containing 69,793 trajectories and OmcA-AP dataset containing 58,543 trajectories, as observed in >1,000 cells each. 107 4.3.3 Simulations of overall electron transport along mem- brane surfaces combine direct electron hopping and physical diusion of cytochromes Note: All simulations described in this section were designed and performed by co-author Sahand Pirbadian. We then used my experimental diusion measurements to inform simulations of overall electron transport (ET). Rather than electron transport from inside to outside the cell, these simulations describe long-distance lateral electron transport along membrane surfaces, as depicted in Figure 4.1. In our recent work [25], we used a mean-eld approach described by Blauch and Saveant [111] to estimate overall ET as a function of cytochrome fractional loading on the surface of mem- brane extensions. However, since my dynamics measurements suggest that redox carrier motion is slower than electron hopping (i.e., t e < t p ), this approach is no longer justied, instead requiring simulations that mimic diusion and hop- ping of each redox carrier on a 2D surface [111]. So, following the Blauch-Saveant model [111], we performed Monte Carlo simulations incorporating random electron hopping and random redox carrier diusion, which together contribute to electron transport across a 2D lattice array representing the roughly cylindrical surface of a whole cell or an outer membrane extension. Key input parameters include the time constant of electron hopping (t e ), the time constant of redox carrier diusion (t p , which is obtained from experimental D phys ), the fractional loading of cytochromes on the lattice (X ), and the dimensions of the 2D lattice itself. The simulation output is the overall ET across the 2D surface, which is calculated from the net number of electrons that cross the length of the lattice over a certain period of time during each simulation. 108 Figure 4.11: Simulation results of overall electron transport (ET) along the surface of whole cells or membrane extensions, based on experimen- tally measured diusion coecients. ET rates on the Y-axis are plotted on a log scale as a function of the fractional loading of redox carriers (X ) on (A) the surface of a whole cell (2 m long and 0.5 m in diameter) or (B) the surface of an outer membrane extension (1 m long and 100 nm in diameter). These results come from simulations using either t e = 10 -5 s (lled shapes) or 10 -6 s (unlled shapes) for a range of experimentally derived diusion coecients D phys = 10 -1 m 2 /s (black squares) or 10 -2 m 2 /s (red circles). Note: All simulations repre- sented in this gure were designed and performed by co-author Sahand Pirbadian. In Figure 4.11, we report our simulation results as overall ET rate as a function of fractional loading for whole cells and membrane extensions. We simulated four scenarios comparing input t e = 10 -5 or 10 -6 s and D phys = 10 -1 or 10 -2 m 2 /s, where t e is estimated by the electron residence time in OM decaheme cytochromes to be approximately 10 -5 to 10 -6 s [109, 122, 123], and D phys was based on my diusion 109 measurements of MtrC and OmcA in this work. We observe a positive correlation between ET rate and fractional loading, due to the greater contribution of direct electron hopping to electron transport (to be precise, t e < t p ). At higher fractional loading, the impact of direct electron hopping is greater, since there is less room for physical diusion. Meanwhile, at lower fractional loading, the contribution of physical diusion becomes more apparent, as direct electron hopping events are less frequent in a landscape with sparsely distributed cytochromes. Previous work estimating the number of cytochromes per S. oneidensis cell via calibrated western blot analysis [110] suggests that we are most likely in the upper range of X (where X > 0.5, if not nearing the limit of X = 1), and based on our simulations (Figure 4.11) an overall lateral ET rate around 10 4 to 10 5 e - /s is possible for electron transport on whole cells and membrane extension surfaces, depending on the exact choice of t e . This estimate is also consistent with our conduction measurements through cells bridging interdigitated electrodes [21]. In [21], the measured electrochemical gating current was around 1 nA or 6 x 10 9 e - /s; the length of the electrode in the interdig- itated electrode array (IDA) was approximately 2.6 x 10 5 m (130 electrodes, each 2 mm long). Here, the overall current is approximated by the electrode edge length multiplied by the number of connections (i.e., cells bridging electrodes) per micron of electrode edge, multiplied by the ET rate per connection. Assuming 1 connec- tion per 10m of electrode edge (which might be an underestimate), the observed ET rate per connection (assuming 1 connection = 1 cell bridging electrodes) was approximately 2.4 x 10 5 e - /s, which is close to the maximum rate (3 x 10 5 e - /s) from these simulations. We note two possible underestimations, namely (1) since the gap between electrodes in the experiment shown in [21] was 5m, it is possible that this gap is bridged by 23 connected cells (of length2m) rather than just 110 1 cell per connection, and (2) it is possible that connections (where cells bridge electrode gaps) occurred somewhat more frequently than 1 connection per 10 m of electrode edge. Together, these possible underestimations likely contribute less than an order of magnitude uncertainty, giving a lower range of ET rates per cell of 10 4 e - /s for the redox conduction measured in [21]. Based on these assumptions, we estimate an ET rate along the length of a whole cell in these electrochemical experiments [21] to be on the scale of 10 4 to 10 5 e - /s, which is consistent with the ET rates from our simulations. In the study of outer membrane extensions in S. oneidensis, another question of interest is whether these proposed bacterial nanowires can support cellular ener- getics, e.g. by assisting with respiration of distant electron acceptors. It has been shown that S. oneidensis oxidizes lactate (a common electron donor and carbon source in studies of S. oneidensis) to acetate, producing 1 ATP molecule and releasing 4 electrons in the process [139, 140]. Previously, the non-growth ATP requirement (i.e., basic ATP requirement for life) for S. oneidensis cells was esti- mated to be about 1.03 mM of ATP per 1 g of ash free dry weight (AFDW) per hour [141]. Assuming an ash weight of 7.5% (ash weight approx. 7.5% for Gram- negative bacterial cells grown in shaken liquid cultures, and typically 5-10% for various species and cell types, according to [142]), we can estimate 1 g dry weight = 0.925 g AFDW. Assuming the mass of a single S. oneidensis cell is similar to E. coli (a well-studied Gram-negative bacterium of similar cell length and diameter), we can use a dry weight to wet weight conversion of 23% (for E. coli) [143] and the mass of a live bacterial cell (i.e., wet weight) to be about 1 x 10 -12 g (for E. coli) [144] to convert 1 g of AFDW 1.08 g dry weight 4.7 x 10 12 cells. Then, using Avogadro's number, we can convert the non-growth (i.e., basic) ATP requirement 111 [141] to 3.7 x 10 4 ATP s -1 cell -1 . Note: This estimate is based on ATP require- ments in laboratory conditions [141] and is possibly an overestimate, since basic cell maintenance in the environment may actually require much fewer ATP. From this study, an electron transport rate along the length of a membrane extension at 10 4 to 10 5 e - /s could thus theoretically support the production of approximately 2,500-25,000 ATP s -1 cell -1 . Thus, based on the estimations above, and assuming only a single point of contact (i.e., 1 porin-cytochrome conduit) with an external electron acceptor, then electron transport along the length of a single nanowire may be sucient to support basic cellular respiration (10 4 ATP s -1 cell -1 ) at the upper range of electron transport rates (10 5 e - /s, or10 4 ATP s -1 cell -1 ). If the electron transport rate is lower (10 4 e - /s, or10 3 ATP s -1 cell -1 ), then an OME with only a single point of contact (i.e., 1 cytochrome) to electron acceptor is likely able to assist with cellular energy requirements, rather than being solely sucient. In both cases, the expected contribution to cellular respiration (in the form of ATP production) may be higher if multiple cytochromes on an OME are in contact with the external electron acceptor. However, to address this question properly will require experiments that measure cell energetics in cells that make OMEs that do or do not contact external electron acceptors, which is currently in progress in the El-Naggar Laboratory. In addition, the study of S. oneidensis outer membrane extensions calls into question how the shape of these structures helps with electron transport (and thus potentially helping with cellular energy requirements). Brie y, the capability of producing such structures (1) allows the cell to extend its cytochrome network by the length of the structure and (2) increases the surface area to volume ratio of the cell, thus increasing its exposure and access to external electron acceptors in the surrounding environment. Furthermore, the most commonly observed vesicle 112 chain morphology is perhaps more advantageous than a simple smooth tubular shape, since the vesicle chain morphology has a greater surface area to volume ratio than a simple smooth tube, increasing access to electron acceptors. 4.4 Conclusion In summary, my experiments in this work demonstrate for the rst time that MtrC and OmcA are mobile along both whole cells and outer membrane extensions in S. oneidensis MR-1. These observations support a collision-exchange model of long-distance electron transport along membrane surfaces, which includes a com- bination of direct electron hopping between adjacent cytochromes and physical diusion that bridges gaps between non-adjacent cytochromes (Figure 4.1). I mea- sured this diusion using single-particle tracking techniques, and found generally restricted diusion behavior by MtrC and OmcA both on the cell surface (overall, D MtrC = 0.0306 ± 0.0082 m 2 /s, R MtrC = 76 nm; and D OmcA = 0.0121 ± 0.0019 m 2 /s, R OmcA = 56 nm) and similarly for on the surface of membrane extensions. These measurements are consistent in magnitude with the diusion of other bac- terial outer membrane proteins, which are on the scale of D = 0.006{0.15 m 2 /s and R = 30{600 nm [66, 114, 115, 128]. To my knowledge, MtrC and OmcA in this study are the rst outer membrane proteins (and EET components) whose diusion was measured in S. oneidensis, and two of few cell surface proteins whose diusion was measured in prokaryotes. Finally, we applied my measurements to simulations of overall ET rate along membrane surfaces, and found that our results are consistent in magnitude with recent measurements of long-distance redox con- duction in S. oneidensis. The techniques used in this work can also be used to study the dynamics of other bacterial surface proteins. In addition, this work 113 motivates future work which the El-Naggar Lab is pursuing, including dynamics measurements of periplasmic components in the S. oneidensis EET network, since their diusion is also likely to contribute to long-distance lateral electron transport parallel to and along the surface of the cell. In the future, it would also be useful to perform further electrochemical experiments which allow us to precisely determine the ET rate along the length of a single cell, which can then be compared to the ET rates from our simulations; these experiments are actively underway in the El-Naggar Lab. 114 Chapter 5 Conclusion To summarize, the main goal of this thesis was to learn more about how biological electron transport works in Shewanella oneidensis MR-1, a bacterium that can breathe rocks, conduct electricity, and make bacterial nanowires. It is now well known that this bacterium harnesses a network of multiheme cytochromes (elec- tron transfer proteins) for the respiration of external electron acceptors, includ- ing minerals and electrodes outside the cell, in a process known as extracellular electron transfer (EET). These cytochromes are also important for long-distance (micrometer-scale) electron transport (ET) along living cells [21]. As they are found along S. oneidensis' micrometer-scale extensions of the cell's outer mem- brane and periplasmic space [23{25], these cytochromes are also thought to enable their proposed function as bacterial nanowires for the respiration of distant electron acceptors. In Chapter 2, I used various imaging techniques to learn more about the pro- posed bacterial nanowires in S. oneidensis. First, I investigated why these exten- sions form. Since their discovery [22], they were rst believed to arise from condi- tions of electron acceptor limitation, particularly oxygen limitation, so subsequent studies focused on the display of these extensions in oxygen-limiting conditions. However, it was unclear to what extent oxygen limitation is responsible for their formation, nor was it known if other conditions contribute to this phenomenon. In this chapter, I designed in vivo time-lapse uorescence microscopy experiments to evaluate the production of outer membrane extensions (OMEs) in a number 115 of physical conditions (including oxygen limitation). I found that cell-to-surface contact is sucient to induce the production of OMEs, including some that reach up to 100m in length (at an elongation rate of 40m/s), irrespective of medium composition, agitation, or aeration. Incidentally, while I demonstrated that oxy- gen limitation is not sucient to induce formation of OMEs, it remains unclear if it plays a supplementary role e.g. in the frequency of OME production (which I was unable to precisely quantify in the context of oxygen abundance). As a bonus, in this chapter I also presented the rst precise quantication of cells producing OMEs (78%) visible over time in perfusion ow culture conditions. In Chapter 2, I also examined the degree of cytochrome-dependent redox activity in S. oneidensis membrane extensions. Here, I mapped the activity of redox centers (e.g. cytochromes) on OMEs in wild type and cytochrome-decient strains, using a combination of heme-dependent staining and transmission electron microscopy. I found that 8 known periplasmic and outer membrane cytochromes are responsible for most of the redox activity detected using this assay: 2.4-fold more OMEs were stained in wild type than in a cytochrome-decient strain lacking those 8 cytochromes; and comparing only OMEs that were stained for heme, wild type OMEs displayed 3.6-fold more staining intensity. In the process, I also probed 3 types of extracellular laments (OMEs, agella, and pili) for these EET com- ponents, and demonstrated that they are limited to OMEs and are not found on agella or pili. As a bonus, these staining and microscopy methods dene a max- imum center-to-center distance (<50 nm) between redox partners on the surface of membrane extensions, which is consistent with recent estimates of cytochrome distribution along OMEs as determined by electron cryotomography [25]. Alto- gether, the work presented in Chapter 2 provides a greater understanding of how bacterial nanowires work, what they look like, and why they form. 116 In Chapters 3-4, I investigated how long-distance electron transport works along membrane surfaces in S. oneidensis. Recent work demonstrated long-distance (micrometer-scale) redox conduction along and between living cells due to the cytochromes also important in EET [21] and distributed along membrane surfaces in S. oneidensis with a heterogeneity from immediately adjacent to up to tens of nanometers apart [25]. Based on this knowledge, we previously proposed that this occurs by a combination of electron hopping and cytochrome diusion, which allows collisions and electron exchange between cytochromes along membranes. However, the dynamics of multiheme cytochromes had never been observed or quantied in vivo, making it dicult to evaluate their hypothesized contribution to the proposed mechanism of long-distance electron transport. In Chapter 3, I developed a labeling scheme in our system that would allow site- specic labeling of individual cytochromes in S. oneidensis, based on the scheme described in [100]. Here, I tagged cell surface cytochromes MtrC and OmcA with a C-terminal biotin acceptor peptide (AP) tag [99] in their respective gene deletion backgrounds. This AP tag could be biotinylated in vivo and subsequently detected by streptavidin conjugates. In labeling controls in western blot and microscopy, I systematically omitted key labeling components and demonstrated the success and specicity of the labeling scheme in our system. As a bonus, my ability to perform in vivo extracellular labeling of MtrC and OmcA via a C-terminal tag also supported recent structural ndings that demonstrated the orientation of MtrC relative to the cell surface [15]. Then, in Chapter 4, I used the labeling scheme established in our system in Chapter 3 to address the original goal of investigating how long-distance (micrometer-scale) electron transport works in S. oneidensis. Using targeted quan- tum dot labeling and single molecule imaging, I visualized individual cytochromes 117 MtrC and OmcA (tagged in Chapter 3), two key components of the S. oneidensis EET network, and observed their mobility along the surface of living cells and their membrane extensions. I then used single particle tracking and diusion anal- yses to describe and quantify this mobility. I observed conned diusion behavior for both quantum dot-labeled MtrC and OmcA along cell surfaces (overall dif- fusion coecients D MtrC = 0.0306 ± 0.0082 m 2 /s, D OmcA = 0.0121 ± 0.0019 m 2 /s) and the membrane extensions thought to function as bacterial nanowires. I found that these dynamics can trace a path for electron transport via overlap of cytochrome trajectories along cells and membrane extensions, consistent with our proposed long-distance conduction mechanism. In addition, my experimental mea- surements now made it possible for us to perform kinetic Monte Carlo simulations of long-distance electron transport along the surface of whole cells and membrane extensions. Overall, in this chapter, I described the rst dynamics measurements of bacterial cell surface cytochromes, and found that this mobility supports our hypothesis of how long-distance electron transport works along membrane surfaces in S. oneidensis. In light of the ndings described in this thesis, several questions remain to frame future studies, particularly relating to biological electron transport and the proposed bacterial nanowires in S. oneidensis: Why do OMEs form? In Chapter 2, I discovered the rst physical condition known to be sucient to induce OME formation, but future work will be required to investigate the precise role of oxygen limitation in OME production, and to determine if anything else (e.g. specic genes/cellular components) in uences their formation. Another related question is whether cells make new lipids during OME production (though at a quick glance this does not appear to be the case, as cells 118 in Chapter 2 appear to shrink in size consistent with the estimated surface area of the newly formed membrane extension). What is the physiological impact of OMEs? Though I demonstrated in Chapter 4 that cytochrome mobility traces a path for long-distance electron transport on the surface of cells and OMEs, it is still unclear if these membrane extensions actu- ally full their proposed role as bacterial nanowires for the respiration of distant electron acceptors (e.g. minerals, electrodes). Experiments that directly evaluate this possibility are required. The El-Naggar Lab recently combined simultaneous in vivo uorescence microscopy and electrochemical experiments to show that a uorescent membrane potential indicator can also be used as a direct reporter of EET in S. oneidensis [145], which will be useful in addressing this question. I have also started developing and troubleshooting experiments using this and other indicators to measure changes in cell energetics as cells form OMEs that do or do not contact distant minerals or electrodes, to be continued in the El-Naggar Lab. How does long-distance electron transport work along bacterial membrane sur- faces? The work presented in Chapter 4 motivates dynamics studies of other cytochromes in S. oneidensis, or cytochrome dynamics in other microbes capable of EET, or studies of diusion by bacterial cell surface proteins in general, which are extremely lacking in Gram-negative bacteria [115]. As a follow up to this work, the El-Naggar Lab is currently investigating the dynamics of periplasmic cytochromes which likely also contribute to long-distance electron transport. It also may be interesting to follow up on the dynamics measurements of MtrC and OmcA described in this thesis, which likely represent a lower bound of their true diusion, e.g. to perform measurements with smaller photostable probes, 2D to 3D curvature correction, etc. In addition, it will be useful to evaluate the results 119 from our simulations of long-distance electron transport (performed by my col- laborator using my dynamics measurements) in light of another study currently being done in the El-Naggar Lab, using electrochemical systems to quantify long- distance electron transport rates in living cells. Furthermore, a better understand- ing of the fractional loading of cytochromes along the cell surface (and membrane extensions) may also help to improve our understanding of long-distance electron transport rates possible along membrane surfaces in S. oneidensis. 120 Bibliography [1] Grace W. Chong, Amruta A. Karbelkar, and Mohamed Y. El-Naggar. Nature's Conductors: What can microbial multi-heme cytochromes teach us about electron transport and biological energy conversion? Cur- rent Opinion in Chemical Biology, 47:7{17, 2018. ISSN 18790402. doi: 10.1016/j.cbpa.2018.06.007. [2] Harry B. Gray and Jay R. Winkler. Electron tunneling through proteins. Quarterly Reviews of Biophysics, 36(3):341{372, 2003. ISSN 0033-5835. doi: 10.1017/S0033583503003913. URL http://www.journals.cambridge.org/abstract S0033583503003913. [3] Liang Shi, Hailiang Dong, Gemma Reguera, Haluk Beyenal, Anhuai Lu, Juan Liu, Han-Qing Yu, and James K. Fredrickson. Extracellular electron transfer mechanisms between microorganisms and minerals. Nature Reviews Microbiology, 14(10):651{662, oct 2016. ISSN 1740-1526. doi: 10.1038/nrmi- cro.2016.93. URL http://www.nature.com/articles/nrmicro.2016.93. [4] Kenneth H. Nealson. Bioelectricity (electromicrobiology) and sustainabil- ity. Microbial Biotechnology, 10(5):1114{1119, 2017. ISSN 17517915. doi: 10.1111/1751-7915.12834. [5] Uwe Schr oder and Falk Harnisch. Life Electric|Nature as a Blueprint for the Development of Microbial Electrochemical Technologies. Joule, 1(2): 244{252, 2017. ISSN 25424351. doi: 10.1016/j.joule.2017.07.010. [6] Syed Zaghum Abbas, Mohd Rafatullah, Norli Ismail, and Muhammad Izzud- din Syakir. A review on sediment microbial fuel cells as a new source of sustainable energy and heavy metal remediation: mechanisms and future prospective. International Journal of Energy Research, 41(9):1242{1264, 2017. ISSN 1099114X. doi: 10.1002/er.3706. [7] Guangyin Zhen, Xueqin Lu, Gopalakrishnan Kumar, P eter Bakonyi, Kaiqin Xu, and Youcai Zhao. Microbial electrolysis cell platform for simultaneous 121 waste biorenery and clean electrofuels generation: Current situation, chal- lenges and future perspectives. Progress in Energy and Combustion Science, 63:119{145, 2017. ISSN 03601285. doi: 10.1016/j.pecs.2017.07.003. [8] Jampala Annie Modestra, Gokuladoss Velvizhi, Kamaja Vamshi Krishna, Kotakonda Arunasri, Piet N L Lens, Yarlagaddavenkata Nancharaiah, S Venkata Mohan, and Ag Silver. Bioelectrochemical Systems for Heavy Metal Removal and Recovery. In Sustainable Heavy Metal Remediation, volume 8, chapter 6, pages 165{198. Springer International Publishing, 2017. ISBN 978-3-319-58621-2. doi: 10.1007/978-3-319-58622-9. URL http://link.springer.com/10.1007/978-3-319-58622-9. [9] Okkyoung Choi and Byoung-In Sang. Extracellular electron transfer from cathode to microbes: application for biofuel production. Biotechnology for Biofuels, 9(1):11, 2016. ISSN 1754-6834. doi: 10.1186/s13068-016-0426-0. URL http://www.biotechnologyforbiofuels.com/content/9/1/11. [10] Charles R Myers and Kenneth H. Nealson. Bacterial Manganese Reduction and Growth with Manganese Oxide as the Sole Electron Acceptor. Science, 240(4857):1319{1321, 1988. [11] Sebastian Beblawy, Thea Bursac, Catarina Paquete, Ricardo Louro, Thomas A Clarke, and Johannes Gescher. Extracellular reduction of solid electron acceptors by Shewanella oneidensis. Molecular Micro- biology, 109(5):571{583, 2018. doi: 10.1111/mmi.14067. URL https://onlinelibrary.wiley.com/doi/pdf/10.1111/mmi.14067. [12] Marcus J. Edwards, Gaye F. White, Colin W.J. Lockwood, Matthew Lawes, Anne Martel, Gemma Harris, David J. Scott, David J. Richard- son, Julea N. Butt, and Thomas A. Clarke. Structural model of a porin-cytochrome electron conduit from the outer membrane of a metal reducing bacterium suggests electron transfer via periplasmic redox partners. Journal of Biological Chemistry, (Ii):jbc.RA118.001850, 2018. ISSN 0021-9258. doi: 10.1074/jbc.RA118.001850. URL http://www.jbc.org/lookup/doi/10.1074/jbc.RA118.001850. [13] Marian Breuer, Kevin M. Rosso, Jochen Blumberger, and Julea N. Butt. Multi-haem cytochromes in Shewanella oneidensis MR-1: structures, func- tions and opportunities. Journal of the Royal Society, Interface, 12(102): 20141117, 2015. ISSN 1742-5662. doi: 10.1098/rsif.2014.1117. URL http://rsif.royalsocietypublishing.org/content/12/102/20141117. [14] Marcus J. Edwards, Gaye F. White, Michael Norman, Alice Tome-Fernandez, Emma V. Ainsworth, Liang Shi, James K. Fredrickson, John M. Zachara, 122 Julea N. Butt, David J. Richardson, and Thomas A. Clarke. Redox Linked Flavin Sites in Extracellular Decaheme Proteins Involved in Microbe-Mineral Electron Transfer. Scientic Reports, 5(1):11677, 2015. ISSN 2045-2322. doi: 10.1038/srep11677. URL http://www.nature.com/articles/srep11677. [15] Marcus J. Edwards, Gaye F. White, Julea N. Butt, David J. Richard- son, and Thomas A. Clarke. The Crystal Structure of a Biolog- ical Insulated Transmembrane Molecular Wire. Cell, 181:1{9, apr 2020. ISSN 00928674. doi: 10.1016/j.cell.2020.03.032. URL https://linkinghub.elsevier.com/retrieve/pii/S0092867420303251. [16] David J. Richardson, Julea N. Butt, James K. Fredrickson, John M. Zachara, Liang Shi, Marcus J. Edwards, Gaye F. White, Nanakow A. Baiden, Andrew J. Gates, Sophie J. Marritt, and Thomas A. Clarke. The `porin-cytochrome' model for microbe-to-mineral electron transfer. Molecu- lar Microbiology, 85(2):201{212, 2012. ISSN 0950382X. doi: 10.1111/j.1365- 2958.2012.08088.x. [17] C.R. Myers and J.M. Myers. The outer membrane cytochromes of Shewanella oneidensis MR-1 are lipoproteins. Letters in Applied Microbiology, 39(5): 466{470, nov 2004. ISSN 0266-8254. doi: 10.1111/j.1472-765X.2004.01611.x. URL http://doi.wiley.com/10.1111/j.1472-765X.2004.01611.x. [18] Dan Coursolle and Jerey A. Gralnick. Modularity of the Mtr respiratory pathway of Shewanella oneidensis strain MR-1. Molecular Microbiology, 77 (4):995{1008, 2010. ISSN 0950382X. doi: 10.1111/j.1365-2958.2010.07266.x. [19] M Schicklberger, G Sturm, and J Gescher. Genomic Plasticity Enables a Secondary Electron Transport Pathway in Shewanella oneidensis. Applied & Environmental Microbiology, 79(4):1150{1159, 2013. doi: 10.1128/AEM.03556-12. URL http://dx.doi.org/10.1128. [20] Jerey A. Gralnick, H. Vali, D. P. Lies, and Dianne K. Newman. Extra- cellular respiration of dimethyl sulfoxide by Shewanella oneidensis strain MR-1. Proceedings of the National Academy of Sciences, 103(12):4669{ 4674, 2006. ISSN 0027-8424. doi: 10.1073/pnas.0505959103. URL http://www.pnas.org/cgi/doi/10.1073/pnas.0505959103. [21] Shuai Xu, Alexandre Barrozo, Leonard M. Tender, Anna I Krylov, and Mohamed Y. El-Naggar. Multiheme Cytochrome Mediated Redox Conduc- tion through Shewanella oneidensis MR-1 Cells. Journal of the American Chemical Society, 2018. ISSN 0002-7863. doi: 10.1021/jacs.8b05104. URL https://pubs.acs.org/sharingguidelines. 123 [22] Yuri A. Gorby, Svetlana Yanina, Jerey S. McLean, Kevin M. Rosso, Dianne Moyles, Alice C. Dohnalkova, Terry J Beveridge, In Seop Chang, Byung Hong Kim, Kyung Shik Kim, David E. Culley, Samantha B. Reed, Margaret F. Romine, Daad A. Saarini, Eric A. Hill, Liang Shi, Dwayne A. Elias, David W. Kennedy, Grigoriy Pinchuk, Kazuya Watanabe, Shun'ichi Ishii, Bruce E. Logan, Kenneth H. Nealson, and James K. Fredrickson. Electrically conductive bacterial nanowires pro- duced by Shewanella oneidensis strain MR-1 and other microorgan- isms. Proceedings of the National Academy of Sciences, 103(30):11358{ 11363, 2006. ISSN 0027-8424. doi: 10.1073/pnas.0905246106. URL http://www.pnas.org/cgi/doi/10.1073/pnas.0905246106. [23] Mohamed Y. El-Naggar, G. Wanger, Kar Man Leung, T. D. Yuzvinsky, G. Southam, J. Yang, W. M. Lau, Kenneth H. Nealson, and Yuri A. Gorby. Electrical transport along bacterial nanowires from Shewanella onei- densis MR-1. Proceedings of the National Academy of Sciences, 107(42): 18127{18131, 2010. ISSN 0027-8424. doi: 10.1073/pnas.1004880107. URL http://www.pnas.org/cgi/doi/10.1073/pnas.1004880107. [24] Sahand Pirbadian, Sarah E. Barchinger, Kar Man Leung, Hye Suk Byun, Yamini Jangir, Rachida A. Bouhenni, Samantha B. Reed, Margaret F. Romine, Daad A. Saarini, Liang Shi, Yuri A. Gorby, John H. Golbeck, and Mohamed Y. El-Naggar. Shewanella oneidensis MR-1 nanowires are outer membrane and periplasmic extensions of the extracellular electron trans- port components. Proceedings of the National Academy of Sciences, 111(35): 12883{12888, 2014. ISSN 0027-8424. doi: 10.1073/pnas.1410551111. URL http://www.pnas.org/cgi/doi/10.1073/pnas.1410551111. [25] Poorna Subramanian, Sahand Pirbadian, Mohamed Y. El-Naggar, and Grant J. Jensen. Ultrastructure of Shewanella oneidensis MR-1 nanowires revealed by electron cryotomography. Proceedings of the National Academy of Sciences, 115(14):103242, 2018. ISSN 0027-8424. doi: 10.1101/103242. URL https://www.biorxiv.org/content/early/2017/01/28/103242. [26] Telma C. Santos, Marta A. Silva, Leonor Morgado, Joana M. Dantas, and Carlos A. Salgueiro. Diving into the redox properties of Geobac- ter sulfurreducens cytochromes: a model for extracellular electron trans- fer. Dalton Transactions, 44(20):9335{9344, 2015. ISSN 1477-9226. doi: 10.1039/C5DT00556F. URL http://xlink.rsc.org/?DOI=C5DT00556F. [27] Bradley G Lusk, Jonathan P. Badalamenti, Prathap Parameswaran, Daniel R. Bond, and C esar I Torres. Draft genome sequence of the gram-positive thermophilic iron reducer Thermincola ferriacetica 124 strain Z-0001. Genome Announcements, 3(5):1072{1073, 2015. doi: 10.1128/genomeA.01072-15.Copyright. [28] Jonathan P. Badalamenti, Zarath M. Summers, Chi Ho Chan, Jerey A. Gralnick, and Daniel R. Bond. Isolation and genomic characterization of `Desulfuromonas soudanensis WTL', a metal- and electrode-respiring bac- terium from anoxic deep subsurface brine. Frontiers in Microbiology, 7(JUN): 1{11, 2016. ISSN 1664302X. doi: 10.3389/fmicb.2016.00913. [29] Mimori Tonomura, Ayaka Ehara, Haruo Suzuki, and Seigo Amachi. Draft Genome Sequence of Anaeromyxobacter sp. strain PSR-1, an arsenate- respiring bacterium isolated from arsenic-contaminated soil. Genome Announcements, 3(3):4{5, 2015. doi: 10.1128/genomeA.00472-15.Copyright. [30] Jessica A. Smith, Muktak Aklujkar, Carla Risso, Ching Leang, Ludovic Giloteaux, and Dawn E. Holmes. Mechanisms involved in Fe(III) respira- tion by the hyperthermophilic archaeon Ferroglobus placidus. Applied and Environmental Microbiology, 81(8):2735{2744, 2015. ISSN 10985336. doi: 10.1128/AEM.04038-14. [31] Michael P. Manzella, Dawn E. Holmes, Jessica M. Rocheleau, Amanda Chung, Gemma Reguera, and Kazem Kashe. The complete genome sequence and emendation of the hyperthermophilic, obligate iron-reducing archaeon \Geoglobus ahangari" strain 234T. Standards in Genomic Sci- ences, 10(1):1{19, 2015. ISSN 19443277. doi: 10.1186/s40793-015-0035-8. URL http://dx.doi.org/10.1186/s40793-015-0035-8. [32] Alex W. Hernsdorf, Yuki Amano, Kazuya Miyakawa, Kotaro Ise, Yohey Suzuki, Karthik Anantharaman, Alexander Probst, David Burstein, Brian C. Thomas, and Jillian F. Baneld. Potential for microbial H 2 and metal transformations associated with novel bacteria and archaea in deep terrestrial subsurface sediments. ISME Journal, 11(8):1915{ 1929, 2017. ISSN 17517370. doi: 10.1038/ismej.2017.39. URL http://dx.doi.org/10.1038/ismej.2017.39. [33] Satoshi Kawaichi, Tetsuya Yamada, Akio Umezawa, Shawn E. McG- lynn, Takehiro Suzuki, Naoshi Dohmae, Takashi Yoshida, Yoshihiko Sako, Nobuhiro Matsushita, Kazuhito Hashimoto, and Ryuhei Nakamura. Anodic and cathodic extracellular electron transfer by the lamentous bacterium Ardenticatena maritima 110S. Frontiers in Microbiology, 9(FEB), 2018. ISSN 1664302X. doi: 10.3389/fmicb.2018.00068. [34] Daniel E Ross, Jerey M Flynn, Daniel B Baron, Jerey A. Gralnick, and Daniel R. Bond. Towards Electrosynthesis in Shewanella: Energetics of 125 Reversing the Mtr Pathway for Reductive Metabolism. PLoS ONE, 6(2): 1{9, 2011. doi: 10.1371/journal.pone.0016649. [35] Kelvin B. Gregory, Daniel R. Bond, and Derek R. Lovley. Graphite elec- trodes as electron donors for anaerobic respiration. Environmental Micro- biology, 6(6):596{604, jun 2004. ISSN 14622912. doi: 10.1111/j.1462- 2920.2004.00593.x. URL https://pubmed.ncbi.nlm.nih.gov/15142248/. [36] Zheming Wang, Zhi Shi, Liang Shi, Gaye F. White, David J. Richardson, Thomas A. Clarke, James K. Fredrickson, and John M. Zachara. Eects of soluble avin on heterogeneous electron transfer between surface-exposed bacterial cytochromes and iron oxides. Geochimica et Cosmochimica Acta, 163:299{310, 2015. ISSN 00167037. doi: 10.1016/j.gca.2015.03.039. [37] Linpeng Yu, Yong Yuan, Christopher Rensing, and Shungui Zhou. Combined spectroelectrochemical and proteomic characterizations of bidirectional Alcaligenes faecalis-electrode electron transfer. Biosensors and Bioelectron- ics, 106:21{28, may 2018. ISSN 0956-5663. doi: 10.1016/J.BIOS.2018.01.032. [38] Nabin Aryal, Pier-Luc Tremblay, Dawid M. Lizak, and Tian Zhang. Per- formance of dierent Sporomusa species for the microbial electrosynthesis of acetate from carbon dioxide. Bioresource Technology, 233:184{190, jun 2017. ISSN 0960-8524. doi: 10.1016/J.BIORTECH.2017.02.128. [39] J org S. Deutzmann, Merve Sahin, and Alfred M. Spormann. Extracellular Enzymes Facilitate Electron Uptake in Biocorrosion and Bioelectrosynthesis. mBio, 6(2):1{8, 2015. ISSN 2150-7511. doi: 10.1128/mBio.00496-15.Editor. [40] Shaoan Cheng, Defeng Xing, Douglas F. Call, and Bruce E. Logan. Direct Biological Conversion of Electrical Current into Methane by Electromethanogenesis. Environmental Science and Technology, 43:3953{3958, 2009. doi: 10.1021/es803531g. URL https://pubs.acs.org/doi/pdf/10.1021/es803531g. [41] Kelly P. Nevin, Trevor L. Woodard, Ashley E. Franks, Zarath M. Sum- mers, and Derek R. Lovley. Microbial electrosynthesis: feeding microbes electricity to convert carbon dioxide and water to multicarbon extracellular organic compounds. mBio, 1(2):e00103{10, may 2010. ISSN 2150-7511. doi: 10.1128/mBio.00103-10. [42] Sof a Carbajosa, Moustaf a Malki, Renaud Caillard, Mar a F. Lopez, F. Javier Palomares, Jos e A. Mart n-Gago, Nuria Rodr guez, Ricardo Amils, Vic- tor M. Fern andez, and Antonio L. De Lacey. Electrochemical growth of Acidithiobacillus ferrooxidans on a graphite electrode for obtaining 126 a biocathode for direct electrocatalytic reduction of oxygen. Biosen- sors and Bioelectronics, 26(2):877{880, oct 2010. ISSN 0956-5663. doi: 10.1016/J.BIOS.2010.07.037. [43] Zarath M. Summers, Jerey A. Gralnick, and Daniel R. Bond. Cultivation of an Obligate Fe(II)-Oxidizing Lithoautotrophic Bacterium Using Electrodes. mBio, 4(1):e00420{12, 2013. ISSN , 2150-7511. doi: 10.1128/mBio.00420- 12.Editor. [44] A Bose, E J Gardel, C Vidoudez, E A Parra, and P R Gir- guis. Electron uptake by iron-oxidizing phototrophic bacteria. Nature Communications, 5:1{7, 2014. doi: 10.1038/ncomms4391. URL https://www.nature.com/articles/ncomms4391.pdf. [45] Juan Liu, Zheming Wang, Sara M Belchik, Marcus J. Edwards, Chongxuan Liu, David W. Kennedy, Eric D Merkley, Mary S Lipton, Julea N. Butt, David J. Richardson, John M. Zachara, James K. Fredrickson, Kevin M. Rosso, Liang Shi, and David Emerson. Identication and characterization of MtoA: a decaheme c-type cytochrome of the neutrophilic Fe(II)-oxidizing bacterium Sideroxydans lithotrophicus ES-1. Frontiers in Microbiology, 3: 1{11, 2012. doi: 10.3389/fmicb.2012.00037. [46] Pascal F. Beese-Vasbender, Jan-Philipp Grote, Julia Garrelfs, Martin Strat- mann, and Karl J.J. Mayrhofer. Selective microbial electrosynthesis of methane by a pure culture of a marine lithoautotrophic archaeon. Biochem- istry, 102:50{55, 2015. doi: 10.1016/j.bioelechem.2014.11.004. [47] Xiao Deng, Ryuhei Nakamura, Kazuhito Hashimoto, and Akihiro Okamoto. Electron Extraction from an Extracellular Electrode by Desulfovibrio fer- rophilus Strain IS5 Without Using Hydrogen as an Electron Carrier. Elec- trochemistry, 83(7):529{531, 2015. ISSN 1344-3542. doi: 10.5796/electro- chemistry.83.529. [48] Zarath M. Summers, Heather E. Fogarty, Ching Leang, Ashley E. Franks, Nikhil S. Malvankar, and Derek R. Lovley. Direct exchange of electrons within aggregates of an evolved syntrophic coculture of anaerobic bacteria. Science, 330(December):1413{1415, 2010. [49] Qiwen Cheng and Douglas F. Call. Hardwiring microbes via direct inter- species electron transfer: mechanisms and applications. Environmental Sci- ence: Processes & Impacts, 18(8):968{980, 2016. ISSN 2050-7887. doi: 10.1039/C6EM00219F. URL http://xlink.rsc.org/?DOI=C6EM00219F. [50] Derek R. Lovley. Electrically conductive pili: Biological function and poten- tial applications in electronics. Current Opinion in Electrochemistry, 4(1): 127 190{198, 2017. ISSN 24519111. doi: 10.1016/j.coelec.2017.08.015. URL https://doi.org/10.1016/j.coelec.2017.08.015. [51] Amelia-Elena Rotaru, Pravin Malla Shrestha, Fanghua Liu, Minita Shrestha, Devesh Shrestha, Mallory Embree, Karsten Zengler, Colin Wardman, Kelly P. Nevin, and Derek R. Lovley. A new model for electron ow during anaerobic digestion: direct interspecies electron transfer to Methanosaeta for the reduction of carbon dioxide to methane. Energy & Environmental Sci- ence, 7(1):408{415, 2014. ISSN 1754-5692. doi: 10.1039/C3EE42189A. URL http://xlink.rsc.org/?DOI=C3EE42189A. [52] Amelia-Elena Rotaru, Pravin Malla Shrestha, Fanghua Liu, Beatrice Markovaite, Shanshan Chen, Kelly P. Nevin, and Derek R. Lovley. Direct interspecies electron transfer between Geobacter metallireducens and Methanosarcina barkeri. Applied and Environmental Microbiology, 80(15): 4599{4605, 2014. ISSN 10985336. doi: 10.1128/AEM.00895-14. [53] Phuc T. Ha, Stephen R. Lindemann, Liang Shi, Alice C. Dohnalkova, James K. Fredrickson, Michael T. Madigan, and Haluk Beyenal. Syntrophic anaerobic photosynthesis via direct interspecies electron transfer. Nature Communications, 8:1{7, 2017. ISSN 20411723. doi: 10.1038/ncomms13924. URL http://dx.doi.org/10.1038/ncomms13924. [54] Shawn E. McGlynn. Energy Metabolism during Anaerobic Methane Oxida- tion in ANME Archaea. Microbes and Environments, 32(1):5{13, 2017. ISSN 1342-6311. doi: 10.1264/jsme2.ME16166. [55] Peer H.A. Timmers, Cornelia U. Welte, Jasper J. Koehorst, Caroline M. Plugge, Mike S.M. Jetten, and Alfons J.M. Stams. Reverse Methanogenesis and Respiration in Methanotrophic Archaea. Archaea, 2017, 2017. ISSN 14723646. doi: 10.1155/2017/1654237. [56] Viola Krukenberg, Dietmar Riedel, Harald R. Gruber-Vodicka, Pier Luigi Buttigieg, Halina E. Tegetmeyer, Antje Boetius, and Gunter Wegener. Gene expression and ultrastructure of meso- and thermophilic methanotrophic consortia. Environmental Microbiol- ogy, 2018. ISSN 14622912. doi: 10.1111/1462-2920.14077. URL http://doi.wiley.com/10.1111/1462-2920.14077. [57] Shawn E. McGlynn, Grayson L. Chadwick, Christopher P. Kempes, and Victoria J. Orphan. Single cell activity reveals direct elec- tron transfer in methanotrophic consortia. Nature, 526(7574):531{ 535, 2015. ISSN 0028-0836. doi: 10.1038/nature15512. URL http://www.nature.com/doifinder/10.1038/nature15512. 128 [58] Gunter Wegener, Viola Krukenberg, Dietmar Riedel, Halina E. Teget- meyer, and Antje Boetius. Intercellular wiring enables electron trans- fer between methanotrophic archaea and bacteria. Nature, 526(7574): 587{590, 2015. ISSN 14764687. doi: 10.1038/nature15733. URL https://www.nature.com/articles/nature15733. [59] Arnulf Kletzin, Thomas Heimerl, Jennifer Flechsler, Laura van Niftrik, Rein- hard Rachel, and Andreas Klingl. Cytochromes c in Archaea: Distribu- tion, maturation, cell architecture, and the special case of Ignicoccus hos- pitalis. Frontiers in Microbiology, 6(MAY), 2015. ISSN 1664302X. doi: 10.3389/fmicb.2015.00439. [60] Connor T. Skennerton, Karuna Chourey, Ramsunder Iyer, Robert L. Hettich, Gene W. Tyson, and Victoria J. Orphan. Methane-fueled syntrophy through extracellular electron transfer: Uncovering the genomic traits conserved within diverse bacterial partners of anaerobic methanotrophic archaea. mBio, 8(4):1{14, 2017. ISSN 21507511. doi: 10.1128/mBio.00530-17. [61] Viola Krukenberg, Katie Harding, Michael Richter, Frank Oliver Gl ockner, Harald R. Gruber-Vodicka, Birgit Adam, Jasmine S. Berg, Katrin Knit- tel, Halina E. Tegetmeyer, Antje Boetius, and Gunter Wegener. Candida- tus Desulfofervidus auxilii, a hydrogenotrophic sulfate-reducing bacterium involved in the thermophilic anaerobic oxidation of methane. Environmental Microbiology, 18(9):3073{3091, 2016. ISSN 14622920. doi: 10.1111/1462- 2920.13283. [62] Frank M. Etzler and Jaroslaw Drelich. Atomic Force Microscopy for Char- acterization of Surfaces, Particles, and Their Interactions. In Developments in Surface Contamination and Cleaning: Detection, Characterization, and Analysis of Contaminants, pages 307{331. Elsevier Inc., jan 2012. ISBN 9781437778830. doi: 10.1016/B978-1-4377-7883-0.00006-7. [63] Samara L. Reck-Peterson, Nathan D. Derr, and Nico Stuurman. Imag- ing single molecules using total internal re ection uorescence microscopy (TIRFM). Cold Spring Harbor Protocols, 5(3):1{12, 2010. ISSN 15596095. doi: 10.1101/pdb.top73. [64] Fabien Pinaud, Samuel Clarke, Assa Sittner, and Maxime Dahan. Prob- ing cellular events, one quantum dot at a time. Nature Methods, 7(4): 275{285, apr 2010. ISSN 1548-7091. doi: 10.1038/nmeth.1444. URL http://www.nature.com/articles/nmeth.1444. [65] Oleg Kovtun, Ian D. Tomlinson, Danielle M. Bailey, Lucas B. Thal, Emily J. Ross, Lauren Harris, Michael P. Frankland, Riley S. Fer- guson, Zachary Glaser, Jonathan Greer, and Sandra J. Rosenthal. 129 Single quantum dot tracking illuminates neuroscience at the nanoscale. Chemical Physics Letters, 706:741{752, aug 2018. ISSN 00092614. doi: 10.1016/j.cplett.2018.06.019. URL /pmc/articles/PMC6157616/ /pmc/articles/PMC6157616/?report=abstract https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6157616/. [66] Ken Ritchie, Yoriko Lill, Chetan Sood, Hochan Lee, and Shunyuan Zhang. Single-molecule imaging in live bacteria cells. Philosophical Transactions of the Royal Society B: Biological Sciences, 368(1611), feb 2013. ISSN 14712970. doi: 10.1098/rstb.2012.0355. [67] Andreas Gahlmann and W. E. Moerner. Exploring bacterial cell biology with single-molecule tracking and super-resolution imaging. Nature Reviews Microbiology, 12(1):9{22, jan 2014. ISSN 1740-1526. doi: 10.1038/nrmi- cro3154. URL http://www.nature.com/articles/nrmicro3154. [68] Grace W. Chong, Sahand Pirbadian, and Mohamed Y. El-Naggar. Surface- Induced Formation and Redox-Dependent Staining of Outer Membrane Extensions in Shewanella oneidensis MR-1. Frontiers in Energy Research, 7:1{9, aug 2019. ISSN 2296-598X. doi: 10.3389/fenrg.2019.00087. URL https://www.frontiersin.org/article/10.3389/fenrg.2019.00087. [69] Alex S. Beliaev, Daad A. Saarini, Jennifer L. McLaughlin, and David Hun- nicutt. MtrC, an outer membrane decahaem c cytochrome required for metal reduction in Shewanella putrefaciens MR-1. Molecular Microbiology, 39(3): 722{730, feb 2001. ISSN 0950-382X. doi: 10.1046/j.1365-2958.2001.02257.x. URL http://doi.wiley.com/10.1046/j.1365-2958.2001.02257.x. [70] Judith M Myers and Charles R Myers. Role for Outer Membrane Cytochromes OmcA and OmcB of Shewanella putrefaciens MR-1 in Reduc- tion of Manganese Dioxide. Applied and Environmental Microbiology, 67(1): 260{269, 2001. doi: 10.1128/AEM.67.1.260-269.2001. [71] Kar Man Leung, Greg Wanger, Mohamed Y El-Naggar, Yuri Gorby, Gor- don Southam, Woon Ming Lau, and Jun Yang. Shewanella oneiden- sis MR-1 Bacterial Nanowires Exhibit p-Type, Tunable Electronic Behav- ior. Nano Letters, 13(6):2407{2411, 2013. doi: 10.1021/nl400237p. URL https://pubs.acs.org/doi/10.1021/nl400237p. [72] Lori A. Zacharo and Mohamed Y. El-Naggar. Redox conduction in biolms: From respiration to living electronics. Current Opinion in Electrochemistry, 4(1):182{189, 2017. ISSN 2451-9103. doi: 10.1016/j.coelec.2017.09.003. URL https://doi.org/10.1016/j.coelec.2017.09.003. 130 [73] Sarah E. Barchinger, Sahand Pirbadian, Christine Sambles, Carol S. Baker, Kar Man Leung, Nigel J. Burroughs, Mohamed Y. El-Naggar, and John H. Golbeck. Regulation of Gene Expression During Elec- tron Acceptor Limitation and Bacterial Nanowire Formation in Shewanella oneidensis MR-1. Applied and Environmental Microbiology, 82(17):5428{ 5443, 2016. ISSN 0099-2240. doi: 10.1128/AEM.01615-16. URL http://aem.asm.org/lookup/doi/10.1128/AEM.01615-16. [74] Charles R Myers and Judith M Myers. Localization of Cytochromes to the Outer Membrane of Anaerobically Grown Shewanella putrefaciens MR- 1. Journal of Bacteriology, 174(11):3429{3438, 1992. ISSN 00219193. doi: 10.1016/0378-1097(93)90480-P. [75] Gerhard Wanner, Kajetan Vogl, and J org Overmann. Ultrastructural char- acterization of the prokaryotic symbiosis in \Chlorochromatium aggregatum". Journal of Bacteriology, 190(10):3721{3730, 2008. ISSN 00219193. doi: 10.1128/JB.00027-08. [76] Gyanendra P. Dubey and Sigal Ben-Yehuda. Intercellular nan- otubes mediate bacterial communication. Cell, 144(4):590{600, 2011. ISSN 00928674. doi: 10.1016/j.cell.2011.01.015. URL http://dx.doi.org/10.1016/j.cell.2011.01.015. [77] Svetlana I. Galkina, Julia M. Romanova, Elizaveta E. Bragina, Irina G. Tiganova, Vladimir I. Stadnichuk, Natalia V. Alekseeva, Vladimir Y. Polyakov, and Thomas Klein. Membrane tubules attach Salmonella Typhimurium to eukaryotic cells and bacteria. FEMS Immunology and Med- ical Microbiology, 61(1):114{124, 2011. ISSN 09288244. doi: 10.1111/j.1574- 695X.2010.00754.x. [78] Ameesha Shetty, Shicheng Chen, Elitza I. Tocheva, Grant J. Jensen, and William J. Hickey. Nanopods: A new bacterial structure and mechanism for deployment of outer membrane vesicles. PLoS ONE, 6(6), 2011. ISSN 19326203. doi: 10.1371/journal.pone.0020725. [79] William D. McCaig, Antonius Koller, and David G. Thanassi. Production of outer membrane vesicles and outer membrane tubes by Francisella novi- cida. Journal of Bacteriology, 195(6):1120{1132, 2013. ISSN 00219193. doi: 10.1128/JB.02007-12. [80] Saida Benomar, David Ranava, Mar a Luz C ardenas, Eric Trably, Yan Rafra, Adrien Ducret, J er^ ome Hamelin, Elisabeth Lojou, Jean Philippe Steyer, and Marie Th er ese Giudici-Orticoni. Nutritional stress induces exchange of cell material and energetic coupling between bacterial 131 species. Nature Communications, 6, 2015. ISSN 20411723. doi: 10.1038/ncomms7283. [81] Samay Pande, Shraddha Shitut, Lisa Freund, Martin Westermann, Felix Bertels, Claudia Colesie, Ilka B. Bischofs, and Christian Kost. Metabolic cross-feeding via intercellular nanotubes among bacteria. Nature Commu- nications, 6:1{13, 2015. ISSN 20411723. doi: 10.1038/ncomms7238. URL http://dx.doi.org/10.1038/ncomms7238. [82] Carla P erez-Cruz, Ornella Carri on, Lidia Delgado, Gemma Martinez, Car- men L opez-Iglesias, and Elena Mercade. New type of outer membrane vesi- cle produced by the Gram-negative bacterium Shewanella vesiculosa M7T: Implications for DNA content. Applied and Environmental Microbiology, 79 (6):1874{1881, 2013. ISSN 00992240. doi: 10.1128/AEM.03657-12. [83] Jonathan P. Remis, Dongguang Wei, Amita Gorur, Marcin Zemla, Jessica Haraga, Simon Allen, H. Ewa Witkowska, J. William Costerton, James E. Berleman, and Manfred Auer. Bacterial social networks: Structure and com- position of Myxococcus xanthus outer membrane vesicle chains. Environmen- tal Microbiology, 16(2):598{610, 2014. ISSN 14622912. doi: 10.1111/1462- 2920.12187. [84] Xueming Wei, Christopher N. Vassallo, Darshankumar T. Pathak, and Daniel Wall. Myxobacteria produce outer membrane-enclosed tubes in unstructured environments. Journal of Bacteriology, 196(10):1807{1814, 2014. ISSN 10985530. doi: 10.1128/JB.00850-13. [85] Gyanendra P. Dubey, Ganesh Babu Malli Mohan, Anna Dubrovsky, Triana Amen, Shai Tsipshtein, Alex Rouvinski, Alex Rosenberg, Daniel Kaganovich, Eilon Sherman, Ohad Medalia, and Sigal Ben-Yehuda. Architecture and Characteristics of Bacterial Nanotubes. Developmental Cell, 36(4):453{ 461, 2016. ISSN 18781551. doi: 10.1016/j.devcel.2016.01.013. URL http://dx.doi.org/10.1016/j.devcel.2016.01.013. [86] Yuri A. Gorby, J. McLean, A. Korenevsky, Kevin M. Rosso, Mohamed Y. El-Naggar, and Terry J Beveridge. Redox-reactive membrane vesi- cles produced by Shewanella. Geobiology, 6(3):232{241, jun 2008. ISSN 1472-4677. doi: 10.1111/j.1472-4669.2008.00158.x. URL http://doi.wiley.com/10.1111/j.1472-4669.2008.00158.x. [87] Xiao Deng, Naoshi Dohmae, Kenneth H. Nealson, Kazuhito Hashimoto, and Akihiro Okamoto. Multi-heme cytochromes pro- vide a pathway for survival in energy-limited environments. Science Advances, (February):1{9, 2018. doi: 10.1126/sciadv.aao5682. URL http://advances.sciencemag.org/cgi/content/short/4/2/eaao5682. 132 [88] Dan Coursolle and Jerey A. Gralnick. Reconstruction of extracellular res- piratory pathways for iron(III) reduction in Shewanella oneidensis strain MR-1. Frontiers in Microbiology, 3(FEB):1{11, 2012. ISSN 1664302X. doi: 10.3389/fmicb.2012.00056. [89] Miriam A. Rosenbaum, Haim Y. Bar, Qasim K. Beg, Daniel Segr e, James Booth, Michael A. Cotta, and Largus T. Angenent. Transcriptional Analysis of Shewanella oneidensis MR-1 with an Electrode Compared to Fe(III)Citrate or Oxygen as Terminal Electron Acceptor. PLoS ONE, 7(2): e30827, feb 2012. ISSN 1932-6203. doi: 10.1371/journal.pone.0030827. URL https://dx.plos.org/10.1371/journal.pone.0030827. [90] Michaela A. TerAvest, Miriam A. Rosenbaum, Nicholas J. Kotloski, Jef- frey A. Gralnick, and Largus T. Angenent. Oxygen allows Shewanella oneidensis MR-1 to overcome mediator washout in a continuously fed bioelectrochemical system. Biotechnology and Bioengineering, 111(4): 692{699, apr 2014. ISSN 00063592. doi: 10.1002/bit.25128. URL http://doi.wiley.com/10.1002/bit.25128. [91] Julia G odeke, Kristina Paul, J urgen Lassak, and Kai M Thormann. Phage- induced lysis enhances biolm formation in Shewanella oneidensis MR-1. The ISME Journal, 5:613{626, 2011. doi: 10.1038/ismej.2010.153. URL www.nature.com/ismej. [92] Lynne Turnbull, Masanori Toyofuku, Amelia L. Hynen, Masaharu Kuro- sawa, Gabriella Pessi, Nicola K. Petty, Sarah R. Osvath, Gerardo C arcamo-Oyarce, Erin S. Gloag, Raz Shimoni, Ulrich Omasits, Satoshi Ito, Xinhui Yap, Leigh G. Monahan, Rosalia Cavaliere, Christian H. Ahrens, Ian G. Charles, Nobuhiko Nomura, Leo Eberl, and Cynthia B. Whitchurch. Explosive cell lysis as a mechanism for the biogenesis of bacterial membrane vesicles and biolms. Nature Communications, 7(1): 11220, sep 2016. ISSN 2041-1723. doi: 10.1038/ncomms11220. URL http://www.nature.com/articles/ncomms11220. [93] Sukhvinder Gill, Ryan Catchpole, and Patrick Forterre. Extra- cellular membrane vesicles in the three domains of life and beyond. FEMS Microbiology Reviews, 43(3):273{303, may 2019. ISSN 1574-6976. doi: 10.1093/femsre/fuy042. URL https://academic.oup.com/femsre/article/43/3/273/5195520. [94] Masanori Toyofuku, Nobuhiko Nomura, and Leo Eberl. Types and ori- gins of bacterial membrane vesicles. Nature Reviews Microbiology, 17(1): 13{24, jan 2019. ISSN 1740-1526. doi: 10.1038/s41579-018-0112-2. URL http://www.nature.com/articles/s41579-018-0112-2. 133 [95] Amanda J McBroom, Alexandra P Johnson, Sreekanth Vemulapalli, and Meta J Kuehn. Outer membrane vesicle production by Escherichia coli is independent of membrane instability. Journal of bacteriology, 188(15):5385{ 92, aug 2006. ISSN 0021-9193. doi: 10.1128/JB.00498-06. [96] Fengbin Wang, Yangqi Gu, J Patrick O'brien, Allon I Hochbaum, Edward H Egelman, Nikhil S Malvankar, Sophia M Yi, Sibel Ebru Yalcin, Vishok Srikanth, Cong Shen, Dennis Vu, and Nicole L Ing. Structure of Microbial Nanowires Reveals Stacked Hemes that Trans- port Electrons over Micrometers Article Structure of Microbial Nanowires Reveals Stacked Hemes that Transport Electrons over Micrometers. Cell, 177:361{369, 2019. doi: 10.1016/j.cell.2019.03.029. URL https://doi.org/10.1016/j.cell.2019.03.029. [97] Matthew D. Yates, Sarah M. Strycharz-Glaven, Joel P. Golden, Jared N. Roy, Stanislav Tsoi, Jerey S. Erickson, Mohamed Y. El-Naggar, Scott Calabrese Barton, and Leonard M. Tender. Measuring conductivity of living Geobacter sulfurreducens biolms. Nature Nanotechnology, 11(11):910{913, 2016. ISSN 17483395. doi: doi:10.1038/nnano.2016.186. [98] Michael E Kovach, Philip H Elzer, D Steven Hill, Gregory T Robertson, Michael A Farris, R Martin Roop, and Kenneth M Peterson. Four new deriva- tives of the broad-host-range cloning vector pBBR1MCS, carrying dierent antibiotic-resistance cassettes. Elsevier Science, 166:175{176, 1995. URL papers2://publication/uuid/4156ADB9-DA4F-41CF-953A-78E72095543C. [99] Dorothy Beckett, Elena Kovaleva, and Peter J. Schatz. A minimal peptide substrate in biotin holoenzyme synthetase-catalyzed biotinylation. Protein Science, 8:921{929, 1999. ISSN 09618368. doi: 10.1110/ps.8.4.921. URL http://doi.wiley.com/10.1110/ps.8.4.921. [100] Mark Howarth, K. Takao, Y. Hayashi, and Alice Y. Ting. Target- ing quantum dots to surface proteins in living cells with biotin lig- ase. Proceedings of the National Academy of Sciences, 102(21):7583{ 7588, 2005. ISSN 0027-8424. doi: 10.1073/pnas.0503125102. URL http://www.pnas.org/cgi/doi/10.1073/pnas.0503125102. [101] Mark Howarth, Wenhao Liu, Sujiet Puthenveetil, Yi Zheng, Lisa F. Mar- shall, Michael M. Schmidt, K. Dane Wittrup, Moungi G. Bawendi, and Alice Y. Ting. Monovalent, reduced-size quantum dots for imaging receptors on living cells. Nature Methods, 5(5):397{399, 2008. ISSN 15487091. doi: 10.1038/nmeth.1206. 134 [102] Mark Howarth and Alice Y. Ting. Imaging proteins in live mammalian cells with biotin ligase and monovalent streptavidin. Nature Protocols, 3(3):534{ 545, 2009. doi: 10.1038/nprot.2008.20.Imaging. [103] Anna D. Corts, Lynn C. Thomason, Ryan T. Gill, and Jerey A. Gralnick. A new recombineering system for precise genome-editing in Shewanella onei- densis strain MR-1 using single-stranded oligonucleotides. Scientic Reports, 9(1):39, dec 2019. ISSN 2045-2322. doi: 10.1038/s41598-018-37025-4. URL http://www.nature.com/articles/s41598-018-37025-4. [104] Annette McDonnel and L. Andrew Staehelin. Detection of Cytochrome f, a c-Class Cytochrome, with Diaminobenzidine in Polyacrylamide Gels. Ana- lytical Biochemistry, 44(198 1):40{44, 1981. [105] C. L. Reardon, A. C. Dohnalkova, P. Nachimuthu, D. W. Kennedy, D. A. Saarini, B. W. Arey, L. Shi, Z. Wang, D. Moore, J. S. Mclean, D. Moyles, M. J. Marshall, J. M. Zachara, J. K. Fredrickson, and A. S. Beliaev. Role of outer-membrane cytochromes MtrC and OmcA in the biomineralization of ferrihydrite by Shewanella oneidensis MR-1. Geobiology, 8(1):56{68, jan 2010. ISSN 14724677. doi: 10.1111/j.1472-4669.2009.00226.x. URL http://doi.wiley.com/10.1111/j.1472-4669.2009.00226.x. [106] Liang Shi, Shuang Deng, Matthew J. Marshall, Zheming Wang, David W. Kennedy, Alice C. Dohnalkova, Heather M. Mottaz, Eric A. Hill, Yuri A. Gorby, Alexander S. Beliaev, David J. Richardson, John M. Zachara, and James K. Fredrickson. Direct involvement of type II secretion system in extracellular translocation of Shewanella oneidensis outer membrane cytochromes MtrC and OmcA. Journal of Bacteriology, 190(15):5512{5516, 2008. ISSN 00219193. doi: 10.1128/JB.00514-08. [107] Bruce E. Logan, Ruggero Rossi, Ala'a Ragab, and Pascal E. Saikaly. Electroactive microorganisms in bioelectrochemical systems. Nature Reviews Microbiology, 17(5):307{319, may 2019. ISSN 17401534. doi: 10.1038/s41579-019-0173-x. [108] Sota Ikeda, Yuki Takamatsu, Miyu Tsuchiya, Keigo Suga, Yugo Tanaka, Atsushi Kouzuma, and Kazuya Watanabe. Shewanella oneidensis MR-1 as a bacterial platform for electro-biotechnology. Essays in Biochemistry, page 20200178, mar 2021. ISSN 0071-1365. doi: 10.1042/EBC20200178. [109] Xiuyun Jiang, Bastian Burger, Fruzsina Gajdos, C Bortolotti, Zdenek Futera, Marian Breuer, and Jochen Blumberger. Kinetics of trifurcated elec- tron ow in the decaheme bacterial proteins MtrC and MtrF. Proceedings of the National Academy of Sciences of the United States of America, 116(9): 135 3425{3430, feb 2019. ISSN 1091-6490. doi: 10.1073/pnas.1818003116. URL http://www.ncbi.nlm.nih.gov/pubmed/30755526. [110] Daniel E. Ross, Susan L. Brantley, and Ming Tien. Kinetic characteriza- tion of OmcA and MtrC, terminal reductases involved in respiratory elec- tron transfer for dissimilatory iron reduction in Shewanella oneidensis MR-1. Applied and Environmental Microbiology, 75(16):5218{5226, aug 2009. ISSN 00992240. doi: 10.1128/AEM.00544-09. [111] David N Blauch, Jean-Michel Sav eant, and H Langmuir. Dynamics of elec- tron hopping in assemblies of redox centers. Percolation and diusion. J. Am. Chem. Soc., 114(9):3323{3332, 1992. doi: 10.1021/ja00035a025. URL https://pubs.acs.org/doi/10.1021/ja00035a025. [112] Catarina M. Paquete, Bruno M. Fonseca, Davide R. Cruz, Tiago M. Pereira, Isabel Pacheco, Cl ~ A<udio M. Soares, and Ricardo O. Louro. Exploring the molecular mechanisms of electron shuttling across the microbe/metal space. Frontiers in Microbiology, 5(JUN):318, jun 2014. ISSN 1664-302X. doi: 10.3389/fmicb.2014.00318. [113] Shuo Zhang, Jiali Chang, Chao Lin, Yiran Pan, Kangping Cui, Xiaoyuan Zhang, Peng Liang, and Xia Huang. Enhancement of methano- genesis via direct interspecies electron transfer between Geobacteraceae and Methanosaetaceae conducted by granular activated carbon. Biore- source Technology, 245(June):132{137, 2017. ISSN 18732976. doi: 10.1016/j.biortech.2017.08.111. [114] Colin Kleanthous, Patrice Rassam, and Christoph G. Baumann. Protein- protein interactions and the spatiotemporal dynamics of bacterial outer membrane proteins, dec 2015. ISSN 1879033X. [115] Paul E. Schavemaker, Arnold J. Boersma, and Bert Poolman. How Important Is Protein Diusion in Prokaryotes? Frontiers in Molecular Biosciences, 5 (93):1{16, nov 2018. ISSN 2296-889X. doi: 10.3389/fmolb.2018.00093. [116] Michael J Saxton and Ken Jacobson. SINGLE-PARTICLE TRACKING: Applications to Membrane Dynamics. Annu. Rev. Biophys. Biomol. Struct, 26:373{99, 1997. URL www.annualreviews.org. [117] Mostyn T. Brown, Bradley C. Steel, Claudio Silvestrin, David A. Wilkinson, Nicolas J. Delalez, Craig N. Lumb, Boguslaw Obara, Judith P. Armitage, and Richard M. Berry. Flagellar hook exibility is essential for bundle formation in swimming Escherichia coli cells. Journal of Bacteriology, 194(13):3495{ 3501, 2012. ISSN 00219193. doi: 10.1128/JB.00209-12. 136 [118] Arnauld Serg e, Nicolas Bertaux, Herv e Rigneault, and Didier Mar- guet. Dynamic multiple-target tracing to probe spatiotemporal cartography of cell membranes. Nature Methods, 5(8):687{694, aug 2008. ISSN 1548-7091. doi: 10.1038/nmeth.1233. URL http://www.nature.com/articles/nmeth.1233. [119] Yunke Zhao and Fabien Pinaud. In Vivo Single-Molecule Tracking of Voltage-Gated Calcium Channels with Split-Fluorescent Proteins in CRISPR-Engineered C. elegans. In Single Molecule Microscopy in Neurobi- ology, pages 11{37. Springer, 2020. [120] Fabien Pinaud, Xavier Michalet, Gopal Iyer, Emmanuel Margeat, Hsiao-Ping Moore, and Shimon Weiss. Dynamic Partitioning of a Glycosyl-Phosphatidylinositol-Anchored Protein in Glycosphingolipid-Rich Microdomains Imaged by Single-Quantum Dot Tracking. Trac, 10(6):691{ 712, jun 2009. ISSN 13989219. doi: 10.1111/j.1600-0854.2009.00902.x. URL http://doi.wiley.com/10.1111/j.1600-0854.2009.00902.x. [121] G.J. Sch utz, H. Schindler, and T. Schmidt. Single-molecule microscopy on model membranes reveals anomalous diusion. Biophysical Journal, 73(2): 1073{1080, aug 1997. ISSN 0006-3495. doi: 10.1016/S0006-3495(97)78139-6. [122] Gaye F. White, Z. Shi, Liang Shi, Z. Wang, Alice C. Dohnalkova, Matthew J. Marshall, James K. Fredrickson, John M. Zachara, Julea N. Butt, David J. Richardson, and Thomas A. Clarke. Rapid electron exchange between surface-exposed bacterial cytochromes and Fe(III) min- erals. Proceedings of the National Academy of Sciences, 110(16):6346{ 6351, 2013. ISSN 0027-8424. doi: 10.1073/pnas.1220074110. URL http://www.pnas.org/cgi/doi/10.1073/pnas.1220074110. [123] Xiuyun Jiang, Jessica H. van Wonderen, Julea N. Butt, Marcus J. Edwards, Thomas A. Clarke, and Jochen Blumberger. Which Multi-Heme Protein Complex Transfers Electrons More Eciently? Comparing MtrCAB from Shewanella with OmcS from Geobac- ter. The Journal of Physical Chemistry Letters, 11:9421{9425, oct 2020. ISSN 1948-7185. doi: 10.1021/acs.jpclett.0c02842. URL https://pubs.acs.org/doi/10.1021/acs.jpclett.0c02842. [124] N. Michael Green. Avidin and Streptavidin. Methods in Enzymology, 184 (C):51{67, jan 1990. ISSN 15577988. doi: 10.1016/0076-6879(90)84259-J. [125] Mark Howarth, Daniel J-F Chinnapen, Kimberly Gerrow, Pieter C Dor- restein, Melanie R Grandy, Neil L Kelleher, Alaa El-Husseini, and Alice Y Ting. A monovalent streptavidin with a single femtomolar biotin binding site. Nat Methods, 3(4):267{273, 2006. doi: 10.1038/NMETHXXX. 137 [126] M. Nirmal, B. O. Dabbousi, M. G. Bawendi, J. J. Macklin, J. K. Trautman, T. D. Harris, and L. E. Brus. Fluorescence intermittency in single cadmium selenide nanocrystals. Nature, 383(6603):802{804, 1996. ISSN 00280836. doi: 10.1038/383802a0. URL https://www.nature.com/articles/383802a0. [127] Hannah H. Tuson and Julie S. Biteen. Unveiling the inner workings of live bacteria using super-resolution microscopy, jan 2015. ISSN 15206882. URL https://pubs.acs.org/sharingguidelines. [128] Jim E. Horne, David J. Brockwell, and Sheena E. Radford. Role of the lipid bilayer in outer membrane protein folding in Gram-negative bacteria, jul 2020. ISSN 1083351X. [129] James A. Dix and A.S. Verkman. Crowding Eects on Diusion in Solutions and Cells. Annual Review of Biophysics, 37(1):247{263, jun 2008. ISSN 1936-122X. doi: 10.1146/annurev.biophys.37.032807.125824. [130] Daniel A. Phillips, Lori A. Zacharo, Cheri M. Hampton, Grace W. Chong, Anthony P. Malanoski, Lauren Ann Metskas, Shuai Xu, Lina J. Bird, Brian J. Eddie, Grant J. Jensen, Lawrence F. Drummy, Mohamed Y. El-Naggar, and Sarah M. Glaven. A prokaryotic membrane sculpting BAR domain protein, jan 2020. URL https://doi.org/10.1101/2020.01.30.926147. [131] J. Deich, E. M. Judd, H. H. McAdams, and W. E. Moerner. Visualiza- tion of the movement of single histidine kinase molecules in live Caulobac- ter cells. Proceedings of the National Academy of Sciences of the United States of America, 101(45):15921{15926, 2004. ISSN 00278424. doi: 10.1073/pnas.0404200101. [132] Marianne Renner, Yegor Domanov, Fanny Sandrin, Ignacio Izeddin, Patri- cia Bassereau, and Antoine Triller. Lateral Diusion on Tubular Mem- branes: Quantication of Measurements Bias. PLoS ONE, 6(9):e25731, sep 2011. ISSN 1932-6203. doi: 10.1371/journal.pone.0025731. URL https://dx.plos.org/10.1371/journal.pone.0025731. [133] Paolo Pierobon, Sarra Achouri, S ebastien Courty, Alexander R. Dunn, James A. Spudich, Maxime Dahan, and Giovanni Cappello. Velocity, pro- cessivity, and individual steps of single myosin V molecules in live cells. Biophysical Journal, 96(10):4268{4275, may 2009. ISSN 15420086. doi: 10.1016/j.bpj.2009.02.045. [134] Libin Abraham, Henry Y. Lu, Rebeca Cardim Falc~ ao, Joshua Scurll, Tim- othy Jou, Brian Irwin, Reza Tafteh, Michael R. Gold, and Daniel Coombs. Limitations of Qdot labelling compared to directly-conjugated probes for 138 single particle tracking of B cell receptor mobility. Scientic Reports, 7(1): 1{13, dec 2017. ISSN 20452322. doi: 10.1038/s41598-017-11563-9. [135] George R. Heath, Mengqiu Li, Isabelle L. Polignano, Joanna L. Richens, Gianluca Catucci, Paul O'Shea, Sheila J. Sadeghi, Gianfranco Gilardi, Julea N. Butt, and Lars J.C. Jeuken. Layer-by-layer assembly of supported lipid bilayer poly-L-lysine multilayers. Biomacromolecules, 17(1):324{335, 2016. ISSN 15264602. doi: 10.1021/acs.biomac.5b01434. [136] Gaye F. White, Marcus J. Edwards, L. Gomez-Perez, David J. Richard- son, Julea N. Butt, and Thomas A. Clarke. Mechanisms of Bacte- rial Extracellular Electron Exchange. In Advances in Bacterial Electron Transport Systems and Their Regulation, volume 68, chapter 3, pages 87{138. Elsevier Ltd, 2016. doi: 10.1016/bs.ampbs.2016.02.002. URL http://linkinghub.elsevier.com/retrieve/pii/S0065291116300029. [137] Marcus J. Edwards, Nanakow A. Baiden, Alexander Johs, Stephen J. Toman- icek, Liyuan Liang, Liang Shi, James K. Fredrickson, John M. Zachara, Andrew J. Gates, Julea N. Butt, David J. Richardson, and Thomas A. Clarke. The X-ray crystal structure of Shewanella oneidensis OmcA reveals new insight at the microbe-mineral interface. FEBS Letters, 588(10):1886{ 1890, 2014. ISSN 18733468. doi: 10.1016/j.febslet.2014.04.013. URL http://dx.doi.org/10.1016/j.febslet.2014.04.013. [138] Ingrid Chamma, Mathieu Letellier, Corey Butler, B eatrice Tessier, Kok- Hong Lim, Isabel Gauthereau, Daniel Choquet, Jean-Baptiste Sibarita, Shel- don Park, Matthieu Sainlos, and Olivier Thoumine. Mapping the dynamics and nanoscale organization of synaptic adhesion proteins using monomeric streptavidin. Nature Communications, 2016. doi: 10.1038/ncomms10773. URL www.nature.com/naturecommunications. [139] Kristopher A. Hunt, Jerey M. Flynn, Bel en Naranjo, Indraneel D. Shikhare, and Jerey A. Gralnick. Substrate-level phosphorylation is the primary source of energy conservation during anaerobic respiration of Shewanella oneidensis strain MR-1. Journal of Bacteriology, 192(13):3345{3351, 2010. ISSN 00219193. doi: 10.1128/JB.00090-10. [140] Aunica L. Kane, Evan D. Brutinel, Heena Joo, Rebecca Maysonet Sanchez, Chelsey M. VanDrisse, Nicholas J. Kotloski, and Jerey A. Gralnick. Formate Metabolism in Shewanella oneidensis Generates Proton Motive Force and Prevents Growth Without an Electron Acceptor. Journal of Bacteriology, 198(8):JB.00927{15, 2016. ISSN 0021-9193. doi: 10.1128/JB.00927-15. URL http://jb.asm.org/lookup/doi/10.1128/JB.00927-15. 139 [141] Grigoriy E. Pinchuk, Eric A. Hill, Oleg V. Geydebrekht, Jessica de Ingeniis, Xiaolin Zhang, Andrei Osterman, James H. Scott, Saman- tha B. Reed, Margaret F. Romine, Allan E. Konopka, Alexander S. Beliaev, Jim K. Fredrickson, and Jennifer L. Reed. Constraint-based model of Shewanella oneidensis MR-1 metabolism: A tool for data anal- ysis and hypothesis generation. PLoS Computational Biology, 6(6):1{8, jun 2010. ISSN 1553734X. doi: 10.1371/journal.pcbi.1000822. URL https://pubmed.ncbi.nlm.nih.gov/20589080/. [142] Johannes A Van Veen and Eldor A Paul. Conversion of Biovolume Mea- surements of Soil Organisms, Grown Under Various Moisture Tensions, to Biomass and Their Nutrient Content. Applied and Environmental Microbi- ology, 37(4):686{692, 1979. [143] Julia Glazyrina, Eva Maria Materne, Thomas Dreher, Dirk Storm, Ste- fan Junne, Thorsten Adams, Gerhard Greller, and Peter Neubauer. High cell density cultivation and recombinant protein production with Escherichia coli in a rocking-motion-type bioreactor. Microbial Cell Factories, 9, may 2010. ISSN 14752859. doi: 10.1186/1475- 2859-9-42. URL https://pubmed.ncbi.nlm.nih.gov/20509968/ https://pubmed.ncbi.nlm.nih.gov/20509968/?dopt=Abstract. [144] Christina L Lewis, Caelli C Craig, and Andre G Senecal. Mass and Density Measurements of Live and Dead Gram-Negative and Gram-Positive Bacterial Populations. Applied and Environmental Microbiology, 80(12):3622{3631, 2014. doi: 10.1128/AEM.00117-14. [145] Sahand Pirbadian, Marko S. Chavez, and Mohamed Y. El-Naggar. Spa- tiotemporal mapping of bacterial membrane potential responses to extracel- lular electron transfer. Proceedings of the National Academy of Sciences of the United States of America, 117(33):20171{20179, aug 2020. ISSN 10916490. doi: 10.1073/PNAS.2000802117. 140

Abstract (if available)

Abstract Electron transfer is a fundamental aspect of life, driving key biological energy conversion processes such as respiration and photosynthesis. While the mechanisms of biological electron transfer over nanometer length scales are now well established, discoveries of fast microbial electron transfer across micron- to centimeter-scale distances are challenging our state of knowledge and are of particular interest for the purpose of wiring microbes to electrodes in bioelectrochemical renewable energy technologies. Dissimilatory metal-reducing bacteria, including Shewanella oneidensis MR-1, can gain energy by extracellular electron transport (EET) to external solids, such as minerals or electrodes, which substitute for oxygen or other soluble electron acceptors in respiration. In S. oneidensis, a network of multiheme cytochromes facilitate EET by forming biological electron conduits that bridge the otherwise insulating inner membrane, periplasm, and outer membrane to external redox-active surfaces. These microbes also form outer membrane extensions (OMEs), from vesicles to micrometer-scale appendages known as bacterial nanowires that are proposed to aid in their use of external electron acceptors. The cytochromes abundant on the cell outer membrane and important in EET are similarly found along its membrane extensions and give these OMEs their conductivity (at least as measured in dry, chemically fixed conditions). ? These proposed bacterial nanowires, which can be several times the cell length, are thereby thought to extend EET to more distant electron acceptors. However, it was still unclear why these extensions form, and to what extent they contribute to respiration in living cells. In Chapter 2, I investigated physical contributors to their formation using in vivo fluorescence microscopy. While previous studies focused on the display of S. oneidensis OMEs as a response to oxygen limitation, I found that cell-to-surface contact is sufficient to trigger the production of OMEs, including some that reach >100 ?m in length, irrespective of medium composition, agitation, or aeration. To visualize the extent of heme redox centers along OMEs, and help distinguish these structures from other extracellular filaments, I also performed histochemical redox-dependent staining with transmission electron microscopy on wild type and cytochrome-deficient strains. I demonstrated that redox-active components are limited to OMEs and not present on other extracellular appendages, such as pili and flagella. I also observed that the loss of 8 functional periplasmic and outer membrane cytochromes significantly decreased both the frequency and intensity of redox-dependent staining found widespread on OMEs. These results improved our understanding of the environmental conditions that influence the formation of S. oneidensis OMEs, as well as the distribution and functionality of EET components along extracellular appendages. ? While the role of multiheme cytochromes in transporting electrons across the cell wall is well established, these cytochromes were also recently found to facilitate long-distance (micrometer-scale) redox conduction along outer membranes and across multiple cells bridging electrodes. Recent studies proposed that long-distance conduction arises from the interplay of electron hopping and cytochrome diffusion, which allows collisions and electron exchange between cytochromes along membranes. However, the diffusive dynamics of the multiheme cytochromes had never been observed or quantified in vivo, making it difficult to assess their hypothesized contribution to the collision-exchange mechanism. In Chapters 3 and 4, I used targeted quantum dot labeling, total internal reflection fluorescence microscopy, and single-particle tracking to quantify the lateral diffusive dynamics of the outer membrane-associated decaheme cytochromes MtrC and OmcA, two key components of EET in S. oneidensis. I observed confined diffusion behavior for both quantum dot-labeled MtrC and OmcA along cell surfaces (diffusion coefficients D (MtrC) = 0.0306 ? 0.0082 ?m?/s, D (OmcA) = 0.0121 ? 0.0019 ?m?/s) and the membrane extensions thought to function as bacterial nanowires. I found that these dynamics can trace a path for electron transport via overlap of cytochrome trajectories, consistent with our recently proposed long-distance conduction mechanism. The measured dynamics informed kinetic Monte Carlo simulations that combine direct electron hopping and redox molecule diffusion, revealing significant electron transport rates along cells and membrane nanowires.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Bacterial nanowires of Shewanella oneidensis MR-1: electron transport mechanism, composition, and role of multiheme cytochromes

PDF

Kinetic Monte Carlo simulations for electron transport in redox proteins: from single cytochromes to redox networks

PDF

Electronic, electrochemical, and spintronic characterization of bacterial electron transport

PDF

Electrochemical studies of outward and inward extracellular electron transfer by microorganisms from diverse environments

PDF

Electrochemical investigations and imaging tools for understanding extracellular electron transfer in phylogenetically diverse bacteria

PDF

From fuel cells to single cells: electrochemical measurements of direct electron transfer at microbial-electrode interfaces

PDF

From cables to biofilms: electronic and electrochemical characterization of electroactive microbial systems

PDF

Electrochemical studies of subsurface microorganisms

PDF

Survival and evolution of Shewanella oneidensis MR-1: applications for microbial fuel cells

PDF

Extracellular electron transport: Investigating the diversity and mechanisms behind an understudied microbial process with global implications

Asset Metadata

Creator Chong, Grace W. (author)

Core Title From single molecules to bacterial nanowires: functional and dynamic imaging of the extracellular electron transfer network in Shewanella oneidensis MR-1

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Molecular Biology

Degree Conferral Date 2021-08

Publication Date 07/06/2021

Defense Date 06/08/2021

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag bacterial nanowires,cytochromes,diffusion,extracellular electron transfer,OAI-PMH Harvest,redox,Shewanella,single particle tracking

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor El-Naggar, Mohamed Y. (committee chair), Nealson, Kenneth H. (committee member), Pinaud, Fabien (committee member), Finkel, Steven E. (committee member)

Creator Email grace.chong006@gmail.com,gracewch@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC15276671

Unique identifier UC15276671

Identifier etd-ChongGrace-9703.pdf (filename)

Legacy Identifier etd-ChongGrace-9703

Document Type Dissertation

Format application/pdf (imt)

Rights Chong, Grace W.

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Repository Email cisadmin@lib.usc.edu