University of Southern California Dissertations and Theses

Investigations in the field of geobiology through the use of digital holographic microscopy

(USC Thesis Other)

Investigations in the field of geobiology through the use of digital holographic microscopy

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content INVESTIGATIONS IN THE FIELD OF GEOBIOLOGY
THROUGH THE USE OF DIGITAL HOLOGRAPHIC MICROSCOPY
By
Casey R. Barr
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
( Geological Sciences)
December 2020
Copyright 2020 Casey R. Barr
ii

DEDICATION

This thesis was written in a time of great injustice, disparity, indignation, greed, falsehoods and
pain. A time in which science and facts are attacked, questioned and ignored to the detriment
of the larger society for the benefit of a few.
The bright horizon of a better future can be hard to see, but it is there if we are willing.

This thesis is dedicated to all people of America.

iii

ACKNOWLEDGMENTS

This thesis would have been impossible without the support of my colleagues, family and
friends.
Foremost, the support and opportunity provided to me by my advisor Dr. Ken Nealson, whose
unparalleled academic career comes to a close with the defense of this body of work. Ken’s
remarkable enthusiasm for science, life and the wellbeing of others is something I will always
cherish and inspire to embody.
My partner, Jasmine Henderson, is solely responsible for the completion of this thesis. Without
her support, patience, and encouragement over the last six years none of this would have been
possible.
I’d also like to acknowledge the many collaborators and colleagues who have helped inspire me
in times of difficulty, shared in my achievements, and with whom I have spent many hours in
philosophical discussions ranging far beyond research and academia. To my brother and closest
collaborator, Dr. Manual Bedrossian: I am so honored to have spent the last four years working
alongside you, and will forever appreciate the education, assistance, and friendship that you
provided as we navigated graduate school together. Thank you to my laboratory family and
past labmates: Pratixa Savalia, Dr. Bonita Lam, Dr. Nancy Merino, Martin Van-Den-Berghe, Dr.
Annette Rowe, Dr. Roman Barco, Karla Abuyen, Dr. Lina Bird, and my team of undergraduates:
Jaclyn Thompson, Mark Liu, Clarissa Tacto and Michelle Tashjian, whose hard work and
dedication was truly inspiring. Thank you to the members of outside laboratories, and
international colleagues including: Dr. Tingting Yang, Leila Mahrokh, Yubin Raut, Dr. Bingran
Cheng, Dr. Radu Popa, Dr. Gijs Kuenen, Dr. Shino Suzuki, Dr. Ken Lohmann and Dr. Serban Sarbu
among others. Finally, thank you to John Curulli, the director of the microscopy core facility at
USC, whose decision to bring me on as Biological Sciences RA for three years allowed me to
fully embrace my love of microscopy.

iv

TABLE OF CONTENTS
DEDICATION ............................................................................................................................. ii
ACKNOWLEDGEMENTS ....................................................................................................... iii
LIST OF TABLES ..................................................................................................................... vi
LIST OF FIGURES .................................................................................................................. vii
LIST OF VIDEOS………………………………………………………………………………x
ABSTRACT ................................................................................................................................xi

CHAPTER I. Digital Holographic Microscopy: In theory, practice, and
consideration…………………….................................................................................................. 1
Determination and quantification of the MZ-DHM instrument capabilities….............................. 5
Limits of the technology................................................................................................................10
History of the MZ-DHM at the University of Southern California...............................................15
Conclusion……………………………………………………………………………………….17
References .....................................................................................................................................18

CHAPTER II. Development and use of DHM in the investigation of microbe-electrode
interactions……………………………………………………………………………………….19
Overview of chapter ......................................................................................................................19
The development, application and observations of digital holographic microscopy in microbe-
cathode interactions.…..................................................................................................................37
Generalized experimental design……………..……………………….........................................44
Methods ……………………………….........................................................................................48
Results……………………………………………………………………………………………49
Conclusion……………………………………………………………………………………….54
References………………………………………………………………………………………..56
Closing remarks………………………………………………………………………………….58

CHAPTER III. The use of DHM in the study of life at high pH: in the field and the
laboratory………………………………………………………………………………………...60
Overview of chapter……………….……………..………………………………………………60
Geological and geochemical background………..………………………………..61
Specifics of the springs at The Cedars and their microbial communities…………68
Experiment I: In situ enumeration of cell densities……………………………………………...74
Experiment II: Motility characterization of enrichment cultures………………………………..78
Experiment III: Chemotaxis of a Cedars isolate…………………………………………………91
Conclusion…………………………………………………………………………………….....96
References………………………………………………………………………………………..97
Closing remarks……………………………………………………………………………….....99

CHAPTER IV. Investigations into magnetotactic bacteria through DHM……………………101
Overview of chapter……………….……………………………………………………………101
Overview of magnetotactic bacteria……………………………………………………………104
Advantages of a DHM/magnetic coil-based experimental design…………………………...…115
Materials and methods………………………………………………………………………….116
v

Results…………………………………………………………………………………………..121
Discussion………………………………………………………………………………………127
References………………………………………………………………………………………131
Closing remarks and future perspectives.……………………………………………..………..133

BIBLIOGRAPHY ...............................................................................................................…..140
Additional Publications…………………………………………………………………………147

vi

LIST OF TABLES

Table 1.1 Summary of MZ-DHM optical determinations…………………………..…………….5
Table 1.2 Summary of orangeBox optical determinations………………………….……….……9
Table 2.1 Analysis of individual Thioclava electrotropha cells that utilize
the electrode surface…………………………….………………………………………………52

Table 2.2 Spearman Correlation results for full cycle bacteria………………………………….53
Table 2.3 Mean speeds of tracked cells by electopotential and time point……………………...54
Table 3.1 In situ enumeration of cells from various springs at The Cedars…………………….77

Table 3.2 Enrichment culture media composition, sample numbers
and predominant motility characterization used in enrichment study…………………………...81

Table 3.3 Frequency Table of nominal variables with regard to presence
of Brownian motion or non-random movement…………………………………………………82

Table 3.4 Analysis of varience table for motility by carbon source,
base media composition, and additive…………………………………………………………..83

Table 3.5 Estimated particle densities using fringe visibility processing……………………….89

Table 3.6 Number of cells and mean average speeds of the pure culture
chemotaxis experiments……………………………………………………..………………..…94

Table 4.1 Mean speeds of tracked cells and magnetite particles at different
magnetic field intensities………………………………………………………………………122

vii

LIST OF FIGURES

Figure A.1 The microbial world………………………………………………………...………xii

Figure 1.1 Fringes of constructive and destructive interference…………………….……………1

Figure 1.2 Mach Zehnder digital holographic microscope designs………………..……………..4

Figure 1.3 US Air Force Resolution Target……………………………………..…….………….6

Figure 1.4 orangeBox DHM instrument……………………………………………………...…..8

Figure 1.5 Limit of detection quantification of Shewanella oneidensis MR-1
cells via DHM……………………………………………………………………………………16

Figure 2.1 Known mechanisms of reduction-based EET…………………………………….…22

Figure 2.2 Reduction-based EET models…………………………………………………...…..24

Figure 2.3 Scales of biological electron conduction processes………………………………….31

Figure 2.4 Rich media grown Thioclava electrotropha ElOx9…………………………………33

Figure 2.5 Full cell electrochemical cell used in Harris et al. (2010)………………………..….38

Figure 2.6 Screen printed electrode………………..……………………………………………39

Figure 2.7 CAD image of laser cut well assembled on sample chamber…………………..……41

Figure 2.8 Various half-cell sample chamber designs……………………………………….….42

Figure 2.9 Final design of EET sample chamber………………………………………………..45

Figure 2.10 Assembled DHM-EET electrochemical cell on the translational stage
during experimental run………………………………………………………………………….46

Figure 2.11 Chronoamperometry data ……………..…………………………………………...47

Figure 2.12 Tracks of T. electrotropha cells under experimental cathodic
potentials over time………………………………………………………………………………50

Figure 2.13 Mean speeds of all tracked motile cells by electropotential over time……………..53

Figure 3.1 Map of The Cedars peridotite and insert of field site locations……………………...62

viii

Figure 3.2 Geological setting of The Cedars active-terrestrial serpentinization site……...…….63

Figure 3.3 Calcium carbonate structures of The Cedars……………………………………...…67

Figure 3.4 Barnes Spring Complex 5 (BS5)…………………………………………………….69

Figure 3.5 Electron micrographs of Serpentinomonas racichei A1………………………..……70

Figure 3.6 Grotto Pool Spring (GPS1)…...……………………………………………………...71

Figure 3.7 Mortar Bed Spring (MBS)…………………………………………………………...73

Figure 3.8 Experimental design for in situ enumeration of microbial cells………………….….76

Figure 3.9 Various enrichment cultures used in motility survey experiment…………………...79

Figure 3.10 Media recipe for CSM 2 Low and CSM2 High, with descriptions of various
additives used in The Cedars enrichment study……………………….…………………………88

Figure 3.11 Relative abundance of amplicon sequence variants of 16S community composition
of select Brownian motion and motile enrichments………………………………………..……85

Figure 3.12 Tracks of select enrichment cultures……………………………………………….86

Figure 3.13 Mean speeds of bacteria tracked in select motile data sets………………………....87

Figure 3.14 Chemotaxis two well sample chamber for DHM………………………………..…92

Figure 3.15 Assembled z planes of tracked bugs across 500 µm depth and xy tracks of select z
planes………………………………………………...…………………………………………..94

Figure 3.16 Relative frequency of tracked cells binned by mean speed………………………...95

Figure 4.1 Diversity of MTB phylogeny, MAI gene sequences and generalized magnetosome
formation…………………………………………………..……………………………………105

Figure 4.2 Various TEM micrographs of magnetotactic bacteria with visible
magnetosomes…………………………………………………………………………………..107

Figure 4.3 General conception of MTB behavior in aquatic environments………………..…..109

Figure 4.4 Experimental system for magnetotaxis-DHM experiments…………..…………....117

Figure 4.5 Tracks of positively identified MTB exhibiting Northern polarity…………...……121

ix

Figure 4.6 Polar histograms of the directional trajectories of tracked cells and magnetite particles
under each magnetic field regime………………………………………………………………123

Figure 4.7 Taxonomic classification of magnetically enriched samples
via Illumina sequencing……………………………………………..………………………….125

Figure 4.8 Mean speed vs B-Field intensity plot of bacteria (coccoid and non-coccoid) and
abiotic magnetite particles………………………………..…………………………………….126

Figure 4.9 Composite image of A) Northern polarity MTB and B) Southern polarity MTB
during 1080° rotation of a 50 µT magnetic field……………………...………………………..127

Figure 4.10: Theoretical sample chamber design for use in MTB-DHM
enrichment experiments…………………………….…………………………………………..135

x

LIST OF VIDEOS

All videos are available on my personal YouTube account:
https://www.youtube.com/channel/UC1RyFvyOV6EZCCnOt27N4iw/

Scan with phone

Video 2.1 Cathode oxidation by Thioclava electrotropha at the ITO electrode surface. -601 mV
(vs Ag/AgCl) at T= 30 mins

Video 3.1 Experimental set up for in situ cell enumeration experiments

Video 4.1 Biotic (left) and Abiotic (right) samples under null-field (NF) conditions.

Video 4.2 Biotic (left) and Abiotic (right) samples under a 50 µT (0.5G) magnetic field.

Video 4.3 Biotic (left) and Abiotic (right) samples under a 40 µT (0.4G) magnetic field

Video 4.4 Biotic (left) and Abiotic (right) samples under a 30 µT (0.3G) magnetic field

Video 4.5 Biotic (left) and Abiotic (right) samples under a 20 µT (0.2G) magnetic field

Video 4.6 Biotic (left) and Abiotic (right) samples under a 10 µT (0.1G) magnetic field

Video 4.7 Rotating magnetic field (0.5G, 1080°)

Video 4.8 Track of MTB displaying Northern polarity during field rotation

Video 4.9 Track of MTB displaying Southern polarity during field rotation

xi

ABSTRACT

This thesis explores the use of Digital Holographic Microscopy (DHM) to observe and characterize some
of the most fundamental behaviors of various geobiologically relevant bacteria. DHM is a non-invasive
volumetric interferometry-based technique with vast potential in the study of single celled
microorganisms both in the laboratory and in the field. DHM instruments provide the ability to
reconstruct three-dimensional space (while maintaining sub-micron resolution) over time, allowing
microbiological researchers the opportunity to investigate the motile response of individual cells as they
navigate the environment under natural ambient conditions or during experimental runs in which
specific behaviors are elicited and described. As the technological innovation of DHM capabilities rapidly
approaches a lower cost threshold and more user-friendly state, this thesis exists primarily as a
collection of research that explores the development and application of DHM in the field of geobiology
and environmental microbiology.
Chapter I provides a background of the DHM system: in theory, development and practice. Highlighting
both the advantages and hurdles of this up-and-coming technique as of the year 2020 and describing a
pathway to building an inexpensive DHM instrument. Chapter II presents an experimental DHM-
electrochemical cell system developed to study the microscale interactions of bacteria capable of
extracellular electron transport (EET) using poised electrode surfaces, providing insight into the
blossoming field of electromicrobiology. Chapter III demonstrates how the DHM platform can be used to
quantify microbial cell density in an extreme environment (pH ≥ 11) where cell numbers are very low (<
10
3
cells/mL) and laboratory methods for the study of the unique organisms endemic to these systems.
Chapter IV investigates the response of magnetotactic bacteria (MTB) from a mixed environmental
sample at realistic magnetic field strengths through the incorporation of DHM and a tri-axial coil system,
xii

allowing for the identification, quantification and characterization of the magnetotaxis response of
individual cells at different environmentally-relevant field strengths for the first time.
In all of these studies, it is clear that DHM provides a novel view of the microbial world, allowing
researchers to view the three-dimensional environment that these simple lifeforms inhabit, navigate
and interact in. As the artwork of Farooq Azam helped illustrate the nature of microbial ecology and
reality of the microscopic environment, DHM is a tool that allows us to view this world directly and in
real-time (Figure A.1). As much of the microbial sciences has turned toward molecular and -omics
techniques, direct observations of individual cells are fundamental in grounding our perceived and
theorized assumptions of the organisms in nature. This thesis attempts to explore exactly that and
provide a launch pad for DHM technology in the microbial sciences.

Figure A.1 The microbial world. A) Microbial control of oceanic carbon flux: The plot thickens. -Farooq Azam,
1998. B) Composite image of DHM data of magnetotactic bacteria (Chapter IV).

1

Chapter I: Digital Holographic Microscopy: In theory, practice,
and consideration

Digital Holographic Microscopy (DHM) is a volumetric imaging technique that is founded on the theory
of off-axis holography. Through the mechanism of optical interferometry, off-axis holography can be
used to describe both phase and amplitude characteristics of objects within a fluid matrix. By combining
both a sample and reference light beam on the detection surface of a camera in a precise alignment a
natural interference pattern is created as an effect of wave propagation physics. As both beams are
sourced from a single laser diode, and thus identical in wavelength, the interference pattern of the two
beams is the result of a wavefront displacement. When the light beams are recombined on the face of a
detector, the resulting pattern of the displacement is described as an interferogram. By adjusting the
displacement (also known as the phase difference) of the two waves, the spatial frequency of the
interference pattern can be changed, resulting in fringes of constructive and destructive interference
which ultimately dictate spatial resolution (Figure 1.1).

Figure 1.1 Fringes of Constructive and Destructive Interference
2

Numerical reconstruction of the wave front is used to translate the information projected onto the
detection plane (i.e. CCD detector) into a hologram via computer. This digital propagation of information
is what lends DHM the “Digital” moniker. By using a computer interface to apply the numerical
reconstruction, DHM facilitates the ability of real time 3D imaging, with data acquisition only
constrained by the shutter speed of the CCD camera. Reconstruction is achieved through the Angular
Spectrum Method, in which the complex wavefront ( 𝛤𝛤 ) is determined using the convolution theorem
and the optical free space propagation term (G) is described in Equation 1.1.
Eq 1.1 𝛤𝛤 ( 𝑥𝑥 , 𝑦𝑦 , 𝑧𝑧 ) = ℱ
− 1
[ ℱ( 𝐼𝐼 ( 𝑥𝑥 , 𝑦𝑦 ) ∙ 𝑅𝑅 ) ∗ 𝐺𝐺 (𝑥𝑥 , 𝑦𝑦 , 𝑧𝑧 )]
In a gross oversimplification for the purposes of this thesis, the inverse Fourier Transform ( ℱ
− 1
) is taken
of the product of Fouier Transform ( ℱ) of the adjusted hologram term ( 𝐼𝐼 ( 𝑥𝑥 , 𝑦𝑦 ) ∙ 𝑅𝑅 ) and the propagation
term ( 𝐺𝐺 ). The adjusted hologram term is simply the product of the hologram 𝐼𝐼 ( 𝑥𝑥 , 𝑦𝑦 ), and the ‘phase
adjustment term’ ( 𝑅𝑅 , which is used to adjust for tilt and other artifacts in reconstructions). The
propagation term (G) describes the diffraction behavior of light as it is transmitted through a material.
This term underlies the process through which we can reconstruct a given focal plane which gives us the
ability to volumetrically resolve our data sets. This numerical reconstruction is done through the data
acquisition software Koala (Lyncée Tec, Switzerland) or the open source DHMx (built by NASA JPL).
In practice, the theory of digital holography becomes Digital Holographic Microscopy through the
incorporation of infinity-corrected objective lenses into the optical pathway
1
. A monochromatic,
coherent laser (405nm in our system) is split through a 50/50 beam-splitter, creating two identical
beams in a perpendicular orientation. These beams are identical and are redirected 90° via a mirror,
with one beam (the sample beam) passing through the experimental medium and the other (the
reference beam) continuing free of interaction. The beams are then passed through identical, yet

1
Infinity-corrected lenses create a parallel optical path, in which the focal length is essentially infinity.
3

separate infinity corrected objective lenses, which supply the magnification required. The light beams
(now infinity-corrected) are recombined through the use of an additional 50/50 beam-splitter, which
directs the beams towards a CCD camera detector. The two beams are incident at the detector plane
but leave the second beam-splitter at an angle relative to each other to achieve the phase
difference/displacement described in the previous section. An optical diagram and the bench-top
instrument are shown in Figure 1.2B/C. This instrument is an updated and custom design built on the
principles described and demonstrated in Kuhn et al. 2014 (Figure 1.2A).
This optical geometry (known as Mach-Zehnder (MZ-DHM)) was selected specifically for its ability to be
used in complex experimental designs. The spatial separation of the reference and sample beams allows
for a larger working surface area in which to stage experiments and ease of adjustment with an XYZ
translation sample stage. Our system utilizes 50x/NA 0.55 objective lenses (Mitutoyo, Japan), selected
for the increased depth of field provided by this numerical aperture. A summary of the following optical
determinations is provided in Table 1.1 and is discussed in further depth below.
4

Figure 1.2 Mach Zehnder Digital Holographic Microscope designs. A.) Original MZ-DHM instrument from (1). B)
Optical diagram of the MZ-DHM built for USC: BS: 50/50 Beam-splitter, M: Mirror, OL: Objective lens, CCD: CCD
Camera. A 405nm coherent illumination source (violet dashed line) is split into the Sample Beam (dark purple
dashed line) and the Reference Beam (blue dashed line). Both beams are incident at the CCD detector plane, with
the angle created by the top beam-splitter providing the off-axis displacement and the resulting diffraction
pattern. C) The MZ-DHM instrument built for this thesis.

5

Table 1.1 Summary of MZ-DHM optical determinations.
Determination and quantification of the MZ-DHM instrument capabilities
Given the illumination source in the MZ-DHM is a 405nm coherent laser, the theoretical diffraction limit
of lateral resolution in our system can be derived as:
Eq 1.2 𝑟𝑟 𝑥𝑥𝑥𝑥
∝ 𝑓𝑓 𝜆𝜆 ,
Eq 1.3 𝑟𝑟 𝑥𝑥𝑥𝑥
=
1. 22𝜆𝜆 2𝑁𝑁 𝑁𝑁 ,
𝑟𝑟 𝑥𝑥𝑥𝑥
is the diffraction-limited lateral resolution; whereas, 𝑓𝑓 is the F-number of the lens system (ratio of
focal length to diameter of the entrance aperture ( 𝑓𝑓 ≈ (2 𝑁𝑁𝑁𝑁 )
− 1
)). 𝜆𝜆 is the wavelength of the
illumination source (405nm) and NA is the numerical aperture of the lens (0.55 for the MZ-DHM). For
the MZ-DHM instrument I built, the theoretical lateral resolution is ~0.45 µm; however, this has yet to
be confirmed experimentally, as the smallest element of the US Air Force (USAF) Resolution Target used
to confirm lateral resolution is 0.78 µm and was easily resolved (Figure 1.3).
6

Figure 1.3 US Air Force Resolution Target. Confirmation of sub-micron resolution on the MZ-DHM. Image on the
left demonstrates the true 117 µm x 117 µm field of view of MZ-DHM instrument. Insert shows magnified area of
element 9 group 3 of target, brown bar is 0.78 µm. The theoretical lateral resolution is 0.45 µm, the actual number
was not determined as we are easily resolving the smallest target elements.

Using the Air Force Target, we were able to experimentally determine the MZ-DHM to have a true
magnification of ~58x, which in combination of our CCD detector and imaging software can be used to
determine the field of view (Eq 1.4).
Eq 1.4 𝐹𝐹𝐹𝐹 𝐹𝐹 = �
𝑝𝑝𝑝𝑝 𝑥𝑥 𝑝𝑝𝑝𝑝 𝑠𝑠 𝑝𝑝 𝑠𝑠 𝑝𝑝 𝑀𝑀𝑁𝑁𝑀𝑀 � ∗ 𝑁𝑁 𝑝𝑝𝑝𝑝 𝑥𝑥 𝑝𝑝𝑝𝑝
In our system, the CCD camera’s per pixel area is 3.45 µm
2
, and the image resolution (2048*2048 pixels)
gives us a pixel size of 3.45 µm and a 𝑁𝑁 𝑝𝑝𝑝𝑝 𝑥𝑥 𝑝𝑝𝑝𝑝 of 2048. Using the true magnification of 58x, our calculated
FOV is 117 µm
2
for the MZ-DHM instrument. The true magnification is slightly above the manufacturer’s
listing due to the fact that a 50x magnification is based on the standard assumption of a 200mm focal
length common to most industry light microscope systems. As the MZ-DHM does not adhere to this
standard focal length, the magnification must be determined experimentally. The theoretical axial
7

resolution, or the minimal distance resolvable between two particles directly aligned in Z space, can be
derived as:
Eq 1.5 𝑟𝑟 𝑠𝑠 ∝ 𝑓𝑓 2
𝜆𝜆 ,
Eq 1.6 𝑟𝑟 𝑠𝑠 =
𝜆𝜆 𝑁𝑁 𝑁𝑁 2

which is calculated to be 1.3 µm for the MZ-DHM. However, due to the controlled nature in which
biomass and microbial density is determined ahead of time in our experiments, this number is not
reflective of the limits of localization (the ability to determine where a particle is in Z space).
Experimentally, the MZ-DHM was found to have a localization limit of .130 µm, which is the RMS error
associated with axial determination of the point spread function of a sub-micron particle.
As previously defined for DHM instruments (2), the depth of field is considered the axial distance from
the central focal plane at which lateral sub-micron resolution is conserved. The US Air Force Resolution
Target was used to experimentally confirm the distance at which sub-micron resolution is lost. For the
MZ-DHM system, sub-micron lateral resolution is conserved for a depth of 600 µm.
An additional DHM instrument was utilized throughout the work encompassed by this thesis. The first-
generation field instrument (known inhouse as orangeBox) was constructed at Caltech in conjunction
with the Jet Propulsion Laboratory (3). This field instrument operates on the same principles as the MZ-
DHM design but is compressed to decrease total volume with all power, laser and camera components
self-contained and protected from disturbance (Figure 1.4). This compression of the optical pathway
limits the functionality of the instrument platform in comparison to the MZ-DHM, requiring the use of
standardized sample chambers and constraining the number of experimental designs that are
compatible with the field instrument. The incorporation of the large XYZ translational stage of the MZ-
DHM and more ‘open-concept’ optical geometry was designed to allow for a larger experimental
8

platform, on which a variety of experimental designs could be staged and investigated. Designed to
withstand the rigors of field work the orangeBox DHM measures 60 x 25 x 12 cm, weighs approximately
10 kg, is waterproof and has slight positive buoyancy (3). Optical statistics for the orangeBox DHM are
listed in Table 1.2 and were determined using the same methodology as described above. For ease of
identification within this thesis, the terms MZ-DHM and orangeBox are to be used to identify each of the
instruments of use.

Figure 1.4 orangeBox DHM instrument. A) Exposed optics, laser and power hardware. B) Assembled orangeBox
field instrument, arrow points to sample chamber insertion point. From (3). The orangeBox measures 60 x 25 x 12
cm in size and weighs approximately 10 kg.

9

Table 1.2 Summary of orangeBox optical determinations. Adapted from (2).

Although these instruments differ in optical pathway, magnification, and resolution (among other
attributes), they both possess notable conserved advantages of DHM instruments. For example, by
virtue of being an instrument based on holography, DHM instruments allow true volumetric imaging and
reconstruction. This gives researchers the ability to accurately acquire and describe a 3D microscale
environment at submicron resolution. In combination with advances in CCD detector technology, the
ability to observe motile microorganisms at 15 frames per second gives us the remarkable ability to
track high speed (>100 µm/sec) bacteria through three-dimensional space. Digital numerical
reconstruction facilitates the possibility of “digital refocusing”, a term we use to describe the ability to
go through a previously acquired hologram data set and adjust the “focal plane” within the 3D data set
(i.e. reconstruct any Z-plane throughout the data set in a 2D fashion) confirming where each bacterium
exists in XYZ and time. As all acquired data are subject to digital refocusing, the actual focal plane at
time of data acquisition is almost irrelevant within reason. If the displayed focal plane (center of the
hologram) is within the vertical dimension of imaging volume of interest, then any information (cells,
particles, etc.) can be reconstructed and tracked. In practical terms, this ability allows the DHM
instruments to be constructed without the use of moving parts, a huge advantage over most
microscopic imaging technologies. Once the initial alignment of the optical pathlengths is completed and
the anticipated resolution is verified, the instrument can be transported and used continuously without
10

need of realignment or adjustment. The MZ-DHM has been shipped commercially along the west coast
of the United States and the orangeBox has traveled extensively across the globe without issue or need
of attention upon arrival.
Limits of the technology as of mid 2020
As with any new technology, DHM has been plagued by many boundaries and bottlenecks to progress,
and while much has been accomplished and the application of the technology continues to be explored,
a few fundamental hurdles remain to be overcome before this technology can be used by the broader
population of scientists. In this section, I outline some of the issues that presently plague the users of
DHM, and although I am a champion of DHM and its uses, there are many potential pitfalls,
assumptions, and shortcomings of the techniques. In effect, these stumbling blocks to wider use in the
scientific community can be broken into three categories: 1. Experimental design, 2. Data acquisition
and 3. Data storage, processing, and tracking.
Experimental design: By nature of being a volumetric technique, Digital Holographic Microscopy has a
number of limitations that are inherent to the optics system, and thus affect the design of any
experiments run on these instruments. One of the most rapidly apparent to a new user of DHM is the
upper limit of detection. As discussed in the section titled ‘History of the MZ-DHM at the University of
Southern California’, the current DHM instruments are designed to and excel in the exploration of
biologically sparse fluid samples. The series of instruments utilized in this thesis were first envisioned as
early predecessors of instruments for eventual use in astrobiological research off-Earth. Specifically,
DHM is viewed as a potential method in which to assay for the presence of life in the fluids of the icy
moons of Jupiter and Saturn. The ability to resolve and image samples low in cellular density makes
DHM an extremely appropriate technique in this quest, as the physics of holography dictates an upper
bound on what we are able to discern in the data. Given the current thinking that if life were to exist in
11

the hypothesized liquid oceans of Europa or Enceladus, the number of cells accessible by a lander or
orbiter (i.e. near the moon’s icy surface or ejected in a plume) is expected to be quite low. As the
working assumption that tectonic forces are actively present and thus capable of providing the energy
necessary for life (through the use of proposed hydrothermal vent systems), life on these moons would
most likely be at highest concentration lower in the water column. The ability to identify life in low
biomass samples is key to exploring these celestial bodies, as access to deep within the moons’ oceans is
well beyond our abilities for the foreseeable future. As mentioned above, the physics of holography
dictate the limits of how dense a sample is capable of resolving. This is due to an overlapping effect and
a low signal to noise value at high density. Common to all 3D imaging techniques, diffraction patterns
exist and propagate the further the focal plane is from a particle in z space. As more particles are added
to the data set, the signal overlap and noise associated with these additional particles drastically
increases making it impossible to resolve. As a result, laboratory experiments must be run at lower cell
densities (ideally on the scale of ~10
4
cells per ml) which may impact the behavior of the bacteria. An
example of the unintentional effects of low biomass samples in targeted experiments is discussed more
in Chapter II: Development and use of DHM in the investigation of microbe-electrode interactions. For in
situ environmental samples, this upper bound is still a limiting factor, but due to the extreme nature of
our field sites and the harsh conditions under which we are interested in exploring motility, we have not
had to address this issue. The remedy to this limitation is simply diluting samples down to DHM
appropriate numbers (most often confirmed visually). The upper bound of the orangeBox system was
determined experimentally by M. Bedrossian to be ~ 6 x 10
6
cells/mL (M. Bedrossian thesis, 2020). With
a smaller field of view, the MZ-DHM has a theoretically lower upper-bound of cell density but has not
been determined experimentally. Although the effect of an upper bound of detection is not the most
challenging aspect facing DHM-based research, it is an important consideration in the planning of
experiments using this technique.
12

Data Acquisition: The initial DHM instruments were built to use the commercially available image
acquisition software Koala (Lyncée Tec, Switzerland), a $20,000 dollar software package that requires
the use of a USB dongle to access the software suite. Caltech had previously purchased the software
license which came with two dongles, one that facilitated data acquisition and one that only allowed for
data processing. This physical limitation was extremely difficult to work around. Between balancing
multiple users trying to actively run experiments along with a queue of data sets waiting for full
numerical reconstruction and z stack construction, this system became quite a burden and resulted in a
major bottleneck for productive workflow. Additionally, Koala has a known issue of dropping random
frames during data capture, as well as a tendency to record at slightly below the advertised 15 FPS
capture rate. In the case of a dropped frame (i.e. the failure to record an individual holographic time
point during data acquisition) the Koala software package substitutes the missing data by simply
duplicating the previously acquired hologram, resulting in a duplicated frame error. By comparing pixel
values of sequential images to identify repeated frames, in addition to observations of flowing particles
and bacteria, we have determined experimentally that the Koala duplicated frame rate is ~30% of all
acquired data. In practice, this dropped/duplicated frame error translates to upwards of 300 repeated
(and scientifically useless) time points per 1000 holograms recorded. This number, which is as
unbelievable as it sounds, has been shown to be true time and time again, requiring the detection and
removal of all duplicated frames from Koala acquired data sets prior to reconstruction. With three out of
every ten frames having been erroneously recorded twice, this error rate resulted in an unnecessary
demand in both computational power and data storage.
As a solution to these issues, the OWLS team at NASA-JPL established a software team to build out the
opensource DHMx acquisition software (https://github.com/dhm-org/dhm_suite). The software
development was begun in January of 2019, with a beta testing version released in June to USC, Caltech
and Portland State. The experiments being run at USC were used as the primary data for
13

troubleshooting and were used to discover bugs in the software. A major part of my effort in the
Summer and Fall of 2019 was involved with the submission of bug reports and subsequent interactions
with the software team. As of writing this, DHMx is published on GitHub and works for single
wavelength data acquisition, but is still a work in progress with regard to some of the more complicated
DHM goals.
Data storage, computation and tracking: It is generally agreed that the single largest hurdle to the
adoption of DHM in the scientific community is the post-experiment steps of data storage, computation
and tracking. A single raw hologram (i.e. the entire 3D wavefront compressed into a two-dimensional
2048*2048 image at 8-bit resolution) is 4.9MB.
At a data capture rate of 15 FPS, we are acquiring 73.5 MB per second. This means a raw data set of
sixty seconds duration = 4,410 MB or 4.4 GB of data. This is prior to reconstruction in three-dimensions.
Depending on the number of Z planes required (i.e the resolution in Z space), each one of the single time
point holograms is multiplied by this number. In the theoretical case of a 60 second, 15FPS data set that
is reconstructed into 10 z planes (at a z resolution of 10 µm increments) the 4.4 GB of raw data becomes
44.1 GBs of reconstructed data. As we need much finer z resolution and deeper field of view to track
micron sized organisms, we often reconstruct up over 100 different Z planes, resulting in a >100x
increase in data volume. A 4.4GB raw data set can easily become a 440+ GB (almost half a terabyte)
reconstructed file. In my work, we have often taken data sets longer than a minute in length so this
number can easily reach over a TB per reconstructed data set. The solution to this is maintaining an
efficient workflow of reconstructions and tracking. As the raw data can always be reconstructed again
and is orders-of-magnitude smaller than a fully reconstructed dataset, we maintain our collective data
library (and backups) solely in the form of the raw data. Our solution has been to reconstruct and store
between 4-10TB of data at a time, work through the tracking of the cells, save the tracked information
and delete the reconstructed data while preserving the raw data and the quantified information.
14

Another bottle neck in this process is simply the computational demands required in the processing of
the data from the raw hologram to the reconstructed data sets. Using a 3.5GHz, 12 core processor, it
takes an average of 300ms per Z slice per hologram to be produced. In the example of our 60 sec, 100 z-
slice reconstruction, it takes approximately 3 seconds per individual hologram, meaning the entire data
set will take 27,000 seconds or 4.5 hours to produce 1-minute of trackable data. Regretfully, all tracking
and data presented in this thesis is the result of manually tracking the cells through the image
processing program ImageJ (imagej.nih.gov). M. Bedrossian (and I to a lesser extent) spent much time
and effort in the attempt to develop a better tracking algorithm or explore machine learning tracking.
Months of time were spent analyzing my data through different tracking algorithms without a tangible
deliverable. The programs developed by M. Bedrossian would work well for some data sets
(homogenous, low biomass samples in which the bacteria were within a fairly tight z plane volume),
while most other data sets returned useless data after hours of analysis. After spending much of the
2018 year in hopes of finding a better methodology and technique, we returned to manual tracking for
the sake of finishing our PhDs. The method in which we manually track the bacteria in our data sets has
changed a number of times throughout the work in this thesis; however, the general pipeline includes
tracking a bug in the xy plane through a coarse resolution of z planes, assembling this data into a “Track”
file and then using high z resolution to determine the exact z plane in which the bug is located.
Depending on the motile behavior of the bug (how much ‘z space’ it explores) it can take half a day to
track a single bacterium. The result of this is that it requires many hours of work to track all the motile
bugs in a data set, yet the bacteria are tracked with extremely high fidelity and the numbers
represented through analysis are of high quality. It is our hope that a passive tracking system (machine
learning or otherwise) is on the horizon and will allow for DHM to make large bounds into the scientific
community as the tracking of data is tedious and time consuming. This being the case, working through
the data we acquired through 4 years of DHM experiments has been a monumental task and one that is
15

still unfinished as of the writing and defense of this thesis. I am extremely proud and confident in the
data presented within this thesis.
History of the MZ-DHM at the University of Southern California
In November of 2016, I was introduced to Digital Holographic Microscopy and began my years-long
collaboration with Manu Bedrossian (Caltech) and Professor Jay Nadaeu (now at Portland State
University). Our first project resulted in the 2017 JoVE publication: Quantifying microorganisms at low
concentrations using digital holographic microscopy (dhm). This publication (on which I am the second
author) was used as an experimental method to determine the limit of detection using a DHM system.
As the microbiologist on the team, I oversaw the biological aspects of experimental design and carried
out the traditional methods of determining cell concentrations. These methods included: using direct
cell counts, colony forming units, and optical density readings to determine cell density. Utilizing an
electronic syringe pump, saline solutions containing known cell concentrations were pumped through a
sample chamber into the orangeBox. Utilizing the DHM instrument, we were successful in quantifying
Shewanella oneidensis MR-1, Shewanella oneidensis ΔflgM (a non-motile mutant), and Serratia
marcescens cultures down to a cell density of 10
2
cells/ml (4). The ability to accurately determine cell
quantities on the order of 10
2
cells per ml (at a 95% confidence interval) was accomplished using < 10
minutes of experimental data, while higher biomass samples can be accurately described in notably less
time. Experimentally, we were able to determine cell densities of 10
3
using only ~1 minute of data, with
cultures at 10
4
and 10
5
cells/mL concentrations being determined in less than 17 seconds of data (4, 6).
This work demonstrates the ability of DHM technology to provide a quick, high throughput and accurate
methodology in the investigation of low biomass fluids, and moreover, the power of DHM as a tool in
fields of extremophile research, astrobiology, and sparsely populated low energy systems.

16

Figure 1.5 Limit of detection quantification of Shewanella oneidensis MR-1 cells via DHM. A) Blue line is the
theoretical relationship between number of cells in the field of view and the density of cells per ml. B)DHM
observation at MR-1 cells at 8x10
5
cells /ml (110 cells in the FOV) and C) DHM observation at 2x10
5
cells/ ml (27
cells in the FOV). Taken from Bedrossian, Barr, et al. 2017 (4).

Following this work, we began a long series of preliminary experiments utilizing an early MZ-DHM
instrument in the investigation of electrokinesis and the field of microbe-electrode interactions. This
DHM work was not the main focus of my research at the time and was a side project that Manu and I
would spend a week or so on per month on. Based on the designs of Harris (5), M. Bedrossian and I
spent most of 2017 developing new sample chamber designs, collecting data and troubleshooting the
overall experimental design. In April of 2018, M. Bedrossian and I drove to Portland State University to
deliver the early MZ-DHM to Professor Nadaeu’s Lab group and run anode reduction experiments on the
system. Unfortunately, this work was compromised by the lack of a functional anaerobic chamber at
PSU. In response to this, we decided to redesign and build a Mach-Zehnder DHM instrument of our own.
The research and development of the new MZ-DHM instrument was funded through my advisor, Dr. Ken
Nealson (USC) and began in earnest in late April 2018. Although the theoretical design of the MZ-DHM is
fairly simple, the development, implementation and build out process was not. From start to finish it
took over nine months. The largest hurdle to overcome was determining which objective lenses to use in
the system, and the restrictions imposed by this decision.
17

Considerations and variables we had to explore included: the impact of coated lenses, objective lens
focal lengths, magnification, and numerical aperture among others. Finding identical (but economic)
lenses that provided the ability to do Digital Holographic Microscopy and gave reasonable field of view
and depth of field determinations was a challenge. Some of the nicest and most expensive lenses we
tried (> $20,000) were not acceptable. We had multiple demo sessions with several different lens
manufacturers throughout the summer of 2018. In the end we found the Mitutoyo Plan Apo 50x NA
0.55 infinity corrected long WD objective lenses (~$2,500 each) provided us the best results and an
acceptable balance of Field-of-View (FOV) and Depth of Field. Once the correct lenses were verified to
work, purchased and implemented, we were free to customize the MZ-DHM optics system to those
lenses, to get the most out of our system as possible. The ultimate result of this labor is the MZ-DHM
built in the Nealson laboratory, but the experience also provided me a crash course in optical
engineering and holography. This instrument was designed and built solely by M. Bedrossian and myself.
We are both extremely proud of this instrument. The general design has since been replicated and
utilized both at Portland State University and NASA-JPL.
Conclusion
As I hope is evident by the existence of this thesis, and the work therein reported, digital holographic
microscopy is a rugged yet unique technique. Capable of providing direct observations of interactions
and behaviors of microorganisms both in the field and in the laboratory setting, DHM can be applied in
many broad fields of research. Although this thesis focuses on the application of DHM in geobiological
investigations, DHM aided research in fields as diverse as food sciences, microbial ecology, and medical
microbiology could easily be envisioned. The ability to observe, identify and track microorganisms in
three-dimensional space at real-time temporal resolution has innumerable applications. Although digital
holographic microscopy will require additional work and technological advances before it can be more
readily used by the larger scientific community, we have brought the technique from proof of concept to
18

targeted application. In a diverse set of investigations related to microbiology, geobiology and
astrobiology, this thesis, in conjunction with that of Manu Bedrossian’s, represents the first steps of an
emerging field of research with immense potential. It has been an honor, privilege, and at times
exceedingly frustrating to work at the forefront of a blossoming new technique and explore its
applications in its infancy.
The work described in this thesis is just some of the directions I explored with DHM.

References:
1. J. Kühn, B. Niraula, K. Liewer, J. K. Wallace, E. Serabyn, E. Graff, J. L. Nadeau, J. Kühn, B. Niraula, K. Liewer,
J. K. Wallace, E. Serabyn, E. Graff, C. Lindensmith, J. L. Nadeau, A Mach-Zender digital holographic
microscope with sub-micrometer resolution for imaging and tracking of marine micro-organisms A Mach-
Zender digital holographic microscope with sub-micrometer resolution for imaging and tracking of marine
micro-organisms. Rev. Sci. Instrum. 85 (2014), doi:10.1063/1.4904449.
2. J. K. Wallace, S. Rider, E. Serabyn, J. Kühn, K. Liewer, J. Deming, G. Showalter, C. Lindensmith, J. Nadeau,
Robust, compact implementation of an off-axis digital holographic microscope. Opt. Express. 23, 17367
(2015).
3. C. A. Lindensmith, S. Rider, M. Bedrossian, J. K. Wallace, E. Serabyn, G. M. Showalter, J. W. Deming, J. L.
Nadeau, A submersible, off-axis holographic microscope for detection of microbial motility and
morphology in aqueous and icy environments. PLoS One. 11 (2016), doi:10.1371/journal.pone.0147700.
4. M. Bedrossian, C. Barr, C. A. Lindensmith, K. Nealson, J. L. Nadeau, Quantifying microorganisms at low
concentrations using digital holographic microscopy (DHM). J. Vis. Exp. 2017, 1–11 (2017).
5. H. W. Harris, M. Y. El-Naggar, O. Bretschger, M. J. Ward, M. F. Romine, A. Y. Obraztsovad, K. H. Nealson,
Electrokinesis is a microbial behavior that requires extracellular electron transport. Proc. Natl. Acad. Sci. U.
S. A. 107, 326–331 (2010).
6. Bedrossian, Manuel M. (2020) A Novel Digital Holographic Microscope (DHM) to Investigate and
Characterize Microbial Motility in Extreme Aquatic Environments. Dissertation (Ph.D.), California Institute
of Technology. doi:10.7907/m3a3-4610

19

Chapter II: Development and use of DHM in the investigation of
microbe-electrode interactions

Overview of chapter
In the decades since the first examples of what would later be known as extracellular electron transfer
(EET) was first reported, hundreds (if not thousands) of experiments, observations, and reviews have
been published on the topic. The vast majority of this work has primarily focused on the metal-reducing
bacteria that belong to the Shewanella or Geobacter genera. In the mid 2010’s, experiments and
enrichment studies of mineral oxidizing bacteria started to explore the other side of EET-facilitated
redox reactions, many of which were accomplished by largely skipping over the earlier mineral-based
experiments found in the metal reducing literature for the more experimentally reproducible and
quantitatively cleaner poised electrode experiments common to most modern EET research. The
experimental design and resulting experiments reported within this chapter reflect a multiyear effort to
incorporate DHM into the field of EET research to provide a methodology for the direct observation and
quantification of microbe-electrode interactions. Although much work, effort and data is not reported
here (specifically a series of anode reduction experiments) this chapter demonstrates the ability to use
DHM to track and analyze the behavior of individual cathode-oxidizing bacteria in three-dimensional
space. Many poised electrode experiments in EET research focus on the output of these interactions
(current production or electron uptake); however, this work is designed to examine the interactions
themselves, specifically looking at sparse density samples in the first few hours of cathode oxidation
experiments. In addition to the fact that these experiments represent the first instance in which DHM
technology is applied to the growing field of electromicrobiology, they also report the first direct
observation of active cathode oxidation by an EET-active bacterium. This chapter includes a brief
20

overview of EET, important developments in the history of this research, and some of the important
caveats of electrode-based experiments (pgs 20-36) followed by a manuscript section in which the
poised electrode-DHM experimental design is discussed and cathode oxidation experiments are
reported (pgs 37-56).
Extracellular Electron Transport
Striking down one the easily accepted assumptions of microbial sciences (that all prokaryotic life is
dependent solely on the physics of diffusion) extracellular electron transport (EET) provides bacteria the
ability to utilize the physics of conduction. EET is the umbrella term used for a number of mechanisms
(some still undescribed) in which bacteria are capable of conducting electrons across their cell
membrane for metabolic use. EET is fundamentally an extension of the electron transport chain beyond
the cytoplasmic membrane and into the outer environment. The initial observations in the 1980s, that
led to the eventual discovery of EET, focused on the geochemical cycling of manganese and the rapid
production and removal of Mn-oxides in aquatic systems (1–3). With the enrichment and cultivation of
novel bacteria capable of transforming insoluble Mn-oxides into soluble Mn
2+
via metabolic reduction,
the biological underpinnings that helped explain the Mn flux issue identified by geochemists started to
become elucidated (4). Initially discovered and investigated in metal reducing model organisms such as
Shewanella oneidensis and Geobacter metallireducens in the late 1980’s, the foundation of EET research
focused primarily on Fe and Mn reduction (4, 5). In the past decades, much has been discovered through
studies of these model organisms in the laboratory and the isolation of other EET-active species. In the
Shewanella and Geobacter models, much has been reported with regard to the genetic basis of EET,
gene regulation, reduction rates, reduction capabilities and electron flow models. As is evident by the
extensive work produced by the research groups led by Ken Nealson and Derek Lovley (and the
prevalence and continued work of their academic offspring over the last 40+ years), the field of EET-
related research has expanded dramatically in many directions. From bioremediation applications to the
21

development of microbial fuel cells and electrode cultivation methods, the fundamental observation
that bacteria were capable of utilizing extracellular minerals for metabolic gain has blossomed into a
field known as electromicrobiology (6).
Originally, extracellular electron transfer was observed to occur only in dissimilatory metal oxide
reducing bacteria (where an insoluble mineral is used as a terminal electron acceptor), in which the
direction of electron flow is towards the external environment. This notion is no longer true, as EET-
active metal oxidizers have now been enriched, isolated, characterized and investigated. That said, the
recent move into oxidation based EET research will undoubtedly teach us much about EET more broadly.
However, at this time the foundations of our understanding of EET processes and EET-capable organisms
have been accomplished predominately through the Shewanella and Geobacter models.
As stated above, reduction-based EET provides a mechanism in which electrons can be conducted from
the electron transport chain to the external environment, allowing the bacteria to utilize terminal
electron acceptors that are not able to be diffused into the interior of the cell (and previously thought to
be biologically unavailable). Through the study of these model organisms, three methods in which cells
can conduct e
-
beyond the confines of the cell membrane have been determined. Utilizing multihaem c-
type cytochromes in all three cases, the reported categories of reduction-based EET electron flow
include: (1) Direct conduction via outermembrane cytochromes, (2) electron conduction along
outermembrane appendages and (3) the production of diffusion-based mediators (Figure 2.1) (7–10).
22

Figure 2.1 Known mechanisms of reduction-based EET. A.) Direct EET via outermembrane cytochromes. B.) Direct
EET via outermembrane extensions impregnated with cytochromes (nanowires). C.) Indirect EET through diffusion-
based co-factors and mediators (flavins).

With over 100 strains in culture and more than 40 described species, the members of the genus
Shewanella comprise one of the most metabolically diverse groups of microorganisms known to science
(11). Identified and isolated from marine and freshwater environments across the globe, the
shewanellae are motile facultative anaerobes belonging to the Deltaproteobacteria. Naturally found in
environments ranging from food spoilage (the first Shewanella species was identified from spoiled
butter), oil field waste sites, and redox gradient dominated systems of stratified water columns and
sediments world-wide (12). The shewanellae have been studied extensively for their ability to use an
astonishing number of terminal electron acceptors that include organic and inorganic compounds,
insoluble minerals, and toxic elements. As a result, the shewanellae have been used to explore
bioremediation and biotechnological applications that include wastewater treatment, radionuclides
(such as urananium VI, technetium VII) and toxic molecules such as chromium (VI) (11). The strain
Shewanella oneidensis MR-1 (formerly classified as Alteromonas putrefaciens MR-1), was first described
in 1988 by Myers and Nealson and was one of the first environmental bacterial species to have its full
23

genome sequenced (4, 13). Although many of the shewanellae are free-living organisms, syntrophic,
epibiont and pathogenic species have also been isolated. Different EET-active Shewanella species have
been shown to have different affinities and redox/chemoattractant sensing capabilities for various
forms of iron and manganese oxides, underscoring the impact of evolutionary history on the cells
preferred insoluble electron sink (14). Given the widespread abundance and diversity of natural
environments that members of the shewanellae have been isolated from, an environmental systematic
approach to correlating habitat parameters and genetic composition has shown that free-living, more
metabolically-diverse, species exhibit a much higher chemoreceptor diversity and environmental sensing
ability than the more metabolically-specialized species (12). This finding is not surprising but reinforces
the applicability of Shewanella species in investigations of signal transduction pathways, sensing
proteins and chemotaxis experiments.
Found in soil and aquatic environments, the members of the genus Geobacter belong to the
Deltaproteobacteria and are often dominant members of anaerobic environments. The Geobacter
species are capable of coupling acetate oxidation (which is produced in large quantities by fermentative
bacteria) and insoluble mineral reduction (commonly Fe(III) oxides ). Long believed to be strict
anaerobes, Geobacter sulfurreducens was shown to be able to use O2 as a terminal electron acceptor at
low concentration, likely a survival strategy for the complex chemical landscapes of soil and sediment
environments (16). Geobacter species have been described as using conductive pili to achieve longer
distance EET conduction, similar in function to the nanowire membrane extensions of Shewanella
species (17). As the shewanellae are predominantly facultative aerobes, and the geobacteraceae are
largely oxygen sensitive or intolerant, it is not surprising that the two main taxonomic clades of known
EET organisms are rarely found in the same environment, potentially providing insight into the
evolutionary history and selection pressures on the development of EET in these two lineages.
24

A reoccurring theme of the Shewanella and Geobacter models of EET is the use of multi-heme
cytochromes as an electron conducting metalloprotein. Genomic analysis of G. sulfurreducens and S.
oneidensis have revealed an extensive number of multi-heme cytochromes located throughout the
inner membrane, periplasm and outer-membrane locales (18). Both organisms utilize a variety of
cytochromes that exhibit a range of redox potentials allowing for the directional conduction of
electrons; however, the EET pathways of Shewanella and Geobacter are distinct. Shewanella conducts
electrons from the inner-membrane (CymA) dehydrogenase to periplasmic cytochromes (STC and FccA).
The periplasmic cytochromes in turn conduct electrons to the MtrABC porin-cytochrome structure
located within the outer membrane. The MtrC and OmcA cytochromes are the location of active
external reduction and used in all three of the aforementioned “traditional EET models”. The Geobacter
model is more complex as it utilizes a series of parallel homolog porin-cytochrome enzymatic structures
in which the regulation of electron flow is not well understood (19). Figure 2.2 illustrates a simplified
version of the known EET pathways in the Shewanella and Geobacter models.

Figure 2.2 Reduction-based EET models. A.) Shewanella MR-1 Mtr Pathway and B.) Geobacter Pcc Pathway (based
on Shi, 2016(19)).

25

One of the largest pivot points (and launch pads of many scientific careers) was the realization and
application of a poised electrode to substitute for the redox potential of the insoluble metal oxides. As
both Fe and Mn-oxides are notoriously difficult and dirty minerals to reproducibly synthesize in the
laboratory, the ability to substitute heterogenous minerals for a clean reproducible electrode surface
unlocked the ability to quantify electron flow rates, electrogenic capabilities, and the development of
microbial fuel cell systems (MFCs) (20). The ability to design more sensitive techniques and methods to
investigate EET reduction via MFC ushered in a new era of research, in which many of the intracellular
processes that contribute to EET were identified.
The double-edged sword of microbe-electrode investigations
Through the use of bioelectrode MFC systems in the laboratory setting, we are able to experimentally
manipulate and quantify redox surface-microbe interactions in ways that are impossible to do using the
redox-active minerals that electrogenic and electrotrophic bacteria utilize in nature. The use of
electrodes can answer more pointed questions (i.e. how long does it take for bacteria X to respond to a
shift in potential, at what potentials are microbe Y most active, studies into the mechanisms of EET,
etc.); which are valuable, but perhaps not totally reflective of how these bacteria behave in the natural
setting.
For instance: traditional electrochemical cell systems utilize 20mM lactate as an electron donor, a simple
low sulfate media and cell concentrations that can be on the order of 10
8
/mL; whereas, the environment
has a large suite of different organic matter compounds that may be utilized, much more complex and
reactive chemistries, and individual species concentrations well below 10
8
per cm
3
. In addition to these
changes, perhaps one of the largest and most important to note is the difference between the redox
potentials we poise on the working electrode surfaces, and the energetic redox potentials the cells
actually utilize during EET in nature. In essence, we are taking a “best guess”, fully aware that the
conditions we impose are not necessarily reflective of those in which the electrogenic bacteria truly exist
26

(cell number, oxide heterogeneity, microenvironments, biofilms, et.), but understanding the pitfalls of
these choices allows us to make note of the ecological implications of the choices we make. Under
standard conditions (25 °C, 1 atm, 1M concentrations, pH =7) the conversion of Gibbs free energy (ΔG°)
to electropotential (E°
cell
) is straightforward using the equation:
Eq. 2.1 ΔG° = -nFE°
cell

where n is the number of electrons and F is the Farrady’s constant. However, when trying to determine
what electropotential to poise an electrode at to mimic an iron or manganese-oxide under conditions
more reflective of the natural environment, this gets much more complex.
One of the largest assumptions in our conversion of the theoretical ΔG
r
to an electropotential is the
mineral form of the insoluble terminal electron acceptor. Although a number of manganese oxides share
the chemical formula MnO
2
, empirical evidence has shown that variations in crystallinity and hydration
impact the rate at which EET-active bacteria are capable of reducing the minerals (21). Both Mn and Fe
minerals are found in a variety of different structural and adsorbed compound capacitance types. In
nature, we know that amorphous minerals (typically biologically produced) are vastly preferred over the
highly crystalline (typically abiotically formed) mineral types when used as a terminal electron acceptor.
Due to the immense variability of mineral types, geometries and associated compounds, it is impossible
to accurately determine the Gibbs free energy potential that is being acquired “in situ”. Adding to this
complexity, is the microscale environment that is occurring at the mineral surface during mineral
reduction. As electrons are being donated to the mineral, reduced forms of the mineral are being
produced, along with protons. This increase in proton and free ion concentration impacts the pH and
chemistry of the system which can alter the assumptions of a simple thermodynamic model. Likewise,
the impact of surface area is an additional consideration that we tend to ignore at this point in
bioelectrode research. Studies of electrogenic bacteria have shown varied preferences in the mineral
27

form of insoluble terminal electron acceptors, although generally mineral reducing bacteria appear to
prefer biologically produced high surface area, amorphous minerals over highly crystallized abiologically
produced mineral forms. Put simply, bacteria tend to prefer smaller mineral grain sizes, and in our
electrode systems, we give the bacteria a giant surface area (>1 cm
2
). Through the experiments in this
chapter, my personal mantra of “ the geometry of the system dictates the response” was repeatedly
reinforced, and fits into the larger biological and evolutionary trend of structure dictating function (be it
the ecological diversity of an environment, the function of enzymes, or the motility structures of
microbes).
Given these reservations, based on the fundamental differences between in situ mineral reduction and
the reduction of steady state electrodes in the laboratory, the use of poised electropotentials to
investigate EET is still an extremely valuable technique, especially with regard to dissecting the
mechanism(s) of EET. For example, electrochemical experiments provide the ability to reproducibly
measure and quantify the flow of electrons, investigate how environmental parameters (pH, salinity,
etc) affect EET rates, and to assay mutants to understand the genetic basis of EET. Electrochemical
studies utilize a range of electropotentials which are nominally characteristic of natural redox potentials
of different insoluble TEA minerals. For Mn-oxide reduction, electrodes are poised between +400 to
+600 mV; whereas, for Fe-oxide reduction, they are in the +100 to +300 mV (vs. Ag/AgCl) range.
Although these values are quite broad, due to the complexities of the natural environment in variables
such as chemistry, geochemical conditions, mineral forms, etc., this approach is not optimal in
describing specific instances within the large range of environments in which insoluble electron
acceptors may be used. However, it does allow us to experimentally manipulate the system and gives
insight into the microbe- redox surface interactions that would be impossible to do using minerals in the
laboratory. This tradeoff is found throughout all microbiology research, as more evidence suggests the
conditions from the laboratory are rarely applicable to the natural setting (ex: maintenance energy
28

values, respiration/substrate utilization rates, even growth phase data doesn’t necessarily reflect the
natural environment).
That said, electrochemical experiments are still valuable, and can teach us much about the specific
organisms we study in the laboratory and their mechanisms of EET. Electrochemical techniques such as
cyclic voltammetry allow us to build redox profiles of EET organisms and give insight into the
mechanisms of electron transfer during reduction and oxidation of poised electrodes. Determining the
midpoint potentials of redox active enzymes and compounds utilized by various EET organisms allows us
to infer important EET electropotentials in which electron flow is maximized. Chronoamperometry is
used to measure current production or electron uptake over time, allowing us to view the EET response
of our experimental organisms and investigations into the nature of direct and indirect electron transfer
techniques.
An unexpected, but important, discovery during the MFC era was that some bacteria were capable of
electrotrophic EET (in which electron flow is initiated by the oxidation of minerals or cathodes external
to the cell with electron flow being conducted inwards into the cell body) and that some bacteria such
as Shewanella are capable of both reduction and oxidation-based EET. Although the majority of work
has focused on the species’ ability to achieve EET-based anode reduction, S. oneidensis MR-1 cells have
also been determined to facilitate inward electron flow via cathode oxidation when a previously
established anode associated biofilm is subjected to cathodic electropotentials (-303 mV vs SHE) under
aerobic conditions. Although inward directed electron flow was shown to be metabolically linked to O 2
reduction and provided the ability to produce ATP and the generation of cellular reducing equivalents
(NAD(P)H and FMN), no net increases in biomass were observed (22). Compared to an open circuit
control that demonstrated a 17% cell loss/day, cathode oxidizing, O
2
reducing treatments showed no
discernible loss of cells. These findings were made during my early PhD work, leaving me with the
impression that many cathode-oxidizing bacteria may utilize insoluble minerals (or cathodes) as a
29

method of maintaining a metabolically stable state (i.e. non-growth related energy gain), yet does not
provide the ability for the cells to actually produce additional biomass (i.e. survival as opposed to
growth). This ability would provide an advantage to microbes that live in complex sedimentary systems
in which pockets of complex chemical and physical conditions (e.g. O
2
concentration, pH gradients,
soluble TEAs, etc.) are abundant and transient.
The mechanisms, diversity and ecophysiology of oxidation-based EET bacteria are in the early phases of
exploration, with many novel cathode-oxidizing bacteria being enriched for (using poised electrodes)
and isolated. Through the pioneering work of Annette Rowe, the electrode-cultivation method of
enriching and isolating novel EET organisms has dramatically underscored the diversity of EET-active
bacteria in the environment (23). Observations of EET-based syntrophies, EET-bacteria that only utilize
soluble mediator molecules (i.e. Pseudomonas aeruginosa), and the presence of the colonial cable
bacteria in stratified marine sediments bring into focus how little is known about EET outside of the
Shewanella and Geobacter models (24–28).

Advantages of EET
With regard to both microbial metabolism and ecology, EET provides a number of innovative solutions
for survival. EET allows bacteria to utilize substrates beyond those that are concentrated within the cell
body, providing an ability to continuously allow electron flow (and ultimately, membrane “charging” and
ATP production) in environments that they may not be able to survive solely through diffusion-based
metabolism. For fermentative organisms, EET provides a mechanism in which electron flow is
maintained, and NAD
+
is regenerated without the production and accumulation of hydrogen or other
fermentation products. Ecologically, the ability to conduct electrons between the outer environment
and the cell body allows EET-active organisms a number of advantages.
30

The first, most obvious, ecological benefit is the ability to utilize substrates in the external environment
that are biologically unavailable to other organisms. By increasing the diversity of potential electron
acceptors and donors beyond that of ‘more traditional’ microbial metabolisms, EET-active bacteria gain
a direct advantage. This ability is not only limited to insoluble electron sources and sinks, but has been
shown to also be used in the reduction of soluble compounds such as DMSO that readily adsorbs onto
surfaces such as silicates and is biologically unavailable to diffusion based metabolisms, or compounds
that become insoluble upon reduction (26). The advantage of using a single mechanism (i.e. EET) to
unlock numerous ‘novel’ electron donors and acceptors may be why EET bacteria tend to exhibit a wide
range of metabolic capabilities and metabolic pathways.
A second, directly related advantage of EET-based metabolism is the competitive advantage of speed.
The near instantaneous conduction of an electron is orders of magnitude faster than the rates of
diffusion non-EET bacteria must rely on. By relying on a conduction-based metabolism, the ability to
rapidly increase the proton motive force gradient within a cell is a massive advantage over non-
electrochemically active lifeforms. The rate of electron flow (and thus ATP production) via EET would be
tied to a single diffusion rate (assuming EET is utilized in either the oxidation or reduction half reaction
but not both); whereas, a metabolism that is reliant solely on diffusion of both the electron donor and
acceptor into the cell is throttled by the availability of both compounds. Given that the rate of electron
conduction is ultimately throttled by the diffusion of protons via the electron-transport chain, and the
inherent biases of laboratory-based EET research (high biomass, rich media grown cells, a homogenous
chemical environment, etc.), the true advantage of utilizing the speed of conduction in nature is difficult
to constrain.
Another physical hurdle that EET provides an ability to overcome is distance. The cable bacteria studied
most famously by the Danish research group led by Lars Peter Nielsen allows for the coupling of redox
pairs (specifically H 2
S oxidation with O
2
reduction) at >10’s of millimeters distance across the oxic/anoxic
31

interface. Cable bacteria form filamentous assemblages of cells (sometimes numbering > 10
4
cells per
filament) allowing electron conduction across distances that are 100’s to 1000’s of times longer than the
nominal cell body length. Belonging to a monophyletic clade within the Desulfobulbaceae family, cable
bacteria are often found in dense networks where the average distance between individual filaments is
on the order of 90 µm (29, 30). Essentially, EET allows cable bacteria to accomplish a reduction-
oxidation reaction in which the half reactions are separated by distances of 10’s to 100’s of millimeters,
tapping into a more energetically favorable TEA, that is beyond the physical reach of any other bacterial
species (Figure 2.3)

Figure 2.3 Scales of biological electron conduction processes. A.) Electron conduction along the electron transport
chain within the cell membrane. B.) Insoluble MnO
2
reduction by a single bacterium. C.) Electron conduction within
a cable bacteria filament. Based on Meysman, 2018 (29).

One of the larger lessons learned from the discovery of the cable bacteria, was the ability for cells to use
one another as sources and sinks of electrons. The electrosyntrophic relationship between metabolic
32

partners has been found in inter-domain methanotrophic consortia (i.e. between anaerobic
methanotrophic archaea and sulfate-reducing bacteria) and between syntrophic anaerobic bacteria (24,
31). The direct interaction, electron conduction, and potential communication between sister cells of a
single species was a paradigm shift in itself, the ability to do this across species and domains opens up a
huge question into how widespread the diversity of EET truly is (28). If the conduction of electrons is so
advantageous ecologically, we would expect to see EET to be widespread among various prokaryotic
lineages. As more and more EET-capable organisms are discovered and investigated, additional variants
on the “classical” EET models of Shewanella and Geobacter have been observed. Observations such as
the discovery of EET-active Gram-positive bacteria, and the use of pigments and nanoparticles of semi-
conductive oxide minerals to aid electron flow, has resulted in a more inclusive and complex description
of EET (6, 32, 33).
Although the majority of research in the past few decades has predominately focused on EET via
Shewanella and Geobacter species, as technology, interest and the field of electromicrobiology has
become more accepted, there remain many unknowns. As mentioned above, the first mechanisms of
cathodic or oxidative-EET are just now being explored. Through cathode-enrichment experiments a
number of pure oxidative-EET organisms have been isolated for further study and characterization. The
use of poised cathodes to enrich for and cultivate EET-active bacteria from marine sediments was shown
to be a novel cultivation technique by Rowe et al, 2015. This pioneering work led to a number of
publications from the Nealson lab on which I am a coauthor. The enrichment technique used in Rowe et
al. (23) and Lam, Barr, Rowe and Nealson (34) can be briefly described as the following: shallow marine
sediments were collected and sieved to remove invertebrates, debris and large grain (>200 µm) sized
rocks. The filtered and homogenized sediments were placed into glass aquaria that was used to replicate
the marine sediment environment. Electrodes were buried 3-5 centimeters below the surface water
interface, at a depth that would be below the oxic/anoxic. Oxygenated filtered seawater was constantly
33

supplied to the experimental aquaria and after a few weeks in which the natural metabolic gradients
inherent to marine sediments could be established. A potentiostat was used to poise specific cathodic
electropotentials upon the surface of the working electrodes. Current production was monitored for a
number of months, until electrodes were harvested to take into the laboratory for further isolation.
Secondary isolation was accomplished in anaerobic chambers using a second poised electrode and
soluble TEAs. Additional enrichment was accomplished using traditional electron donors and acceptors,
with final isolation of individual colonies on solid media. The thesis of Bonita Lam explores the impact of
cathodic potential on the taxonomy of isolated electrotrophic bacteria, and the isolation of novel EET-
active bacterial groups (35). As the electropotential is mimicking a redox potential, it is not surprising
that metabolic capability was closely associated with enrichment potential.
One of the first and most hardy cathode-enriched strains to be isolated was a Thioclava species now
described as Thioclava electrotropha ElOx9. This thiosulfate oxidizing bacterium was enriched and
isolated from a cathode poised at -203 mV (vs SHE) using marine sediments from Catalina Island,
California. The species is capable of oxidizing insoluble elemental S
0
via EET. A chemolithoautotroph, T.
electrotropha is also capable of organoheterotrophy (36). I am a coauthor on the characterization and
species description publication (see additional publications) which was the senior thesis of USC
undergraduate Rachel Chang.

Figure 2.4 Rich Media Grown Thioclava electrotropha ElOx9. Electron micrographs taken by C.Barr.
34

The Thioclava electrotropha species has garnered much interest in the EET-research community, and has
been further electrochemically described by Karbelkar, et. al (2019). Ongoing investigations by the Rowe
research group at the University of Cincinnati have focused on elevating T. electrotropha ElOx9 towards
becoming a model organism in oxidative EET studies. When utilizing a cathode poised at -278 mV vs SHE
as an e
-
D coupled with nitrate reduction, ElOx9 was determined to yield an electron uptake rate of
~1.4x10
4
e
-
/sec/cell, and a Coulombic efficiency of 43.9% (37). Although this electron uptake rate is ~1-2
orders of magnitude less than the electron donation rates seen in anode reduction experiments with S.
oneidensis MR-1, the Coulombic efficiency is similar to other sulfur-oxidizing bacteria biocathodes (Pous,
2014). As the number of cathode-active EET bacteria that are in culture and under current investigation
has increased in the last few years, a few early trends appear to be coming into focus. Models of
oxidative-EET mechanisms include: oxidation via direct contact with redox proteins, electron uptake
through small diffusible molecular mediators, and enzymatic mediators that are endogenously produced
yet attached to the electrode surface (37–39). In the case of T. electrotropha ElOx9, electrochemical
studies suggest that EET in this organism is primarily through direct contact with a formal midpoint
potential of -94 mV vs SHE. ElOx9 does not appear to produce any soluble redox mediators, as
electrochemical analysis of spent media did not elicit an increase in electron uptake (37).
Given the numerous lines of EET-related research currently being pursued, there remains a number of
fundamental questions and observations that need to be made. As current production (or electron
uptake) in electrochemical cell experiments tends to be the most sought-after quantifiable data point,
most electrochemistry experiments require high biomass quantities and multi-day experiments.
Although these experiments can teach us much about the nature of long-term EET metabolisms in a
stable environment, they do not necessarily reflect the microscale interactions that occur in nature
between EET-active microbes and insoluble electron sources/sinks. Much like the fact that the true role
of flavins as a redox mediator molecule has been hotly debated, the observations of laboratory
35

experiments (which use flavin and biomass concentrations unseen in nature) may not truly represent
the importance of that ability for EET-microbes in the natural environment. As an insoluble mineral does
not produce a diffusion gradient, the basic question of how electrogenic and electrotrophic bacteria find
these redox active minerals or poised steady state electrodes is unknown.
A series of publications authored by previous Nealson Lab member H. Wayne Harris, first attempted to
explore the nature of microbe-electrode interactions through direct observation of various Shewanella
species and mutants during mineral and anode reduction experiments. In Harris et al. (40), the motile
behavioral response termed ‘electrokinesis’ was first described. Electrokinesis is defined as the increase
in swimming speeds and motile activity of redox/electropotential sensitive bacteria in the presence of a
redox active surface. Although a gradient is created during the reduction of metal oxides (a gradient of
soluble ions and protons emanating from the reduced surface of the minerals), in poised electrode
systems, as electrons are donated to the anode, no gradient of reduced products is established. The
fundamental question of how EET-bacteria find/utilize and congregate near insoluble electron acceptors
is unknown. Measurements of swimming speeds for select Shewanella species and mutants were made
at a number of poised potentials, and yielded an apparent correlation, in agreement with the notion
that a higher anodic potential could result in a higher electron flow, and thus a higher membrane
potential and perhaps greater motility. The nature of this correlation needs to be confirmed with higher
resolution. The electrokinesis response has also been determined to result in the congregation of
electrogenic bacteria near redox-active/poised surfaces and leads to eventual colonization and biofilm
formation (41). Using mutants of S. oneidensis MR-1, it was determined that both chemotaxis genes
(mcp cache) and EET associated proteins (MtrBC/omcA complex) are required for cells to congregate
around insoluble electron acceptors (14). The aforementioned EET model organism, Thioclava
electrotropha ElOx9 is the subject of the experiments detailed within this chapter in which digital
holographic microscopy (DHM) is used to view early interactions of cathode oxidizing bacteria with
36

poised electrodes. As DHM provides the ability to volumetrically determine cell location and swimming
trajectories, the ability to observe and quantify how initial interactions between wildtype EET microbes
and poised surfaces hopes to illuminate dynamics and behavior of microbe-redox surfaces in the
environment. Here I describe the preliminary results obtained in the study of the interaction between an
electrotrophic microbe and a negatively charged electrode (cathode) and present the various
approaches and development of an experimental design to be used in future DHM-EET experiments.
This work represents the first reporting and characterization of electrotrophic motility behavior during
cathode oxidation.
37

The development, application and observations of digital holographic microscopy in
microbe-cathode interactions

This chapter describes the design, development and implementation of an experimental apparatus in
which digital holographic microscopy is used to explore microbe-electrode interactions at the single cell
level. The work included here is the product of a multiyear project, in which many directions, prototypes
and designs were implemented, tested and evaluated. The result of which is a method and design that is
both robust and adaptable. This work lays the foundation for the next generation of EET experiments in
which the behavior of microbes in the presence of anodic and cathodic electrodes can be directly
observed and quantified. As the field of EET-related research continues to grow, so does the taxonomic
diversity of known electrogenic and electrotrophic bacteria on Earth. The use of poised electrodes to
mimic reduction-oxidation potentials across a broad range of metabolisms is a valuable sterile technique
in the search for life. By modulating the electropotentials to mimic various plausible redox chemistries,
electrodes may be one of the most diverse, yet chemically aseptic methods of eliciting a motile
response. Combined with the sensitivity for low biomass samples inherent to DHM technology, the
meeting of electrochemical techniques and volumetric imaging gives insight into a more “natural” view
of extracellular electron transfer in almost any environment where microbes reside.
As discussed previously, the electrokinesis motile response was described in 2010 by Harris et al. Using a
modified microbial fuel cell design (which utilized both a cathode and anode chamber), electropotenitals
were poised on an insulated carbon fiber filament. The fiber was housed within the anaerobic half of the
microbial fuel cell design, which included the use of glass capillary tubes to provide a method of
observation on a standard light microscope (Figure 2.5). This design was based on the early two
chamber electrochemical cell models used heavily in the Nealson lab in the late 2000’s. Although
functional, the two chambered design is fairly complex, relying on a proton permeable membrane to
allow for diffusion of protons from the anoxic anodic half to the oxic cathode compartment. As the
38

initial research was done with various Shewanella species under anode reduction conditions,
maintaining anaerobic conditions in the anode compartment was paramount in order to assure data
was valid.

Figure 2.5 Full Cell Electrochemical Cell used in Harris (2010).

The first attempts in applying DHM to the study of electrodes (May, 2018) utilized the experimental
design of Harris (2010) and S. oneidensis MR-1. This system was not applicable for use with the DHM for
a number of reasons: 1) the capillary tubes introduced non-uniform changes in the light path that led to
artifacts in both phase and amplitude reconstruction, 2) the two-compartment system did not lend itself
well to our experiments (or to further miniaturization) and 3) it was not possible to maintain anaerobic
conditions in the DHM system when interfaced with the system, as shown in Figure 2.5. One solution to
this problem was to move to an organism that was not negatively impacted by the presence of oxygen. I
decided to switch organisms to Thioclava electrotropha, a bacterium I had worked with previously in
cathode oxidation experiments. By using a cathode oxidizer, the issue of maintaining anaerobic
conditions and the two-chamber system was no longer necessary. As the majority of EET-related
research conducted during the early to mid-2010’s in the Nealson lab had transitioned to using three-
39

electrode electrochemical cells, and I was no longer worried about maintaining anaerobic conditions,
the first prototype three-electrode cells were built and tested during the Fall of 2018. These
electrochemical cells still utilized the poised carbon fiber working electrode common to the Harris
publications, but was sandwiched between two standard microscope slides offset by PDMS polymer
shims that we had manufactured ourselves, and only sterilizable through cleaning with ethanol. The
well-sections of these sample chambers were the same plastic pieces produced for the original literature
as they were autoclavable and reusable. Like the original electrokinesis work, this sample chamber
design relied heavily on silicon grease to seal the system both near the working electrode and at the
reference/counter electrode. A new addition at this time was the incorporation of single-use screen
printed carbon electrodes (Pine Research, Durham, NC) which contained both a carbon electrode
surface (used as the counter electrode) and a Ag/AgCl counter electrode (Figure 2.6).

Figure 2.6 Screen Printed Electrode. Fabricated by Pine Research (Durham, NC) used for both the reference and
counter electrodes in the final sample chamber design.

Although this experimental design was functional, it was still clumsy and not nearly as ‘plug and play’ as I
had hoped to achieve for the experimental sample chamber design. The decision to move to indium tin-
oxide (ITO) coated glass electrodes had been discussed previously but had not been tried due to
concerns from the Caltech team regarding photon penetration and potential scattering of the beam due
40

to the ITO coating. Much of the concurrent work in the Nealson Lab (produced by A. Rowe, B. Lam, and
I) utilized ITO glass as a method of reducing electrochemical noise associated with high surface area.
Although electrodes constructed of carbon felt, carbon cloth, or graphite offered high surface-area (a
feature that many EET bugs preferred much like their affinity for high -surface area, amorphous
minerals), they result in high noise to signal ratios that can obscure electrochemical analysis. ITO is glass
that is coated with a thin layer of indium tin-oxide, allowing for the conduction of electrons with
minimal surface topography. These smooth electrodes are ‘clean’ electrochemically and capable of
discerning features in electrochemical data that would be lost in traditional electrode materials. Late
one evening, while working on the research that would be utilized in the JoVE publication (see Other
Publications), I assembled two simple sample chambers in which the top layer of glass (the ‘ceiling’ of
the sample chamber) comprised of either an ITO coated coverslip, or the standard microscope slide thick
ITO coated glass we had been using in the Nealson Lab. Optically, neither coated glass samples
introduced discernable differences in resolution or reconstructed holograms. Due to the fragile nature
(and higher cost) of the coverslip design, we decided to move forward prototyping an ITO sample
chamber using the standard glass thickness of 1 mm.
The change to an ITO coated electrode was critical for a number of reasons. Although DHM is volumetric
in nature, the ‘shadow’ produced by the poised carbon fibers left us with a zone of unresolvable data. By
switching to an ITO electrode we could view the entire sample volume unobstructed and had the ability
to track all cells within the FOV. Incorporating an ITO electrode as the ceiling surface of the sample
chamber addressed my concern about the settling of dead cells (and potential to obstruct real electrode
usage or accumulation rates) at the bottom of the sample chamber. As ITO electrodes had proven to
withstand sterilization via autoclaving, it was now possible to make a hardier sample chamber to
investigate electrokinesis. Much of the Winter of 2018 and Spring of 2019 was dedicated to
experimenting with sample chamber design. Some of the first design improvements included the
41

implementation of standardized plastic shims, laser cut reservoir wells, marine epoxy and Luer lock
needles. Although the PDMS shims used in the first prototype were malleable and optically translucent,
the issue of sterilization and reproducibility had reduced my enthusiasm in using them long term. As the
Caltech group had begun to use laser cut plastic sheeting to develop sample chambers for the
orangeBox field instrument, the ability to securely adhere autoclave-proof plastic shims to glass surfaces
was adopted for use in our sample chamber design. As our experiment did not require the precision of
laser cut wells used in the field instrument, the ability to quickly manufacture suitable sample chambers
using shims constructed of 0.0125 inch thick plastic (McMaster-Carr part # 9513K55) allowed for an
increase in reproducibility and uniformity between individual chambers. Another significant
improvement was the adoption of standardized reservoir wells. The wells were cut (using the CalTech
Fabrication Lab’s CO
2
laser cutter) out of a 0.06-inch-thick piece of plastic (McMaster-Carr part
9513K93). Measuring 25 mm x 25 mm with a circular hole of 0.5 inch diameter, the wells provide a large
surface in which the counter and reference electrodes can be submerged. Due to the round nature of
the center hole and the cohesive and adhesive properties of water, a convex ‘droplet’ of media can be
formed ensuring both electrodes are in constant contact with the fluid and the electrochemical circuit
with the working electrode is maintained (Figure 2.7).

Figure 2.7 CAD Image of Laser Cut Well Assembled on Sample Chamber. Well dimensions: 25 mm x 25 mm with
0.5 inch diameter center cut.

The use of marine epoxy allowed for the construction of complex sample chambers prior to
sterilization. The two chamber cells used in Harris et al. required the sterilization of multiple pieces
42

immediately prior to experimental runs. By using autoclave proof marine epoxy, 90% of the final sample
chamber could be built and ran through autoclave sterilization prior to use, allowing for ease of
switching between sample chambers and the running of back to back experiments with little down time.
The use of epoxy also provided the ability to permanently secure Luer lock needles into the sample
chamber opposing the laser cut well. This provided a simple way of introducing cells and media into the
sample chamber, while avoiding air pockets or bubbles on the glass surfaces. Additionally, by
permanently sealing one side of the chamber, the amount of inherent drift in the sample volume was
dramatically reduced (Figure 2.8).

Figure 2.8 Various Half Cell Sample Chamber Designs. A.) Large ITO surface (SA = 400 mm
2
). B.) Limited ITO
surface (SA = 225 mm
2
). C.) First prototype with Kapton tape to keep EET activity away from the sample well and in
a more easily observable area (SA = 140 mm
2
).
What became increasingly apparent was that the geometry of the system dictated the response of
bacteria. Since we were experimenting with low cell densities on the order of 10,000 cells per mL, and
the initial design contained a larger working electrode surface (on the scale of 400 mm
2
), although the
bacteria were interacting with the electrode surface throughout the sample chamber, the organisms
43

were quite spread out and the number of active cells per Field-of-View (FOV) was low, with many FOV
showing no active motility. Due to the increased complexity inherent to microbe-electrode experiments
and the necessity of replicates to determine any discernible trend or even proof of experimental
success/failure, the evolution of experimental designs is a months-long process. After a few weeks of
running experiments with the first ITO sample chamber design, it became obvious that we could do
better. First attempts utilized limiting the sample volume (i.e. making the plastic shims a larger
proportion of the sample chamber, Figure 2.8B); however, due to the fact that the well/reference and
counter electrode side of the sample chamber was not readily accessible to pass under the objective
lens while maintaining relative focus within the sample chamber, this design did not solve the problem.
If we view the electrochemical circuit primarily as field lines between the counter and working
electrode, it makes sense that many of the bacteria would accumulate at the point of the working
electrode nearest the counter. Once this was experimentally determined to be true, the usage of Kapton
tape was adopted into the sample chamber design. Kapton tape (chemically designated as 4,4'-
oxydiphenylene-pyromellitimide) is an insulating tape often used in electronic devices. Developed by
DuPont in the 1960’s, this polyimide film has a large temperature stability range of -269 to + 400° C (thus
able to survive autoclave sterilization) and is used in many spacecraft and satellites (including the James
Web Space Telescope’s sunshield). The first iteration of the Kapton amended sample chamber limited
the working electrode surface to just that which was not covered by the plastic shim (Figure 2.8C). One
of the immediately recognized benefits of using Kapton tape was the ability to easily determine the
surface of the ITO glass. As Kapton reduces light transmission, being able to differentiate and reproduce
focus near the ITO surface rapidly decreased the time needed to confirm where the focal plan was
within the sample chamber. Although the inclusion of Kapton did somewhat improve the bacterial
response, a more direct and ‘point source’ geometry was desired. During the Spring of 2019, I decided
to start implementing a ‘viewing port’ design to limit the amount of working electrode surface exposed
44

to the media and bacterial culture, in hopes of accumulating more cells per FOV (Figure 2.9A). Using the
0.25 inch diameter of a standard hole punch, the effective area of the working electrode was reduced
29% to ~41 mm
2
(compared to the original Kapton tape design). This ‘view port’ sample chamber
designed allowed for a relatively restricted area of EET-activity, but the ability to scan across a number
of FOV. This design became the final iteration of the electrokinesis sample chamber and was utilized in
the majority of the experiments reported within this chapter under both cathodic and anodic conditions.
Over time, it became apparent that behavior of the microbes in our electrode system differed from that
reported by Harris et al. (2010). Unsurprisingly, our relatively large planer electrode surface elicits a
different response than the point source experiments in which electrokinesis was first reported.
Although not reported here, additional experiments involving anode reduction by Shewanella oneidensis
MR-1 were conducted in the Winter of 2019.
Generalized Experimental Design
Sample chambers were assembled in a multiday process so as to allow the adherence of tape covered
plastic shims to the glass and the curing of marine epoxy. To ensure optimal imaging conditions, all glass
surfaces were cleaned with 70% ethanol and/or chloroform prior to assembly. Once assembled (yet
prior to the epoxy step), sample chambers were left sandwiched in a vise overnight to ensure complete
bonding of the doubled sided tape to the glass surfaces. The next day, marine epoxy was applied to all
seams and to affix the Luer lock needle in place and left to dry on the benchtop. Once cured, these
sample chambers are able to be sterilized via autoclave, and for the most part have been shown to be
reusable with proper cleaning and re-sterilization. The re-usability of sample chambers was a feature
that I wanted to explore as the cost of each sample chamber came out to ~ 4.00 $ each, and I wanted to
move towards a more reproducible and lower waste generating procedure. That said, for all of the
experiments in this thesis chapter, new sample chambers were used for each experiment and were
never reused to obtain data.
45

Figure 2.9: Final Design of EET Sample Chamber. Utilizing view port design in which the working electrode surface
area is approximately 41 mm
2
. A.) ITO glass with Kapton tape view port. B.) Assembled sample chamber prior to
epoxy. C.) Fully assembled sample chamber for experimental run.

Figure 2.10 illustrates a fully assembled viewing port sample chamber resting on the XYZ translational
stage during an experimental run. In brief: cells are injected via syringe through the Luer lock needle
until the sample chamber body and reservoir well are full. A sterile Luer lock cap is attached to the end
of the needle to prevent flow within the sample volume. A screen-printed electrode (with attached
insulated copper wire leads) is placed into contact with the solution and immobilized through the use of
sterile vacuum grease. The leads are attached to the appropriate alligator clips of the potentiostat
46

device for the reference and counter electrodes. An exposed corner of the ITO working electrode was
connected to the working electrode lead of the potentiostat and often utilized aluminum foil to help
ensure optimal contact was maintained.

Figure 2.10 Assembled DHM-EET electrochemical cell on the translational stage during experimental run. Figure
on right shows the MZ-DHM in use in an anaerobic chamber during anode reduction experiments.

Chronoamperometry data
Chronoamperometry is an electrochemical technique in which the current generated (or electron
uptake) occurring on a poised electrode surface is measured over time. This technique is utilized for
both EET-experimentations in the laboratory setting and in electrode enrichment experiments from the
environment. A Gamry Reference 600 potentiostat and the Gamry Framework software (Gamry
Instruments, Warminster, PA) were used to establish working electrode potentials and monitor current/
electrochemical cell health. Electron flow reflected in the electron uptake (under cathodic conditions,
-µA)) was quite low for all experiments due to the low biomass nature of the experiments (Figure 2.11).
Although the use of ITO electrodes drastically diminishes the background noise associated with higher
surface area materials, the use of longer alligator clip wires between the sample chamber and the
potentiostat increased the baseline noise in the electrochemical data. No real current generation or
47

change in electron uptake was discernable, as these experiments do not observe the multiday time
periods associated with long term electrode utilization at high cell densities. The nature of the low
biomass, short duration and low signal to noise ratio inherent to these experiments result in the
chronoamperometry data being used solely as a method of evaluating the health of the electrochemical
cell during experimental runs. Monitoring the current produced (for these experiments at the nA scale)
ensures that the closed electrochemical circuit was maintained, and the working electrode potential was
held constant. Figure 2.11 displays typical cathodic chronoamperometry data obtained through the first
hour of these experiments.

Figure 2.11 Chronoamperometry data. Data reflects typical electron uptake values through the first hour of
experimentation. Due to low biomass numbers (which limits the number of microbe-electrode interactions
occurring per second) chronoamperometry is used primarily as a method of evaluating the conditions of the
electrochemical cell and integrity of the closed circuit.

Data acquisition, processing, reconstructions and tracking
T
0
holograms were acquired using Koala (for cathode experiments) or the DHMx software suite (for
anode experiments) prior to the establishment of an electropotential, and taken at 10 FPS frame rates at
48

T
0.5 hours
,T
1hour
, T
2 hours
. Additional data not reported here was taken every 30 minutes for up to 8 hours
after establishment of the working electrode potential. Holograms were acquired for ~ 60 seconds
duration and taken in triplicate at three random fields of view within the sample chamber. Processing
was done via the HologramDenoise MatLab scripts developed by M. Bedrossian and reconstructions
were done in accordance to the acquisition software used. Replicate data sets were analyzed,
normalized and combined by time point.
For Koala-acquired data, the software reconstructs the hologram in the detector space (i.e. as it appears
at the surface of the CCD detector which is 1 cm
2
), thus the information must be converted from the cm
scale down to the appropriate micron scale. Lateral reconstruction parameters are simply the true
magnification of the system, whereas axial reconstruction parameters are a function of mag
2
and will
change with different objective lenses and DHM instruments. For the DHMx software suite, the JPL team
chose to avoid this issue by reconstructing holograms in the image space (i.e. on the micron scale by
using the known FOV dimensions), which is then used to calculate the lateral and axial magnification
parameters to be applied during reconstruction. The software also includes a better peak detection
capability that provides a higher degree of tilt correction in the reconstructed holograms. Although the
two reconstruction software suites are capable of achieving the same desired outcome, the DHMx
software provides a more straight-forward method of doing so with a higher fidelity reconstruction.
Methods
Since the electrokinesis motile response was initially observed and classified through the use of anode
reducing bacteria, one of the fundamental questions I wanted to explore was the behavior of cathode
oxidizing bacteria. Using the recently isolated and characterized sulfur oxidizing bacterium Thioclava
electrotropha, the utilization of ITO and poised platinum wire working electrodes was observed.
Cathode oxidation experiments were conducted on the bench top, as T. electrotropha is capable of
49

coupling mineral/cathode oxidation with oxygen reduction. Tested cathodic electropotentials include
open circuit (control), -401 mV, -501 mV, and -601 mV (vs Ag/AgCl). Experiments using coated Pt wire as
a point source electrode were conducted solely under a -601 mV electropotential. An additional group
of experiments (not reported here) observed the prevalence of motility after removal of the
electropotential and in the presence of elemental sulfur minerals.
Cell culture and medias
T. electrotropha ElOx9 cultures were grown aerobically in Difco
TM
Marine (DM) broth from frozen (-80°C)
stock as a rich growth medium at 30° C for 24 hours until a cell density on the order of 10
8
cells per mL.
Cells were pelleted via centrifuge and resuspended (7200 X g, 3 X washed) in a low-sulfate Salt Water
Base (SWB) minimal media that contained no additional electron donor or acceptor. Used for
electrochemical experiments and enrichment, the base media contains 342 mM NaCl, 14.8 mM
MgCl
2
6H
2
O, 0.1 mM CaCl
2
2H
2
O, 6.7mM KCl, 10mM NH
4
Cl, 1 mM Na
2
SO
4
and 1 mM phosphate buffer
(pH = 6.5). Resuspended cells were diluted to a DHM appropriate concentration of ~10
4
cells per mL
with additional SWB media and were injected into sterile experimental sample chambers.
Results
All motile cells that were found within the top 120 µm were tracked, with some cells below that cut off
being tracked dependent on the reconstruction software used. Tracks and z axis heights were then
normalized to reflect distance from the ITO surface. Motility statistics were determined using the
MatLab scripts developed by the Caltech DHM team (available at M. Bedrossian’s Github:
https://github.com/mbedross/manualTracking). Tracks of cells at the poised experimental
electropotentials over time are shown in Figure 2.12. Video 2.1 shows the amplitude reconstruction of a
-601 mV experiment at the ITO surface at T= 30 min, shown in real time at 10 fps.

50

Figure 2.12 Tracks of T. electrotropha cells under various cathodic potentials over time. ITO electrode surface is
located at Z=0. Z axis locations have been normalized to reflect distance from the ITO surface.

Individual bacterial cells that were determined to have resided at the electrode surface (for > 5 frames,
or 0.5 seconds) were noted and subjected to additional analysis. These microbe-electrode interactions
were categorized into one of two groups: full cycle or half cycle. Full cycle interactions describe bacteria
51

that were motile prior to and after residing at the electrode surface for a period of time. These
interactions allow for a fuller determination of the cell’s behavior, as both the mean Speed
prior
and mean
Speed
post
electrode interaction can be quantified in addition to Δ
speed
. The residence time or ‘time on
electrode’ was determined by the number of frames that the cell was localized at the electrode surface
with no discernable active motility. Cells were characterized as ‘half cycle’ if only one of the Speed
variables could be determined due to the cell being present at the electrode surface prior to the start of
data acquisition, or having failed to leave the surface by the final frames of the acquired hologram. Both
full cycle and half cycle cells are listed in Table 2.1. Importantly, only cells that were determined to show
non-random movement at some point in the data sets were tracked. There were cells that appeared
near the electrode surface, but if they exhibited no direct motility, they were not tracked. Only cells that
displayed active swimming prior-to or after touching the electrode surface were tracked and quantified.

52

Table 2.1. Analysis of individual Thioclava electrotropha cells that utilize the electrode surface. The top half of
the table represents the ‘full cycle’ cells; whereas, cells below the double line are categorized as ‘half cycle’ and
thus are unable to determine Δ
speed
or the true residence time.

To determine if any statistically significant relationship between the electropotential, electrode
residence time or the Δ
SPEED
could be discerned, a Spearman correlation analysis was conducted (Table
2.2). The only correlation of statistical significance was between the electropotential and residence time
(p =0.005); however, this designation may be an artifact of the over representation of -601 mV in the full
cycle bacteria cohort (n= 14 of 20) and single representation of a full cycle -401 mV bacteria. That said,
the correlation of longer residence times at less cathodic electropotentials is in theory understandable,
as a more negative (more cathodic) potential would facilitate a higher electron uptake, and potentials
that are less negative would require more time to uptake the same number of electrons. The lack of
53

correlation of residence time or electropotential with Δ
SPEED
is notable and implies that the
manifestation of the electron uptake (through increased speed via the PMF) is not readily discernible.
Table 2.2. Spearman Correlation results for full cycle bacteria.

The adjusted mean swimming speeds of the electrode utilizing bacteria was determined by removing
the velocity rates recorded during the residence time of the cells on the electrode (i.e. Brownian
motion). All cells were then grouped by electropotential and time point (Figure 2.13 and Table 2.3).

Figure 2.13 Mean speeds of all tracked motile cells by electropotential over time. Error bars reflect standard
deviation.

54

Table 2.3. Mean speeds of tracked cells by potential and time point.

To determine if either electropotential or time point significantly predicted the determined mean speed,
linear regression analysis was conducted. The results of the linear regression model were significant
with an R
2
= 0.16, predicting that approximately 16% of the variance in mean speed can be explained by
electropotential and time point. The most predominant factor influencing mean speed was determined
to be electropotential, with -501 mV eliciting the highest mean rates of swimming (p<0.001).
Conclusion
As Thioclava electrotropha is just now starting to be used as a model organism in the emerging field of
oxidation-based EET, this work not only provides the first direct visual observations of the species during
microbe-electrode interactions, but also of oxidative EET in general. By incorporating DHM technology
into the field of extracellular electron transfer research, we provide a unique and robust ability to
determine single cell dynamics during electron transfer events. Although much of the previous
laboratory research has elucidated the molecular and genomic underpinnings of EET under ideal
conditions, experiments in which sparse samples and the motile behavior of individual cells can be
quantified allows us to observe a more natural behavior that more realistically represents the conditions
faced by the bacterium in the natural environment. The experiments described here more closely reflect
55

the struggle of life in low energy systems as they navigate the environment in search of a metabolically
available electron donor (which must be more difficult in the case of gradient-less insoluble substrates).
This work reports the development and implementation of a technique in which EET-active bacteria can
be observed, tracked and quantified at the single cell level, as they interact with redox-active surfaces.
This breakthrough allows for the exploration of the nature of microbe-electrode interactions in three
dimensions and furthers our understanding by providing the first account of the in-situ behavior of
cathode oxidizing bacteria. Although some of the initial aims of this line of research have not been
elucidated (but are expected upon completion of analysis of the larger data set), these results have
shown that: 1) We have successfully developed and implemented a volumetric DHM-
electromicrobiology experimental design in which imaging of cells during electron transfer events are
capable, 2) we are able to quantify the motility rates of individual cells prior to and after interacting with
a poised electrode surface, in addition to determining the localization time at the electrode surface, 3)
we can compare motility rates across different working electrode voltages. This work offers preliminary
insights into the individual cell dynamics of electron transport events and provides an experimental
design in which to further study these interactions.

References:
1. W. E. Dean, W. S. Moore, K. H. Nealson, Manganese cycles and the origin of manganese nodules, Oneida
Lake, New York, USA. Chem. Geol. 34, 53–64 (1981).
2. B. Tebo, R. Rosson, K. Nealson, in Black Sea Oceanography. NATO ASI Series, M. J. . İzdar E., Ed. (Springer,
Dordrecht, vol 351., 1991).
3. K. H. Nealson, C. R. Myers, B. B. Wimpee, Isolation and identification of manganese-reducing bacteria and
estimates of microbial Mn(IV)-reducing potential in the Black Sea. Deep. Res. Part A. 38, S907–S920 (1991).
4. C. R. Myers, K. H. Nealson, Microbial reduction of manganese oxides: Interactions with iron and sulfur.
Geochim. Cosmochim. Acta. 52, 2727–2732 (1988).
5. D. R. Lovley, E. J. Phillips, Rapid assay for microbially reducible ferric iron in aquatic sediments. Appl.
Environ. Microbiol. 53, 1536–1540 (1987).
56

6. K. H. Nealson, A. R. Rowe, Electromicrobiology: realities, grand challenges, goals and predictions. Microb.
Biotechnol. 9, 595–600 (2016).
7. D. J. Richardson, J. N. Butt, J. K. Fredrickson, J. M. Zachara, L. Shi, M. J. Edwards, G. White, N. Baiden, A. J.
Gates, S. J. Marritt, T. A. Clarke, The “porin-cytochrome” model for microbe-to-mineral electron transfer.
Mol. Microbiol. 85, 201–212 (2012).
8. Y. Liu, Z. Wang, J. Liu, C. Levar, M. J. Edwards, J. T. Babauta, D. W. Kennedy, Z. Shi, H. Beyenal, D. R. Bond, T.
A. Clarke, J. N. Butt, D. J. Richardson, K. M. Rosso, J. M. Zachara, J. K. Fredrickson, L. Shi, A trans-outer
membrane porin-cytochrome protein complex for extracellular electron transfer by Geobacter
sulfurreducens PCA. Environ. Microbiol. Rep. 6, 776–785 (2014).
9. M. Y. El-Naggar, G. Wanger, K. M. Leung, T. D. Yuzvinsky, G. Southam, J. Yang, W. M. Lau, K. H. Nealson, Y.
A. Gorby, Electrical transport along bacterial nanowires from Shewanella oneidensis MR-1. Proc. Natl.
Acad. Sci. U. S. A. 107, 18127–18131 (2010).
10. J. A. Gralnick, On conducting electron traffic across the periplasm. Biochem. Soc. Trans. 40, 1178–1180
(2012).
11. H. H. Hau, J. A. Gralnick, Ecology and Biotechnology of the Genus Shewanella . Annu. Rev. Microbiol. 61,
237–258 (2007).
12. J. K. Fredrickson, M. F. Romine, A. S. Beliaev, J. M. Auchtung, M. E. Driscoll, T. S. Gardner, K. H. Nealson, A.
L. Osterman, G. Pinchuk, J. L. Reed, D. A. Rodionov, J. L. M. Rodrigues, D. A. Saffarini, M. H. Serres, A. M.
Spormann, I. B. Zhulin, J. M. Tiedje, Towards environmental systems biology of Shewanella. Nat. Rev.
Microbiol. 6, 592–603 (2008).
13. J. F. Heidelberg, I. T. Paulsen, K. E. Nelson, E. J. Gaidos, W. C. Nelson, T. D. Read, J. A. Eisen, R. Seshadri, N.
Ward, B. Methe, R. A. Clayton, T. Meyer, A. Tsapin, J. Scott, M. Beanan, L. Brinkac, S. Daugherty, R. T.
DeBoy, R. J. Dodson, A. S. Durkin, D. H. Haft, J. F. Kolonay, R. Madupu, J. D. Peterson, L. A. Umayam, O.
White, A. M. Wolf, J. Vamathevan, J. Weidman, M. Impraim, K. Lee, K. Berry, C. Lee, J. Mueller, H. Khouri, J.
Gill, T. R. Utterback, L. A. McDonald, T. V. Feldblyum, H. O. Smith, J. C. Venter, K. H. Nealson, C. M. Fraser,
Genome sequence of the dissimilatory metal ion-reducing bacterium Shewanella oneidensis. Nat.
Biotechnol. 20, 1118–1123 (2002).
14. H. W. Harris, I. Sánchez-Andrea, J. S. McLean, E. C. Salas, W. Tran, M. Y. El-Naggar, K. H. Nealson, Redox
sensing within the genus Shewanella. Front. Microbiol. 8, 1–11 (2018).
16. W. C. Lin, M. V. Coppi, D. R. Lovley, Geobacter sulfurreducens Can Grow with Oxygen as a Terminal
Electron Acceptor. Appl. Environ. Microbiol. 70, 2525–2528 (2004).
17. R. Mahadevan, B. Palsson, D. R. Lovley, In situ to in silico and back: Elucidating the physiology and ecology
of Geobacter spp. using genome-scale modelling. Nat. Rev. Microbiol. 9, 39–50 (2011).
18. G. W. Chong, A. A. Karbelkar, M. Y. El-Naggar, Nature’s conductors: what can microbial multi-heme
cytochromes teach us about electron transport and biological energy conversion? Curr. Opin. Chem. Biol.
47, 7–17 (2018).
19. L. Shi, H. Dong, G. Reguera, H. Beyenal, A. Lu, J. Liu, H. Q. Yu, J. K. Fredrickson, Extracellular electron
transfer mechanisms between microorganisms and minerals. Nat. Rev. Microbiol. 14, 651–662 (2016).
20. B. H. Kim, H. J. Kim, M. S. Hyun, D. H. Park, Direct electrode reaction of Fe(III)-reducing bacterium,
Shewanella putrefaciens. J. Microbiol. Biotechnol. 9, 127–131 (1999).
21. D. J. Burdige, S. P. Dhakar, K. H. Nealson, Effects of manganese oxide mineralogy on microbial and chemical
manganese reduction. Geomicrobiol. J. 10, 27–48 (1992).
22. A. R. Rowe, P. Rajeev, A. Jain, S. Pirbadian, A. Okamoto, J. A. Gralnick, M. Y. El-Naggar, K. H. Nealson,
57

Tracking electron uptake from a cathode into Shewanella cells: Implications for energy acquisition from
solid-substrate electron donors. MBio. 9, 1–19 (2018).
23. A. R. Rowe, P. Chellamuthu, B. Lam, A. Okamoto, K. H. Nealson, Marine sediments microbes capable of
electrode oxidation as a surrogate for lithotrophic insoluble substrate metabolism. Front. Microbiol. 6, 1–
15 (2015).
24. Z. M. Summers, H. E. Fogarty, C. Leang, A. E. Franks, N. S. Malvankar, D. R. Lovley, Direct Exchange of
Electrons Within Aggregates of an Evolved Syntrophic Coculture of Anaerobic Bacteria. Science (80-. ). 330,
1413–1415 (2010).
25. G. Wegener, V. Krukenberg, D. Riedel, H. E. Tegetmeyer, A. Boetius, Intercellular wiring enables electron
transfer between methanotrophic archaea and bacteria. Nature. 526, 587–590 (2015).
26. J. A. Gralnick, D. K. Newman, Extracellular respiration. Mol. Microbiol. 65, 1–11 (2007).
27. C. Pfeffer, S. Larsen, J. Song, M. Dong, F. Besenbacher, R. L. Meyer, K. U. Kjeldsen, L. Schreiber, Y. A. Gorby,
M. Y. El-Naggar, K. M. Leung, A. Schramm, N. Risgaard-Petersen, L. P. Nielsen, Filamentous bacteria
transport electrons over centimetre distances. Nature. 491, 218–221 (2012).
28. M. Y. El-Naggar, S. E. Finkel, Live wires. Scientist (2013), pp. 38–43.
29. F. J. R. Meysman, Cable Bacteria Take a New Breath Using Long-Distance Electricity. Trends Microbiol. 26,
411–422 (2018).
30. D. Trojan, L. Schreiber, J. T. Bjerg, A. Bøggild, T. Yang, K. U. Kjeldsen, A. Schramm, A taxonomic framework
for cable bacteria and proposal of the candidate genera Electrothrix and Electronema. Syst. Appl.
Microbiol. 39, 297–306 (2016).
31. S. E. McGlynn, G. L. Chadwick, C. P. Kempes, V. J. Orphan, Single cell activity reveals direct electron transfer
in methanotrophic consortia. Nature. 526, 531–535 (2015).
32. S. Kato, R. Nakamura, F. Kai, K. Watanabe, K. Hashimoto, Respiratory interactions of soil bacteria with
(semi)conductive iron-oxide minerals. Environ. Microbiol. 12, 3114–3123 (2010).
33. K. C. Wrighton, J. C. Thrash, R. A. Melnyk, J. P. Bigi, K. G. Byrne-Bailey, J. P. Remis, D. Schichnes, M. Auer, C.
J. Chang, J. D. Coates, Evidence for direct electron transfer by a gram-positive bacterium isolated from a
microbial fuel cell. Appl. Environ. Microbiol. 77, 7633–7639 (2011).
34. B. R. Lam, C. R. Barr, A. R. Rowe, K. H. Nealson, Differences in Applied Redox Potential on Cathodes Enrich
for Diverse Electrochemically Active Microbial Isolates From a Marine Sediment. Front. Microbiol. 10, 1–17
(2019).
35. B. R. Lam, thesis, University of Southern California (2018).
36. R. Chang, L. Bird, C. Barr, M. Osburn, E. Wilbanks, K. Nealson, A. Rowe, Thioclava electrotropha sp. Nov., a
versatile electrode and sulfur-oxidizing bacterium from marine sediments. Int. J. Syst. Evol. Microbiol. 68,
1652–1658 (2018).
37. A. A. Karbelkar, A. R. Rowe, M. Y. El-Naggar, An electrochemical investigation of interfacial electron uptake
by the sulfur oxidizing bacterium Thioclava electrotropha ElOx9. Electrochim. Acta. 324 (2019),
doi:10.1016/j.electacta.2019.134838.
38. S. Carbajosa, M. Malki, R. Caillard, M. F. Lopez, F. J. Palomares, J. A. Martín-Gago, N. Rodríguez, R. Amils, V.
M. Fernández, A. L. De Lacey, Electrochemical growth of Acidithiobacillus ferrooxidans on a graphite
electrode for obtaining a biocathode for direct electrocatalytic reduction of oxygen. Biosens. Bioelectron.
26, 877–880 (2010).
58

39. P. F. Beese-Vasbender, S. Nayak, A. Erbe, M. Stratmann, K. J. J. Mayrhofer, Electrochemical characterization
of direct electron uptake in electrical microbially influenced corrosion of iron by the lithoautotrophic SRB
Desulfopila corrodens strain IS4. Electrochim. Acta. 167, 321–329 (2015).
40. H. W. Harris, M. Y. El-Naggar, O. Bretschger, M. J. Ward, M. F. Romine, A. Y. Obraztsovad, K. H. Nealson,
Electrokinesis is a microbial behavior that requires extracellular electron transport. Proc. Natl. Acad. Sci. U.
S. A. 107, 326–331 (2010).
41. H. W. Harris, M. Y. El-Naggar, K. H. Nealson, Shewanella oneidensis MR-1 chemotaxis proteins and
electron-transport chain components essential for congregation near insoluble electron acceptors.
Biochem. Soc. Trans. 40, 1167–1177 (2012).

Closing remarks on DHM and electrode work
The nature of electromicrobiology is extremely difficult especially from a “re-pro-du-ci-bility” (-Gijs
Kuenen), aspect. Microbiology in general is a science that is subject to immense variation, mutation,
contamination and unexplained failures requiring the utmost attention to ensure findings are valid.
Coupled with the inherent difficulties of electrochemistry, the resulting meeting and exploration of an
interaction of a physical and biological science makes electromicrobiology a rewarding yet non-trivial
pursuit. As the tools of science and modern microbiology constrain the methodology in which to
investigate microbial behaviors (like EET) in the natural environment, the nature of laboratory-based
electromicrobiology is artificial. The “behavior” we elicit may occur under conditions fundamentally off
by orders of magnitude (e.g. organisms at high cell densities, rich media chemistries, flavin
concentrations, or magnetic field intensity, etc.) it is important to keep the observations and findings of
these experimental behaviors in the context of organisms in the environment.
In the case of the original electrokinesis work (Harris, 2010) the run and reverse motile behavior
expected for a point source gradient was readily apparent. Perhaps due to the geometry of our working
electrode (planar ITO glass vs. the face of a sheared wire) that behavior was not observed in our
experiments. Due to the fact that we are using a different organism than that used in the anode work, it
is hard to discern if this difference is due the difference in species or that of the experimental design.
Although not included in the work reported here, a series of parallel experiments observing Shewanella
59

oneidensis MR-1 under anode reduction conditions was conducted in an anaerobic chamber using the
same experimental design. Upon analysis, the MR-1 experiments should help illuminate this question of
experimental geometry or species. Additionally, a series of experiments were conducted with both MR-1
and T. electrotropha that more reflect the original experimental design by using the sheared face of an
insulated platinum wire (30 µm diameter, Omega Engineering part SPPL-001) as the poised working
electrode. As these experiments are analyzed, additional insight is hoped to be gained as the results are
put into the context of reduction vs oxidation-based EET events. The work described above reports the
first instance of DHM utilization in EET research, proof that the experimental design is valid, and
represents the beginning of EET-DHM research on Earth while laying the foundation for proposed poised
electrode-DHM utilizing experiments in future space missions.

60

Chapter III: The use of DHM in the study of life at high pH, in
the field and the laboratory

Overview of chapter
Over the past decade and a half, many advances have been made in our understanding of the microbial
community that inhabits The Cedars, one of the most extreme and enigmatic field sites on Earth.
Through the work of past members of the Nealson Lab, and contributions by our collaborators both
domestic and international, we have been able to observe, describe and gain insight into the
geochemistry and microbiology that makes this unique ecosystem one of the leading extreme
environments for astrobiological research. The scientific goal of this chapter is to demonstrate the
applicability and versatility of DHM technology in the study of an astrobiologically relevant field site
through a series of independent experiments both in the field and in the laboratory. The experiments
described in the manuscript portion of this chapter were designed to explore potential future research
directions while simultaneously highlighting the technology of DHM. From in situ cell enumeration
experiments in the field to chemotaxis experiments with pure culture isolates in the laboratory, this
chapter aims to explore The Cedars’ microbial community and observe and quantify trends of motility
exhibited by prokaryotic life at high pH. The following thesis chapter contains an overview of The Cedars
field site (pgs 61-73) and the experiments conducted with DHM experiments described and reported
(pgs 74-96).
List of minerals mentioned in the following section:
Peridotite: Ultramafic igneous rocks composed of olivine and pyroxene
Olivine: (Mg, Fe)
2
SiO
4

Pyroxene: Silicate minerals typically composed of a XY(Si,Al)
2
O
6
geometry
Clinopyroxene: pyroxenes with a monoclinic crystal structure
Orthopyroxene: pyroxenes with an orthorhombic crystal structure
Spinal harzburgite: peridotite mineral consisting of ~75% olivine + 25% orthopyrocene/clinopyroxene
61

Dunite: peridotite mineral consisting of > 90% olivine
Serpentine: (Mg,Fe)
3
Si
2
O
5
(OH)
4

Serpentinite: metamorphic rock composed of serpentine group minerals, produced by the serpentinization
hydration reaction.
Chromite: (FeCr
2
O
4
)
Magnesite: (MgCO
3
)
Calcite: (CaCO
3
)

Geological and geochemical background:
The Cedars Peridotite is located in Sonoma County, California (038°37’37.53”N, 123°07’21.51”W). This
ophiolite body is composed primarily of Late Cretaceous shales, sandstone and marine sediments within
the Franciscan subduction complex of the Coast Ranges (Figure 3.1) (1, 2). Sometimes denoted as the
Cazadero ultramafic mass in early reports, The Cedars is an eugeosynclinal assemblage of sandstones,
greenstone, chert and foraminiferal limestone (3).
First mapped by the USGS in the mid-1920s, research at the Cedars focused primarily upon botanical
and naturalist observations with an emphasis on collecting and describing the unique local flora. Dr.
Robert Coleman (a leading authority on the geology of serpentines) and Ivan Barnes (namesake of the
Barnes Spring Complex) both spent considerable time at The Cedars through the 1960’s. Barnes’ 1967
Science publication gave evidence that low temperature serpentinization was actively occurring at the
site, one of the first known instances of an ongoing example of this geologically important reaction (4).
Previous to 1969, ultramafic and mafic rock assemblages within the Coast Range were believed to have
originated as igneous intrusions; however, the presence of metamorphic blueschists at The Cedars and
other sites, gave evidence of high pressure/low temperature conditions now known to exist in
subduction zones (5, 6). In 1969, Edgar Bailey (USGS) and colleagues proposed that The Cedars, among
other ultramafic rocks within the Coast Range, were sections of oceanic basement comprised of the late-
Mesozoic Great Valley sequence and not native to the Franciscan Complex (7).
62

Figure 3.1 Map of The Cedars Peridotite and insert of field site locations. USGS altitudes and lineation map
produced by E.H. Bailey in the mid 1960’s. From (3). Field site location map from (1). Note: The spring denoted as
CS1 (listed in the early literature as Campsite Spring) has been changed circa 2015 to the Mortar Bed Spring (MBS).
63

Figure 3.2 Geological setting of The Cedars active-terrestrial serpentinization site. A.) The Cedars Peridotite
Ophiolite, from Morrill et al., 2013 (1). B.) Transect of fault structure across The Cedar Peridotite From SW to NE.
Also from (1). C). Photograph of The Cedars’ Main Canyon, in which the Austin Creek and the general field site is
located. From Raiche, 2009 (8).

The geology of The Cedars exhibits a layering gradient of olivine/orthopyroxene composition within the
harzburgite peridotite and the presence of chromite aggregates, both of which indicate high
temperature ductile flow in the upper mantle. Notably, The Cedars peridotite does not contain the
upper layers of gabbro and basalt commonly seen in ophiolites. These layers are hypothesized to have
been removed by faulting. The Cedars is thought to be a section of the mantle tectonite, previously
belonging to the larger Coast Range Ophiolite (3). The Cedars Peridotite is composed primarily of olivine,
clinopyroxene and othropyroxene in various proportions. Spinal harzburgite (75% olivine + 25%
orthopyrocene/clinopyroxene) and subordinate dunite (100% olivine) dominate the rock composition.
Chromite is found as an accessory mineral (< 0.5%) within the obducted rock (2, 9). The peridotite body
is altered to serpentinite 100% along the perimeter of the surrounding rock and found at 5-20% within
the central ophiolite (1, 2). The hydration reaction responsible for the alteration of the peridotite is the
64

serpentinization reaction (Eq. 3.1-3.4). This reaction is endemic to marine tectonic plate boundaries as it
occurs when seawater interacts with deep ultramafic rocks. Locations of terrestrial serpentinization
reactions are much less common but have been found on all three major Northern Hemisphere
continents. The serpentinization reaction is accompanied by increases in rock volume, which can result
in serpentinite extrusions (4). Highly alkaline and geochemically characteristic fluids are extruded from
the rocks in multiple locations throughout the peridotite. The chemistry of this hydration reaction and
its byproducts create a unique environment reflected in both the native microorganisms and the
implications of their existence. Much work has been produced looking at the geochemistry, geobiology
and microbial ecology of The Cedars environment.
The generalized serpentinization hydration reaction can be described as:
Eq. 3.1 Fe
2
SiO
4
+ 5Mg
2
SiO
4
+ 9H
2
O → 3Mg
3
Si
2
O
5
(OH)
4
+ Mg(OH)
2
+ 2Fe(OH)
2

Eq. 3.2 Mg
2
SiO
4
+ Mg
2
SiO
3
+ 2H
2
O → 3Mg
3
Si
2
O
5
(OH)
4

Note: Pyroxene denotes both ortho- and clinopyroxene minerals.
Theoretical thermodynamic calculations provided in (10) indicated that in the presence of both
serpentine and brucite, the alteration reaction can be described as:

Eq. 3.3 Fe
6
Si
4
O
10
(OH)
8
+ 6Mg(OH)
2
↔ Mg
6
Si
4
O
10
(OH)
8
+ 6FE(OH)
2

Eq. 3.4 6Fe(OH)
2
↔ 2Fe
3
O
4
+ 2H
2(aq)
+ 4H
2
O

Of note in equation 3.4 is the production of hydrogen gas. The resulting altered fluid creates an
environment in which the reduction of carbon (if present) to methane is possible, due to the H
2
gas,
higher temperatures and Fe catalyst(s) (1, 10). The Cedars, however, is CO
2
-poor (1). Models predict that
in Mg-rich serpentine environments that contain low Al quantities (such as The Cedars), Ca-silicate
minerals are unstable and that under these conditions the reaction in Eq. 3.5 proceeds to the right;
producing Ca
2+
ions and increasing the pH of the system through the consumption of protons (11).
65

Eq. 3.5 3CaMgSi
2
O
6
+ 6H
+
(aq)
↔ Mg
3
Si
2
O
5
(OH)
4
+3Ca
2+
(aq)
+ H
2
O + 4SiO
2

Characteristic and typical of serpentinization sites, the altered fluids produced in The Cedars springs are
ultrabasic and saturated in Ca
2+
(1).
Geochemically, the composition of both the fluids and gas that emerge from The Cedars spring systems
are consistent between years with little variation in the pH, temperature, E
h
and conductivity (1). The
main spring heads that have been studied over multiple years consist of the Barnes Spring Complex
(BSC) and the Grotto Pool Spring (GPS) (both of which are discussed in depth later). The fluids that
emerge from the GPS spring head is comprised primarily of deep groundwater held within the
Franciscan Subduction Complex with a large portion of abiotically produced methane. The fluid source
for the BSC spring are a mix of the deep water found in the GPS system and meteoric groundwater, in a
ratio of ~15% : 85%, found in the BSC system, reflecting differences in the microbial communities
discussed in detail below (beginning on page 86) (12).
The major elemental constituents (Na
+
, K
+
, Ca
2+
, Mg
2+
, Cl
-
) of The Cedars springs fluid are concurrent with
other terrestrial serpentinization sites worldwide. Likewise, The Cedars system is representative of all
terrestrial serpentinization springs in high Ca
2+
(~1 mM). The ultra-basic springs of The Cedars are over
saturated in Ca
2+
and carbonate levels, resulting in an Ω > 1 for both calcite and aragonite. Morrill et al.
(1) determined that Ca
2+
concentrations were over an order of magnitude higher than necessary to be in
equilibria with carbonate levels. The over saturation of Ca
2+
results in the abiotic precipitation of CaCO
3

when the serpentinization fluids (i.e., pH 11.4-11.9) interact with atmospheric CO
2
forming a thin film of
calcite across the surface of the springs. As new fluids and gases escape the spring head, the
precipitated carbonate builds up along the edges of the spring eventually forming the layered mineral
structures shown in Figure 3.3. These carbonate structures are dominated by cyanobacteria (Shun’Ichi
Ishii, personal communication) and exhibit a seasonality in growth and presence. The carbonate
structures are dissolved and washed away during the wet months (November-April), as increased rain
66

input raises local creek levels. During the dry months, precipitation continues, and the calcite structures
build upon themselves until the rains return.
The springs of The Cedars system produce fluids of pH ≥ 11.5 due to the extremely reducing
geochemical composition of the spring fluids, with E
h
values of -640 to -580mV often being observed at
the spring heads (1). The gases produced at The Cedars have been previously determined to be
dominated by N
2
, H
2
, and CH
4
. The fluids are very low in both dissolved organic carbon and total
inorganic carbon as well as ammonium, phosphate and traditional terminal electron acceptors (oxygen,
nitrate, sulfate) (1, 12, 13). From both a thermodynamics standpoint and known metabolic paradigms,
the environment produced in The Cedars springs is generally hostile to life making the sparse
community of prokaryotic organisms endemic to the fluids a bit of an enigma. From the outset, one of
the largest hurdles in the scientific investigation of The Cedars microbial population has been the
naturally low abundance of cells. As a result, previous work has had to rely on the collection and filtering
of 10’s to 100’s of liters of fluid to increase microbial concentrations to a value more appropriate for
traditional molecular and -omics based methodologies. Given Digital Holographic Microscopy’s (DHM)
applicability and technological advantages in imaging and describing low density samples, I proposed
using The Cedars as a test site to attempt in situ cell quantification experiments and a series of
laboratory-based explorations of enriched and isolated Cedars bacteria using the orangeBox DHM field
instrument.

67

Figure 3.3 Calcium carbonate structures of The Cedars. Photographs A.-C. demonstrate the thin film of calcium
carbonate that forms at the water/atmosphere interface. When supersaturated Ca
2+
-rich fluids come into contact
with atmospheric CO
2
, calcite abiotically precipitates. Photograph D. highlights a typical area of extruding
serpentinization fluids. pH meters are used to determine the source of the fluids within the rocks. Images E.-H.
show the accumulation and layering of calcite that occurs during the dry months of the year. These carbonate
structures are dominated by cyanobacteria and other photosynthetic organisms. Image I. shows a weathered
calcite mound that has been mildly eroded by season rains. Pictures by C. Barr, J. Henderson and M. Bedrossian.
68

Specifics of springs at The Cedars and their microbial communities:
My work has been focused on three specific spring sites within The Cedars complex that have been
previously explored in multi-year investigations (Figure 3.1). The Barnes Spring Complex (BSC) is located
on a tributary of Austin Creek in the Main Canyon (elevation: 282m, N: 38°37.282’, W: 123°7.987’) and
consists of multiple discharge points that have been documented and numbered. BS5 is the largest and
most easily accessible discharge point thus garnering the most research activity within the Barnes
Springs Complex (Figure 3.4). The mouth of the spring is located at the base of a pool ~ 80 cm width x 40
cm depth. Due to low flow rates (measured at 0.1-0.25 L/min) of the BS5 spring, the pool is a stratified
system: oxic surface waters (with a redox potential of -40 mV E
h
) overlay the anoxic serpentinization
fluids (≤ -550 mV E
h
) produced at the spring head. The fluids emerging from BS5 are a mix of deep
Franciscan Complex ground water and shallow meteoric inputs (1, 12, 14). The fluids within the BS5 pool
(both surface and bottom water) have been measured at 11.5 ± 0.1 pH through multiple years. Due to
the large catch pool above the BS5 spring head, an air-surface interface of between 2,500-4,000 cm
2
is
present depending on the time of year. The catch pool has a measurable Eh gradient (-40mV at the
surface and -550 mV at the bottom), implying that atmospheric O
2
diffuses into the system. However,
dissolved O
2
is below the limit of detection of a standard oxygen electrode, suggesting depletion via
abiotic and biological usage (66). The chemical composition of the gas bubbles released from the BS5
spring contained: 53.6% N
2
, 34.0% H
2
, and 6.1% CH
4
. This ecosystem is exposed to sunlight, oxygen-
containing air from above, as well as environmental debris and organic detritus (vegetation, insects,
etc.) (1). Metagenomic analysis of the BS5 microbiome uncovered a microbial diversity and metabolic
versatility that reflects both chemistries of the deep water and shallow water constituents. Oxygen
respiration and obligate requirements of CaCO
3
in the bacteria of the mixed fed waters of BS5 are not
observed in the deep water-fed springs (GPS1). The metabolic potential of BSC appears to be dominated
by bacteria capable of hydrogen-based autotrophy.
69

Figure 3.4 Barnes Spring Complex 5 (BS5). A.) Thin film calcite overlaying the BS5 catch pool before and after B.)
removal to expose spring head. C.) The greater Barnes Spring Complex structure. D.) M. Bedrossian and C.Barr
taking field measurements with the orangeBox field DHM. Pictures by J. Henderson

The prokaryotic community of BS5 is similar to other terrestrial serpentinization sites (CVA, the
Tablelands; (15, 16)) and is dominated by phylotypes belonging to the Betaproteobacteria and Clostridia
classes. The most abundant taxon in BSC system is the Serpentinomonas genera (Figure 3.5) (Suzuki,
2017), of which 3 members are in laboratory culture. The three isolated strains represent 2 species of
the recently proposed genus. The characterization of the three isolates (referred to as: A1,H1, and B1), is
discussed in a manuscript on which I am an author, (Bird et al. (17), see Additional Publications). In the
manuscript we propose a genus description for Serpentinomonas as:
Rod shaped, motile cells 1-3 µm in length, with a single polar flagellum. Are
microaerophilic. They form small light colored (opaque creamy) colonies on plates
(<1mm) and tend to cause clumping of the calcium carbonate precipitates when grown
in liquid medium. Cells can grow autotrophically on formate or hydrogen and
heterotrophically on a range of organic compounds including acetate, butyrate, lactate
and sugars such as glucose. The cells stain gram negative. The Catalase assay is positive.
Sensitive to the antibiotics kanamycin (50 µg/ml) and gentamicin (10 µg/ml).
Phylogentically the genus is a member of the beta subclass of the Proteobacteria. The
70

base composition of the DNA of the type strain of the type species is 66.6 mol% G+C.
The type species is Serpentinomonas raichei. The type strain is A1.

Figure 3.5 Electron micrographs of Serpentinomonas racichei A1. Micrographs of S.
racichei attached to a mineral surface and on Whatman filter paper. Images by C.Barr.

Two distinct Serp. species A1 (S.raichei) and B1 (S.mccroryi) were isolated from the BS1 spring; whereas,
H1 (a sub strain of Serpentinomonas raichei) was isolated from BS5. The two sub strains of S.raichei (A1
and H1) share more than 70% genome similarity (109). A1 can grow on cyclohexane and utilize
thiosulfate as both an electron donor and acceptor. H1 is capable of metabolizing glucose via both
fermentation and respiration, as an electron donor. H1 is also capable of utilizing nitrate as a TEA,
whereas, A1 is not (Bird et al in prep). The strain B1 was designated as Serpentinomonas mccroryi, the
strain only shares 50% genome similarity with S.raicchei and was designated a novel species. The strain
cannot grow autotrophically on thiosulfate, and is unable to use it as an electron acceptor. S.mccroryi
can utilize nitrate as a TEA. (Bird in prep). Members of the Serpentinomonas require low levels of oxygen
and calcium carbonate to survive initially, reflecting the heterogenous environment of the BSC system
and the mixed chemistries of the serpentinization fluid and meteoric water. The Serpentinomonas
isolates described above are all motile organisms with single polar flagellum and are used in the
targeted chemotaxis experiments described here (see page 124).
71

The Grotto Pool Spring (GPS1) fluids are comprised solely of deep Franciscan Complex ground water
(Figure 3.6). GPS1 (elevation: 273m, N: 38°37.268’, W: 123°8.014’) consistently records the lowest redox
potentials and highest pH (≤ -650 mV E
h
, pH = 11.9+) of all The Cedars springs. The Grotto Pool Spring is
also located along Austin Creek, but discharges above the surface of the surrounding waters giving
unimpeded access to the deep anoxic fluids. The gases produced at the head of the GPS1 spring have
not been analyzed fully due to the geometry and structure of the spring (14). The gas composition is
believed to be similar to that of the Mortar Bed Spring (discussed below); however, the H
2
gas
concentration of GPS1 was theoretically determined to be close to equilibria (<4%) with that of the
spring fluids (1).

Figure 3.6 The Grotto Pool Spring (GPS1): A.) the Grotto Pool catch basin above which the GPS1 springhead is
located. B.) C.Barr and M. Bedrossian at GPS1. C.) N. Merino and C.Barr, building up a plastic sheeting dam in GPS1
to isolate the springhead from the creek fluids for sample collection. Pictures by J. Henderson.

Metagenomic analysis of GPS1 samples revealed a largely unknown microbial community. The deep-
water ecosystem is low in cell number but is predominantly represented by bacteria in the groups
Candidate Phylum OD1, Chloroflexi and Firmicutes. One metabolism that has been proposed on the
72

basis of metagenomic (genome assembly) in the group Chloroflexi is acetogenesis, utilizing hydrogen
and inorganic carbon (in the form of either bicarbonate or carbon monoxide). A number of uncultured
and un-described microbial groups (most notable and largest representation: Cand.Div. OD1) represent
~1/3 of the members of the GPS1 fluid community (14). Candidate Phylum OD1, the ‘Parcubacteria’ are
a collection of uncultured bacteria that exhibit small genomes (typically <1 Mb). These organisms are
thought to be highly specialized, as their streamlined genomes often lack many of the traditional
metabolic and catabolic functions that are broadly conserved in most life forms. The Parcubacteria are
thought to be ectosymbionts, incapable of a free-living lifestyle. In general, the Parcubacteria lack both
respiration genes (implying potential fermentation), and the genes required for amino acid, nucleotide,
vitamin and lipid biosynthesis (18). The OD1 representatives in The Cedars are unique in that they
appear to be some of the smallest reported ( ≤0.5 mB) (12). The GPS1 community is one of the most
enigmatic and largely undescribed ecosystems reported, and the chemistry of the deep water and the
evolutionary pressures of maintaining life in that system have made the study of the biome of GPS1 a
unique opportunity to explore the evolution of genomes and syntrophic microorganisms.
Mortar Bed Springs (MBS, also listed as the Campsite Spring in some early literature), is located further
down Austin creek (elevation: 269m, N: 38°37.9’, W: 123°7.59’) (Figure 3.7). MBS is a submerged spring
that discharges in the bottom of the creek. It is hypothesized to be fed primarily by the deep water fluids
as it has a similar chemistry to the Grotto Pool Spring (19). The pH of the emerging fluids has been
recorded at 11.8, with a redox potential of -613 mV. The gas composition (37% H
2
. 51% N
2
, 7.4 % CH
4
) is
also suggestive of a deep water origin of the MBS fluids.
73

Figure 3.7 Mortar Bed Spring (MBS). A.) Overview of MBS site with field equipment kept dry under blue tarp. B.)
orangeBox field DHM and IV-Bag set up for enumeration of bacteria in situ. Photographs by M. Bedrossian.

Due to the low availability of traditional terminal electron acceptors (TEA) in the deep fluids, it was
hypothesized that bacteria from the deep water may be capable of insoluble mineral reduction through
extracellular electron transport (EET, covered extensively in Chapter II: Development and use of DHM in
the investigation of microbe-electrode interactions, pgs 19-35). Rowe et al. (19) provided direct evidence
of the existence of EET-capable bacteria in The Cedars environment; however, the majority of EET-
associated strains were not representative of the bulk community composition. This finding suggests
that EET-active bacteria exist in the environment but are not represented in the dominant taxa.
Given the many directions that research at The Cedars environment is capable of exploring, there are
certain questions that Digital Holographic Microscopy can give unique insight into. The experiments,
observations and conclusions contained in this chapter, while varied in scope, are held together via the
shared technology of DHM. The work contained in this chapter reflects my desire to explore an
astrobiologically relevant site (both in situ and from the laboratory) through the use of DHM.
To this end, during the summers of 2017 and 2018, the orangeBox DHM instrument was taken to The
Cedars to examine the microbiota of the three springs discussed above. As is often the case with field
74

work, the 2017 trip was largely devoted to solving logistical problems of using the DHM in the field
(sample acquisition, sample analysis, etc.), and the return trip (2018) yielded most of the field data that
is discussed below. This work addresses two key questions concerning The Cedars sites: 1) the
abundance of bacteria at the different sites, measured in real time, 2) the motility of the microbiota –
focusing on laboratory analyses of enrichment cultures obtained from samples from The Cedars, and 3)
motility and chemotaxis of a pure culture isolate from The Cedars.
Experiment I: In situ enumeration of cell densities
In 2018, we conducted experiments aimed at quantifying the microbial cell density of the three springs
at The Cedars. The methods used were similar to those outlined in our previous JoVE publication
(Bedrossian, Barr et al., 2017) (20), with the following changes. The goal was to attain in-situ
enumeration of cell densities within the fluids of three springs in The Cedars. This was important for two
reasons: 1) the densities were known to be low (in the range of 10
3
cells per ml (or less)), numbers hard
to quantify without filtration and microscopic counting. Knowing these numbers in real time would
allow one to calculate the time needed to concentrate the cells for harvesting; and 2) getting a sense of
the motility of the populations in the different ponds might allow some insight into the levels of nutrient
limitation.
For these experiments, sterile gas-impermeable tubing was inserted directly into the spring heads of the
Grotto Pool Spring, Barnes Spring 5 and the Mortar Board Spring. A peristaltic pump was used to move
the extruding fluids to a sterile IV bag which was hung next to the orangeBox field instrument. The IV
design was required in order to prevent the exsolving gas bubbles from reaching the sample chambers
and to dampen the pulsating momentum created by the peristaltic pump, both of which would affect
biomass calculations.
75

Gasses were trapped in the IV bag, while the fluids were gravity fed through the sample chambers using
a brass flow regulator (Figure 3.8). An additional drainage line of tubing was used to keep a consistent
volume within the IV bag as the peristaltic pump was adding fluid to the bag faster than the desired
experimental flow. The flow rate of The Cedars fluids through the sample chamber were determined
using a graduated cylinder and stopwatch. Determining and establishing an appropriate flow rate was
important to ensure that each bacterial cell that passed through the FOV (360 µm x 360 µm) was
recorded in a minimum of two frames. Data was acquired at a camera frame rate of 15 FPS for one
minute. A Toughbook laptop was used to control the orangeBox field instrument and data was recorded
directly to the internal hard drive located within the DHM.
In order to determine if the observed numbers were statistically significant, we performed 7 replicates
per site, resulting in over 7 minutes of data (approximately 6,000 holograms) for each spring. See Video
3.1 to see the experimental design in action.
Processing and calculating cell densities
Raw holograms were converted to 32-bit images and the background noise was processed by medium
subtraction, allowing for more pronounced and identifiable bacteria within the data set. The number of
cells in each hologram series (i.e. data set) were counted as they entered the FOV. By knowing the flow
rate, camera speed and number of holograms acquired, the total volume of fluid is quantifiable.
Approximate cell densities are determined by dividing the number of cells detected by the total volume
imaged. The results are listed in Table 3.1.

76

Figure 3.8: Experimental design for in situ enumeration of microbial cells. A.) Experimental equipment assembled
and in use at the Mortar Bed Spring (MBS). B.) Bubbles of gases produced through the serpentinization reaction
(H
2
, N
2
and CH
4
dominated) seen within the IV collection bag. C.) IV collection bag attached to two well sample
chamber. D.) Close up of the two well sample chamber with inlet and outlet tubing. E.) C. Barr assembling
experimental apparatus at the Grotto Pool Spring (GPS1). F.) M. Bedrossian and C. Barr determining flow rates
prior to data acquisition. G.) Illustrative schematic of the experimental apparatus: I. Sterile, gas impermeable
tubing is secured to the spring head, II. A peristaltic pump brings fluids to III. A sterile IV bag used to collect fluids
and gasses, IV. Extra tubing used to control volume within the IV bag, V. Flow regulator used to establish low flow
rate through the VI. Sample chamber located within DHM field instrument, VII. Outflow of fluids from sample
chamber post-data acquisition. Photographs by J. Henderson and M. Bedrossian.

77

Table 3.1 In situ enumeration of cells from various springs at The Cedars.

The method in which we were able to directly sample and determine cell densities from serpentinization
springs is novel for several reasons. Given previous methods used to determine biomass counts (direct
sampling using filtered 10 mL samples and the use of tangential flow filters that accumulate 100’s of
liters worth of cells) have provided a rough range of estimates, the DHM and IV bag method used in our
experimental design provides accurate cell numbers with only seven minutes worth of data, which
comprises of 7 replicates, leading to higher statistical confidence. The flow through method of cell
enumeration provides the opportunity to sample directly The Cedars springs fluids without introducing
atmospheric CO
2
or oxygen, providing the most realistic and unaltered view of the microorganisms that
inhabit The Cedars environment. This technique provides an unprecedented ability to achieve reliable
cell determinations in the field, in a non-invasive and straight forward methodology. As this
experimental design was constructed with the primary focus on cell enumeration, motility rates and
swimming dynamics of individual microbial cells within the samples were not assessed. However, using
phase and amplitude reconstructions of the acquired holograms allows for us to confirm the
identification of observed particles as bacterial cells or abiotic mineral precipitation, adding an
additional level of certainty.
The relative biomass of the three springs (GPS1, BS5, MBS) reinforces our previous findings and
understandings of The Cedars spring system (12). The Grotto Pool Spring (GPS1) was determined to
contain the lowest biomass of all springs, a reflection of the mysterious deep-water community that is
limited both in diversity and metabolic capabilities. Likewise, the Mortar Bed Spring has been
78

hypothesized to be predominantly made up of the deep ground water fluid source, which is supported
by evidence that the biomass numbers are on the same order of magnitude as GPS1 (19). The most
variable and highest cell density spring was the Barnes Spring 5. Given the large catch basin, bimodal
fluid source and larger chemical heterogeneity of the spring, the higher biomass determinations reflect
the diversity of lifeforms and metabolic ecotypes that have been observed in BS5 (12). Although these
findings do not challenge our initial assumptions, the ability to determine biomass concentrations on the
scale of 10
2
cells/ mL at a 95% confidence interval in the field is impressive. The low volume of sample
fluid required to determine these concentrations in situ also illustrates the capabilities of DHM. The
average flow rate in the experiments was < 0.5mL per minute, meaning that our total seven replicate
determinations were accomplished with only 3.5mL of sample per site. For reference, the current NASA
guidelines for future astrobiological instrumentation on the proposed Europa lander is a minimal limit of
detection of 100 cells per mL with a minimal localization of 200 nm, both of which are achievable
through DHM instruments (21). These findings once again demonstrate the applicability of DHM
technology to the in situ study of low biomass aquatic environments.
Experiment II: Motility characterization of enrichment cultures:
Utilizing the high sample throughput and short time spans inherent to DHM experiments, one of the
laboratory-based questions I was interested in exploring was utilizing DHM to take a ‘shotgun’ survey of
enrichment cultures and determine if correlations between media composition and motility behaviors
can be determined. The results of that experimental suite are reported here.

79

Figure 3.9: Various enrichment cultures used in Experiment II.

Enrichment medias and growth conditions
Enrichment cultures from the MBS spring were assayed for motility to determine if any enrichment
conditions resulted in changes in the amount or nature of the cell motility. Two chemically distinct
batches (denoted as CSM2 Low and CSM2 High) of CSM2, a growth medium developed by Shino Suzuki
and Gijs Kuenen (12), was amended with various additions to test enrichment parameters. These
included: carbon sources (4 mM), amino acids, supernatant from Serpentinomonas cultures, nucleotides
and alginic acid (see Figure 3.10). 30 mL of pH 10.5 (adjusted with 1 M NaOH) CSM2 base media were
added to 100mL serum bottles and sterilized by autoclaving. The media bottles were purged with N
2
gas
and filter-sterilized experimental constituents were added via syringe. The samples we investigated and
discussed in this thesis were supplemented with H
2
gas and filtered air to achieve a final O
2
concentration of approximately 10 µM. Enrichment cultures were inoculated at The Cedars using fluids
pumped directly from the BS5 spring and stored at ambient temperature until we were back at the
University of Southern California. Once in the laboratory samples were incubated at 20°C.
80

Figure 3.10 Media recipe for CSM 2 Low and CSM2 High, with descriptions of various additives used in The
Cedars enrichment study.

81

Sampling and hologram data acquisition for motility characterization via DHM
Using the orangeBox field DHM, samples were taken from the serum bottles via 1 mL syringe. A Luer
lock 0.2 µm syringe filter was added to the syringe and used to create the filtrate used in the reference
channel of the two well sample chambers (identical to those used in the field). The enriched biomass
was then injected into the sample well. If samples appeared too dense for adequate DHM imaging, they
were diluted with filtrate. Holograms were acquired at 5 and 10 frames per second for approximately 1
minute’s time. The raw holograms underwent the normal processing pipeline of median subtraction to
remove background noise and accentuate motile cells. Medium subtracted holograms were then visually
examined to determine the extent to which non-random motility was readily apparent. Enrichment
samples were categorized based on the observed presence or absence of non-random movement and
designated as displaying predominately ‘Brownian Motion’ or ‘Direct Motility’ behavior. Results of these
observations are displayed in Table 3.2 with brown boxes representing conditions in which the majority
of cells displayed Brownian motion; whereas, green boxes represent conditions in which direct motility
was obvious and abundant.
Table 3.2: Enrichment Culture Media Composition, Sample Numbers and predominant motility characterization
used in this experiment.

Analysis of enrichment factors on the presence of non-random movement
Sample enrichment conditions that varied across microcosms and were tested to determine significance
with regard to motile characterization included: carbon source (6 different sugars), base media
82

composition (high and low), and additive type (4 supplement types). Table 3.3 displays the frequency of
each variable in both the Brownian motion and non-random movement categories. Notable findings
include the enrichments amended with lactate were largely motile; whereas, citrate cultures were more
often found to be dominated by Brownian motion. Most apparent was the distribution of enrichment
culture by base media composition, with the CSM2 High samples more often displaying non-random
movement.
Table 3.3: Frequency table of nominal variables with regard to presence of Brownian motion or non-random
movement. Percentages represent statistical composition of each variable within the motility observation type.
* represents total break down of all data sets.

To determine if the impact of each variable type was statistically significant on the presence of motility,
a three-way analysis of variance (ANOVA) statistical test was conducted. The results of which are listed
in Table 3.4. Using an alpha value of 0.05, it was determined that results of the ANOVA were significant,
with an overall F ratio (9, 36) = 2.19, and p = .046. These values indicate that there are significant
differences in motility among the different carbon sources, v/m concentrations, and additive
83

combinations; however, when broken down by variable type, only the v/m variable was statistically
significant (p = 0.004). It is important to note, that as each combination was only represented by a single
sample (n =1), the robustness of this analysis is impossible to determine. Additional experiments with
replicate samples must be run to determine true statistical significance. That said, this analysis provides
a starting point.

Table 3.4 Analysis of Variance table for motility by carbon source, base media composition, and additive.

16S sequencing and analysis
To determine if a correlation between motility and community composition could be observed, samples
were also collected for 16S sequencing. DNA was extracted from microcosms by filtering a 10 mL sample
through a 0.1 μm Omnipore membrane (25 mm diameter, Millipore, United States) housed in
a PerFluoroAlkoxy filter holder (Advantec, United States). Subsequently, the filter was immediately
stored at -80ºC inside a bead beating tube from the ZymoBIOMICS DNA Miniprep Kit (Zymo Research,
United States). The protocol from the kit was used to extract DNA, which was then sequenced by Zymo
Research Corporation. Briefly, the V4 region of the 16S rRNA gene was sequenced using the Illumina
MiSeq platform with the 2 x 300 bp paired-end sequencing kit. Raw reads were then quality filtered and
trimmed using DADA2 and sorted to amplicon sequence variants (ASV), or unique sequences (22). Reads
were taxonomically identified using the SILVA v132 database (22, 23). Afterwards, phyloseq in R was
84

used to remove contaminant ASVs observed in a negative control of filtered clean bench air during DNA
extraction (24). ASVs classified as Eukaryota, Mitochondria, and Chloroplast were also removed. The
remaining ASVs were then filtered using a prevalence threshold as 2% of the total samples. The final
read count ranged from 12,287 – 46,803. Due to the constraints of time and money, not all sampling
combinations were sequenced, although both DHM data and samples for sequencing were taken.
Samples that displayed low or negligible growth at the time of hologram acquisition were not submitted
for sequencing. Sequencing results are provided in Figure 3.11.
As previous experiments had determined, the community composition of The Cedars BS5 ecosystem
tends to be dominated by a few genera of bacteria (specifically the Hydrogenophaga/Serpentinomonas
group). Other dominant represented groups include the Alishewanella, an isolate of which has been
under active characterization In the Nealson Laboratory. Unfortunately, due to our small sample size
and incomplete data sets (samples that were not sequenced) we are not able to report any statistically
significant results with regard to community composition and enrichment conditions. To further probe
differences between enrichments, three-dimensional tracking of select motile samples were conducted
and used to determine average swimming rates between motile samples.
Tracking and data analysis
Select samples designated as “motile” were then tracked using amplitude reconstructions, achieved
through the DHM Reconstruction ImageJ plugin (25). Tracking was done with the FIJI manual tracking
function by USC undergraduate researcher, Mark Liu (26). Motility statistics and tracks were assembled
through the Manual Tracking MatLab scripts created by M. Bedrossian. 3D tracks of select enrichment
samples are displayed in Figure 3.12, while mean speeds (and standard deviations) of tracked bacteria
for the motile samples are presented in Figure 3.13.

85

Figure 3.11 Relative abundance of amplicon sequence variants of 16S community composition of select
Brownian motion and Motile enrichments. At the A.) Class B.) Familial and C). Genus level.

86

Figure 3.12 Tracks of Select Enrichment Cultures. A) Sample 309: Lactate, CSM2 High, Nucleotides. B) Sample 283:
Acetate, CSM2 High, Supernatant. C) Sample 333: Formate, CSM2 High, Nucleotides. D) Sample 317: Citrate,
CSM2 High, Nucleotides. E) Sample 301: Glucose, CSM2, Nucleotides. F) Sample 304: Glucose, CSM2 High, Alginic
Acid. G) Sample 312: Lactate, CSM2 High, Alginic Acid. H) Sample 328: Propionate, CSM2 High, Alginic Acid. I)
87

Sample 320: Citrate, CSM2 High, Alginic Acid. J) Sample 144: Lactate, CSM2 Low, Alginic Acid K) Sample 160:
Propionate, CSM2 Low, Alginic Acid. L) Sample 139: Lactate, CSM2 Low, Supernatant. M) Sample 141: Lactate,
CSM2 Low, Nucleotides.

Figure 3.13 Mean speeds of bacteria tracked in select motile data sets. Data sets are grouped by carbon source,
base media and addition type. Error bars represent standard deviation.
88

As noted above, the average speeds of tracked microbes does appear to differ with enrichment
conditions. Not surprisingly, a range of motile behaviors is also noted between samples, as some
bacteria demonstrate a broader range of environment exploration (larger distances travelled in all 3
axes) while others show much more localized motility patterns. These observations may reflect actual
differences in search strategy between the organisms; however, other factors including the local
chemistries within the enrichment microcosms or the growth phase of the bacteria observed (among
others) may also impact these findings.
In an effort to determine if relative cell density could be correlated to motility and the dominant species
of enriched samples, approximations of particle densities were derived for each data set through the
use of a particle detection script that combines the residual fringe visibility and DBSCAN clustering
algorithms (27). The residual fringe visibility algorithm focuses on areas in which the interference fringes
are most disturbed (i.e. where the light path is altered as photons pass through the microorganisms).
Areas of ‘normal’ fringe visibility (where no bacteria exist) are removed, boosting the signal to noise
ratio of the data set. This allows standard pixel thresholding to occur (i.e. definition and segmentation of
the digital data) which can be used to define areas of interest, resulting in a high throughput and low
data intensity method of highlighting particles within a hologram. Using the DBSCAN plugin (Density-
Based Spatial Clustering of Applications with Noise) clustering algorithm, a commonly used standard in
digital image thresholding, boundaries of ‘point clouds’ are determined. A point cloud is a list of 3
dimensional coordinates that represent the physical space boundaries of areas of interest (in our case
the XYZ coordinates of pixels which contain information positive for bacterial presence). The output of
the DBSCAN algorithm is an indexed list of positively identified point clouds (number of individual clouds
identified), the corresponding spatial coordinates of each cloud, and the removal of ‘negative’ point
clouds that are determined to be simply the manifestation of noise. The indexed list of positively
identified point clouds is used to define the number of particle (or cell) clusters, while the mathematical
89

mean average of the point cloud boundaries (taken individually in the X, Y and Z axes) is used to
determine the centroid of each cloud and can be used to determine the individual number of particles
or cells within the hologram data set.
The estimated particle densities for both motile and Brownian motion enrichment samples were
determined by calculating the average number of particles within five median-subtracted holograms
from each data set. The holograms were selected at random and reconstructed at 5 µm intervals
through the entire sample volume within the field of view (~360 µm x 360 µm x 550 µm). Average
number of particles per FOV and the standard deviation among the five selected holograms is also listed
in Table 3.5.
Table 3.5: Estimated Particle Densities using Fringe Visibility Processing

Binary logistical regression was conducted on the motile sample cohort to examine whether particle
density and mean speeds variables have a significant effect on the dominant taxa observed in the
90

enrichment 16S data (either Serpentinomonas or Alishewanella). Neither factor was found to have a
statistically significant impact on dominant taxa: particle density (p-value = 0.332) and mean speed (p-
value=0.526).
The results of the enrichment motility study are a bit difficult to put into context as only select samples
have been tracked and characterized fully. As additional data sets are tracked and quantified, broader
correlations between enrichment media composition and motility are expected to be parsed out. What
can be said for certain is that a large fraction of the enriched microorganisms from the BS5 spring are
motile. As the different enrichment media chemistries select for different groups of organisms (as
evident in the taxonomic analysis), in conjunction with the fact that our enrichment media contains
much higher organic concentrations than the spring fluids, the observations reported here are not able
to be put into much context regarding the original diversity of the BS5 microbial community.
However, the dramatic and widespread presence of motile bacteria found in many of the enrichment
cultures highlight the importance of motility in a traditionally difficult environment to achieve the
chemiosmotic basis of flagellar motility. When The Cedars organisms are provided a more organic-rich
environment, the number of cells exhibiting non-Brownian motility rapidly increases and swimming
characteristics are differentiable. The ability to be motile in low energy systems is an expensive but
necessary bioenergetic cost to life as active search strategies are required to survive in patchy chemical
environments. So, although these findings are not applicable to the behavior of the organisms in the
limiting conditions of the field site, a large portion of the cells observed in the enrichment cultures are
motile organisms, suggesting that the conservation of flagellar motility within the microbial community
is an important aspect of survival in the extreme environment of The Cedars.

91

Experiment III: Chemotaxis of a Cedars isolate:
Media and culture
Chemotaxis experiments utilized a previously isolated and characterized bacteria Serpentinomonas H1.
(28). The basal growth media consists of: 0.2 mM Na
2
SO
4
, 1.5 mM NH
4
Cl, 0.199 mM MgCl
2
, 0.23 mM
K
2
HPO
4
, 5 mM CaCl
2
, 1x ATCC Vitamin solution, 1x ATCC mineral solution, and 15 mM CAPS buffer (pH
11). Na
2
SO
4
, NH
4
Cl, and MgCl
2
are stored as a combined 100X stock solution. Calcium, phosphate,
vitamins, and minerals were added as separate sterile solutions after autoclaving. The basal media can
be augmented for heterotrophic growth (with the addition of 3mM acetate) or autotrophic growth
(substitution of 5mM CaCl
2
with 20mM CaCO
3
). Under heterotrophic growth conditions, cultures were
incubated in sealed serum vials with a medium:headspace ratio of 3.5:6.5, with a headspace containing
30% air, with a balance of N
2
. For autotrophic growth, 50% H
2
is added to the headspace mix.
Revival of glycerol stocks were accomplished by adding 20 mM calcium carbonate (as a suspension) to
the heterotrophic media and the addition of 50% H2 to the headspace to create a mixotrophic medium.
For each revival, a single 1 ml aliquot was used.
Revived cultures in mixotrophic media were incubated at 20 C for two weeks and monitored by direct
counts until density approached 10
6
cells/ mL. 1% of the culture volume was then used to inoculate the
heterotrophic medium which were also incubated at 20° C for approximately one week prior to
chemotaxis experiments. At the time of experimentation, heterotrophically grown cells (at an ~OD
600

0.4-0.5) were diluted into a sterile and re-gassed bottle of 0.2 µm-filtered filtrate from a previous
heterotrophic culture to a DHM appropriate density of ~ 10
4
cells/mL. Qualitative analysis of correct
magnitude of cell density and cell motility behavior were confirmed visually on the DHM prior to
chemotaxis experiments.
Sample Chamber Design
92

The sample chamber design utilized in the chemotaxis experiments is identical to that described for the
field experiments; however, the sample well has an additional channel in which a pulled glass pipette
can be inserted (Figure 3.14). The glass pulled pipettes were supplied by Iulia Hanczarek at Portland
State University. The pipettes were pulled to a final diameter of approximately 25 microns and were
sterilized in an autoclave prior to use. To accommodate the pulled pipettes, the plastic shim used for the
chemotaxis sample chambers were 0.06” thickness. The process for building the sample chamber is the
same as described for the two well sample chambers.

Figure 3.14 Chemotaxis Two Well Sample Chamber. CAD files and final epoxied chamber. Note the sample
channel (inner channel) has an additional port in which a sterile pulled pipette tip filled with a low agar chemotaxis
plug is inserted and sealed with sterile vacuum grease.
Experimental design:
Experimental salt media (used for the chemotaxis plug) was created prior to experimental runs by
amending the filtrate of a previous heterotrophically grown culture with the appropriate salts. The
amended salt media were filter sterilized a second time after the addition of the salts, stored in sterile
serum bottles and gassed with the appropriate N
2
/ filtered air ratio. During the experimental run,
amended medium was added to 5% freshly autoclaved agar to create a soft agar, which was directly
injected into the pulled pipettes. Salts used for these experiments included CaCl
2
, and NaCl. Additional
93

salts (KCl and MgCl
2
) were tested but results were inconclusive and irreproducible, thus not reported
here. The chemotaxis plug (at the end of the pipette) contained a final salt concentration of 10 mM,
with the soft agar consisting of a final concentration of 0.04%. Agar loaded pipettes were fitted into the
center well of the sample chamber and sealed with sterile vacuum grease. Serpentinomonas cultures
were injected into the sample chamber and sealed with Luer-lock caps to prevent drift. The reference
channel of the sample chamber was filled with 0.2 µm filtered media taken from the heterotrophically
grown experimental cultures.
Imaging conditions:
Holograms were acquired on a bench top Common Mode DHM using the Koala software package
(Lyncée Tec, Switzerland). A range of data acquisition rates were tested including: 15 frames per second,
5 frames per second, 1 frame every 10 seconds and 1 frame every 30 seconds. Ultimately, the 5 fps data
sets were determined to be optimal and were used for tracking and quantification.
Processing and Tracking:
Processing was similar to that mentioned previously. To illustrate another (yet more coarse resolution)
use of DHM technology and analysis, two-dimensional tracking (xy axis) was done at z axis increments of
25 µm intervals (through Fiji) and the DHM reconstruction Plugin (25, 26). All cells near each z plane
were tracked and assigned a static z value resulting in a sample volume of 21 independent z planes
covering a total depth of 500 µm.

94

Table 3.6: Number of cells and mean average speeds of the pure culture chemo(halo)taxis experiments.

Figure 3.15 Assembled z planes of tracked Serpentinomonas H1 cells across 500 µm depth and XY tracks
of select z planes.

The findings of the chemotaxis experiments revealed two main data points which can be used to assess
the impact of the chemoattractant: 1) number of cells observed/tracked (Table 3.6); and 2) the relative
95

frequency of cells when binned by mean speed (Figure 3.16). Compared to the control of non-amended
log-phase cells, the addition of CaCl
2
resulted in a 65% increase in the number of cells; whereas the
addition of NaCl resulted in a decrease of 52% in tracked cell numbers. When normalized for the
number of observations, the relative frequency of average speeds shows that both the NaCl and CaCl
2

treatments resulted in a lower average speed compared to the control. The number of cells is
dramatically increased in the CaCl
2
experiments, but we do not see a quantifiable increase in cell
swimming rates, implying the plug of high CaCl
2
concentration elicited a positive taxis motile response,
and not a kinesis response. The NaCl experiments shifted the mean swimming rates down and resulted
in lower cell numbers in the FOV, eliciting a negative taxis response, once again confirming the
physiological observations that organisms of The Cedars springs are adapted to thrive and prefer low
Na
+
concentrations.

Figure 3.16 Relative Frequency of Tracked Cells Binned by Mean Speed. Chemotaxis experiments using filtered
growth media, 10 mM NaCl and 10 mM CaCl
2
.

96

Given that this experiment inherently has a lower quantitative fidelity (due to the 3D assembly of xy
tracking over true xyz tracking of individual cells), it provides another method in which experiments such
as bulk chemotaxis or fluid flow dynamics can be described. In these experiments, determining the
collective behavior of the cells as a whole was more important than that of individual microbes. That
said, due to the nature of DHM, the ability to track each cell in a true 3D fashion is still possible, as the
raw holograms can always be reconstructed again for further and finer-resolution analysis.
Conclusion
The results presented here clearly demonstrate several potential uses for DHM approaches to
observing, quantifying, and understanding many aspects of microbial life, including motility, and tactic
abilities, and the potential applications to studies in both geobiology and astrobiology are exciting. The
utilization of DHM to explore alkaliphilic microorganisms in the field, during enrichment cultures and in
the study of pure cultivars enabled the observation of multiple aspects of microbial life in The Cedars.
Although much is still to be learned through additional experiments, we have provided a proof of
concept and preliminary results that indicate future directions of research. In an era dominated by the
molecular techniques of the -omics revolution, the ability to directly observe and characterize the
behavior of elusive microorganisms in situ and in the laboratory can provide insight into some of the
fundamental unknowns and natural ecophysiology of unique lifeforms. The abundant presence of active
microbial motility in the serpentinizing fluids of The Cedars attests to the importance of motility in even
the most thermodynamically challenging environments. As the survey of extreme lifeforms on Earth and
the search for life beyond the confines of our planet continues, DHM provides a useful tool in exploring
non-traditional microbial communities.

97

References
1. P. L. Morrill, J. G. Kuenen, O. J. Johnson, S. Suzuki, A. Rietze, A. L. Sessions, M. L. Fogel, K. H.
Nealson, Geochemistry and geobiology of a present-day serpentinization site in California: The
Cedars. Geochim. Cosmochim. Acta. 109, 222–240 (2013).
2. R. G. Coleman, in In Ophiolites and Oceanic Crust: New Insights from Field Studies and the Ocean
Drilling Program., and A. N. Y. Dilek, E. Moores, D. Elthon, Ed. (Geological Society of America,
Boulder, CO, Geological.), pp. 351–364.
3. M. C. Blake, E. H. Bailey, C. M. Wentworth, The Cedars Ultramafic Mass, Sonoma County,
California. USGS Open-File Rep. 2012–1164, 1–16 (2012).
4. G. Barnes, Ivan, LaMarche, V.C., Himmelberg, Geochemical Evidence of Present-Day
Serpentinization. Sci. (New York, N.Y). 156, 830–832 (1967).
5. Nicholas L. Taliaferro, Franciscan-Knoxville Problem. Am. Assoc. Pet. Geol. Bull. 27, 12–12 (1943).
6. R. G. Coleman, D. E. Lee, Metamorphic aragonite in the glaucophane schists of Cazadero,
California. Am. J. Sci. 260, 577–595 (1962).
7. E. H. Bailey, J. Blake, M.C., D. L. Jones, On-Land Mesozoic oceanic crust in California Coast Ranges.
U.S. Geol. Surv. Prof. Pap. 700, C70–C81 (1970).
8. R. Raiche, The Cedars: Sonoma County’s Hidden Treasure. Fr. 37, 3–15 (2009).
9. C. Oze, S. Fendorf, D. K. Bird, R. G. Coleman, Chromium geochemistry in serpentinized ultramafic
rocks and serpentine soils from the Franciscan complex of California. Am. J. Sci. 304, 67–101
(2004).
10. N. H. Sleep, D. K. Bird, E. C. Pope, Serpentinite and the dawn of life. Philos. Trans. R. Soc. B Biol.
Sci. 366, 2857–2869 (2011).
11. R. B. Frost, J. S. Beard, On silica activity and serpentinization. J. Petrol. 48, 1351–1368 (2007).
12. S. Suzuki, S. Ishii, T. Hoshino, A. Rietze, A. Tenney, P. L. Morrill, F. Inagaki, J. G. Kuenen, K. H.
Nealson, Unusual metabolic diversity of hyperalkaliphilic microbial communities associated with
subterranean serpentinization at the Cedars. ISME J. 11, 2584–2598 (2017).
13. E. Kohl, Lukas. Cumming, Exploring the metabolic potential of microbial communities in ultra-
basic, reducing springs at The Cedars, CA, USA...., 1203–1220 (2016).
14. S. Suzuki, S. Ishii, A. Wu, A. Cheung, A. Tenney, G. Wanger, J. G. Kuenen, K. H. Nealson, Microbial
diversity in the Cedars, an ultrabasic, ultrareducing, and low salinity serpentinizing ecosystem.
Proc. Natl. Acad. Sci. U. S. A. 110, 15336–15341 (2013).
15. J. M. Marques, P. M. Carreira, M. R. Carvalho, M. J. Matias, F. E. Goff, M. J. Basto, R. C. Graça, L.
Aires-Barros, L. Rocha, Origins of high pH mineral waters from ultramafic rocks, Central Portugal.
Appl. Geochemistry. 23, 3278–3289 (2008).
16. I. Tiago, A. Veríssimo, Microbial and functional diversity of a subterrestrial high pH groundwater
associated to serpentinization. Environ. Microbiol. 15, 1687–1706 (2013).
17. L. Bird, J. G. Kuenen, M. Osburn, N. Tomioka, S. Ishii, C. Barr, K. H. Nealson, S. Suzuki,
98

Serpentinomonas raichei and S. mccroryi, gen. nov., sp. nov., two hyperalkaliphilic, hydrogen-
oxidizing microaerobic bacteria isolated from terrestrial serpentinizing springs (in prep).
18. W. C. Nelson, J. C. Stegen, The reduced genomes of Parcubacteria (OD1) contain signatures of a
symbiotic lifestyle. Front. Microbiol. 6, 1–14 (2015).
19. A. R. Rowe, M. Yoshimura, D. E. LaRowe, L. J. Bird, J. P. Amend, K. Hashimoto, K. H. Nealson, A.
Okamoto, In situ electrochemical enrichment and isolation of a magnetite-reducing bacterium
from a high pH serpentinizing spring. Environ. Microbiol. 19, 2272–2285 (2017).
20. M. Bedrossian, C. Barr, C. A. Lindensmith, K. Nealson, J. L. Nadeau, Quantifying microorganisms at
low concentrations using digital holographic microscopy (DHM). J. Vis. Exp. 2017, 1–11 (2017).
21. K. Hand, A. Murray, J. Garvin, W. Brinkerhoff, B. Christner, K. Edgett, B. Ehlmann, C. German, A.
Hayes, T. Hoehler, S. Horst, J. Lunine, K. Nealson, C. Paranicas, B. Schmidt, D. Smith, A. Rhoden,
M. J. Russell, A. Templeton, P. Willis, R. Yingst, C. Phillips, M. Cable, K. Craft, A. Hofmann, T.
Nordheim, R. Pappalardo, and the P. E. T. (2017), “Report of the Europa Lander Science
Definition Team” (2017).
22. B. J. Callahan, P. J. McMurdie, M. J. Rosen, A. W. Han, A. J. A. Johnson, S. P. Holmes, DADA2:
High-resolution sample inference from Illumina amplicon data. Nat. Methods. 13, 581–583
(2016).
23. B. J. Callahan, P. J. McMurdie, S. P. Holmes, Exact sequence variants should replace operational
taxonomic units in marker-gene data analysis. ISME J. 11, 2639–2643 (2017).
24. P. J. McMurdie, S. Holmes, Phyloseq: An R Package for Reproducible Interactive Analysis and
Graphics of Microbiome Census Data. PLoS One. 8 (2013), doi:10.1371/journal.pone.0061217.
25. D. Cohoe, I. Hanczarek, J. K. Wallace, J. Nadeau, Multiwavelength Digital Holographic Imaging and
Phase Unwrapping of Protozoa Using Custom Fiji Plug-ins. Front. Phys. 7, 1–14 (2019).
26. J. Schindelin, I. Arganda-Carreras, E. Frise, V. Kaynig, M. Longair, T. Pietzsch, S. Preibisch, C.
Rueden, S. Saalfeld, B. Schmid, J. Y. Tinevez, D. J. White, V. Hartenstein, K. Eliceiri, P. Tomancak,
A. Cardona, Fiji: An open-source platform for biological-image analysis. Nat. Methods. 9, 676–682
(2012).
27. M. Ester, H.-P. Kriegel, J. Sander, X. Xu, A Density-Based Algorithm for Discovering Clusters in
Large Spatial Databases with Noise. Proc. 2nd Int. Conf. Knowl. Discov. Data Min., 226–231
(1996).
28. S. Suzuki, J. G. Kuenen, K. Schipper, S. Van Der Velde, S. Ishii, A. Wu, D. Y. Sorokin, A. Tenney, X.
Meng, P. L. Morrill, Y. Kamagata, G. Muyzer, K. H. Nealson, Physiological and genomic features of
highly alkaliphilic hydrogen-utilizing Betaproteobacteria from a continental serpentinizing site.
Nat. Commun. 5 (2014), doi:10.1038/ncomms4900.

99

Closing remarks on DHM and The Cedars work:
The microbial community of The Cedars represents one of the most enigmatic ecosystems on Earth.
Underscoring how little we know about the lifeforms that inhabit our own planet, we investigate The
Cedars in hopes that it can teach us about how life may persist elsewhere. The experiments listed in the
chapter were designed to highlight the applicability of DHM to an analogue field site for astrobiological
investigation.
The most fundamental (and still unanswered) question faced by the microorganisms of The Cedars
serpentinization fluids is how they are able to maintain a chemiosmotic gradient in such a reducing
external environment. The challenge of maintaining a proton motive force impacts both the cell’s
abilities to create ATP, but also the mechanism through which polar flagellated cells are capable of
achieving motility. As motile bacteria must allocate this shared resource amongst the functions of ATP
synthesis, motility and metabolite transport, many alkaliphilic and marine organisms have substituted
Na
+
as the ion of choice, or in conjunction with H
+
to overcome this hurdle. However, we have reached
the conclusion through multiple lines of inquiry that this must not be the case for the microorganism of
The Cedars. Our work using multiple Cedars isolates have shown an extremely low tolerance for Na
+

concentrations, and the chemotaxis experiments described in the above chapter clearly demonstrate a
negative taxis response to the presence of Na
+
. Although the use of Ca
2+
, as a chemiosmotic ion has
been suggested previously, no clear evidence of an organism utilizing this anion in this capacity has been
definitively shown. Given the highly saturated Ca
2+
concentration of Cedars spring fluids, the
requirement of Ca
2+
to sustain growth of enriched and isolated members of The Cedars community, and
its positive taxis response in the experiments described above, it is tempting to suggest that this may be
the case for the bacteria of the Cedars environment. However, we do not see an increase in relative
frequency of swimming speeds associated with higher Ca
2+
concentration that we would expect if the
100

cells were translocating Ca
2+
for flagellar motility, at least with Serpentinomonas H1. As more Cedars
isolates have been brought into culture, repeating these experiments with new organisms will be an
important step. Due to the heavy emphasis of the Barnes Spring community, envisioned future work
would likely focus more on the nature of the other Cedars microbial communities (GPS1 and MBS) in
which the deep groundwater fluids play a dominant factor. What is undisputedly evident, is the high
presence of motility in The Cedars microbial community. As the first successful isolates (the
Serpentinomonas spp) obtained from the environment were confirmed to be motile, we were not sure
how widespread the ability was among the other organisms endemic to the springs. The observed
motility of the Paenibacillus and Alishewanella isolated strains, and the variety of motile behaviors seen
in the enrichment experiments underscore the importance of non-random movement even under some
of the harshest external environments. In a location in which we could assume the cost of motility may
outweigh the benefits, we see a thriving community of motile organisms, providing credence to the use
of DHM and motility as a biosignature in the search for life off planet.

101

Chapter VI: Investigations into Magnetotactic Bacteria through
DHM

Overview of chapter:
As DHM facilitates the ability to observe and document the natural motile behaviors of bacteria in a
practically non-invasive method, the technology is readily applicable to the study of magnetotactic
bacteria (MTB) and the magnetotaxis motile response that is characteristic of these organisms. The
exploration of magnetotactic bacteria from within environmental samples via DHM discussed within this
chapter reflects a novel and valuable approach in MTB research. Because of the poor success to date in
the cultivation of the MTB, research has focused primarily on the molecular biology of the
magnetosome organelle in select model organisms. Thus, much of the ecophysiology and fundamental
behavior(s) of MTB in the natural environments remains unknown. To best illustrate the advantages of
incorporating DHM technology into the study of MTB, an overview of MTB research and the lines of
inquiry that have dominated the field over the past decades is provided, followed by the experimental
description and results obtained in our novel DHM-based methodology. The use of a novel iron-free
DHM in conjunction with a triaxial magnetic coil system provides the ability to elicit and characterize the
magnetotaxis response of MTB from within a heterogenous population at magnetic field intensities at
and below that of the Earth’s nominal magnetic field strength (≤ 0.5 G or 50 µT).
Although MTB have long been assumed to detect and orient to the Earth’s magnetic field (hypothesized
since the time of discovery of MTB in the 1970’s), this work represents the first direct observation and
quantification of MTB behavior at realistic field strengths. As mentioned later in the chapter, the
discrepancy noted between the magnetotaxis response of environmental samples and that of pure
culture laboratory strains of MTB highlights the value of the DHM-magnetic coil system and underscores
the need for additional in situ experiments of MTB if we hope to truly understand the role and behavior
102

of these enigmatic and diverse organisms in the natural setting. The investigation of magnetotactic
bacteria at realistic magnetic field strengths is the ultimate example of geobiology as it explores the
existence and relationship between magnetotactic microorganisms and the global magnetic field.
Although this work provides direct evidence of the inherent variability in the magnetotaxis response of
MTB collected from a single site, we provide unequivocal confirmation of the technological advantages
of DHM when coupled to a magnetic coil system in the study of magnetotactic bacteria.
Background of research:
The research covered in this chapter was born of a collaboration between the USC-Caltech DHM team
(M. Bedrossian and myself) and Ken Lohmann, Prof. of Biological Sciences at the University of North
Carolina, Chapel Hill through our mutual contacts at the Air Force Office of Scientific Research (Willard
Larkin & Patrick Bradshaw). The DHM-magnetic coil experiments were conducted at UNC Chapel Hill in
Dr. Lohmann’s laboratory during November of 2019. Experiments included investigations of the
magnetotaxis response in pure culture Magnetospirillum AMB-1 (wildtype and the magnetotaxis
mutants ΔMAI), environmental samples from Baldwin Lake, California (1), and sub-micron magnetite
particles as an abiotic control. The work reported in this chapter would have been impossible without
the help of Leila Mahrokh, Tingting Yang, Radu Popa, Mark Liu, and Ken Lohmann.
Synopsis of DHM experiments:
We report here the behavior of an enriched natural population of magnetotactic bacteria (MTB) from a
freshwater lake, using a non-magnetic digital holographic microscopy (DHM) instrument in conjunction
with a magnetic coil system to quantify the magnetotactic responses of individual MTB cells. The natural
population contained a mixture of morphologically distinct, magnetotactic cells that displayed both
northern and southern polarity. The magnetotaxis response of MTB were quantified under a variety of
magnetic field strengths below those normally encountered at the surface of the Earth (~50 µT), and
103

provided the following conclusions: 1) some MTB respond to magnetic fields as low as 10 µT, 2) the
response of the bacteria to changes in the direction of the magnetic field is rapid, even at low magnetic
field strengths, and 3) the DHM system provides a powerful tool for the examination of motility and
taxis under natural conditions, with the ability to quantify responses without the need for strong
magnets or magnetic interference from the instruments.
Overview of magnetotactic bacteria:
Magnetotactic bacteria are a diverse group of highly motile gram-negative microorganisms which exhibit
the common ability to sense and orientate along magnetic field (B-fields) lines. Although the discovery
of MTB is usually attributed to Richard Blakemore in 1975 (30), the existence of MTB was initially
reported by the Italian physician Salvatore Bellini in 1963 (2)(3), (which were translated in 2009 through
the efforts of Richard Frankel). Bellini observed and described freshwater bacteria that he termed
‘magnetosensitive’ (4). Today, the motile response of a microorganism stimulated by the presence of a
magnetic field is known as magnetotaxis. Ubiquitous in aquatic sediment environments at or below the
oxic/anoxic interface, MTB are largely considered ‘gradient’ organisms that thrive in geochemically
stratified environments that are difficult to reproduce in the laboratory (5). As it is relatively easy to
identify the presence of MTB from within an environment using strong bar magnets, the inability to
successfully cultivate the majority of MTB in the laboratory has provided a large scientific hurdle,
influencing the current state of MTB research and shaping the lines of inquiry that have been pursued.
Out of necessity, the few pure culture model organisms that have been cultivated have dominated the
focus of MTB research, providing highly detailed studies into the specifics of a few magnetotactic
microbes (predominately the Alpha-proteobacteria strains belonging to the Magnetospirillum genus or
the Deltaproteobacterium Desulfovibrio RS-1) for which genetic tools have been developed. However, as
more modern techniques (provided largely by the -omics revolution) have been utilized in the study of
MTB, additional phylogenetic taxa have been identified to contain MTB. Currently, the
104

Alphaproteobacteria, Gammaprotetobacteria, Deltaproteobacteria, Nitrospira and the candidate
divisions WS3 (candidate phyla Latescibacteria) and OP3 (Omnitrophica) have all been observed to
contain extant MTB species (6) (Figure 4.1A). Although much knowledge has been gained by the study of
MTB cultivars, almost certainly, many of the findings that have been previously deemed universal
among MTB will not be applicable to the diverse clades of MTB now recognized (and growing).
105

Figure 4.1 Diversity of MTB phylogeny, MAI gene sequences and generalized magnetosome formation. A.)
Phylogeny of identified MTB (cultured and uncultured) from the Proteobacteria, Nitrospirae and Candidate OP3
(Planctomycetes-Verrucomicrobia-Chalamydiae (PVC)) phylums, based on neighbor-joining analyses. MTB are in
bold. From Lefevre and Wu (7). B.) Sequencing of mam genes from various MTB isolates. Colored arrows represent
mam genes that are common to all MTB; whereas, grey arrows and letters represent mam genes that are found in
some MTB but not others, from (26). C.) magnetosome formation based on the Magnetospirillum model: I. Cell
prior to magnetosome formation, II. Invagination of the membrane to form a lipid bilayer membrane and the
establishment of the magnetosome filament (orange dashed line). III. Import of soluble Fe into the magnetosome
106

vesicle, redox driven control of Fe-mineral biosynthesis, and movement of the magnetosome organelles towards
the magnetosome filament. IV. Final alignment of the magnetosome organelles on the filament, allowing for the
establishment of the single dipole action of the chain and the ability for the cells to align to geomagnetic field lines.

The magnetosome organelle
Microbial magnetic field sensing is dependent upon the presence of intracellular organelles called
magnetosomes: membrane-encased crystalline magnetic minerals that function as a compass needle,
allowing the bacterium to orient in the preferred B-field direction. Magnetosomes consist of pure single-
domain crystals of the magnetic minerals magnetite (Fe
3
O
4
) or greigite (Fe
3
S
4
) that are encased within a
lipid-bilayer membrane vesicle. Mature crystals are typically within the 50-120 nm diameter size and are
biomineralized within the magnetosome membrane (2)(8). Often found in a single chain or aligned in
multiple chains within the cytoplasm, the morphology and arrangement of the magnetosome crystals
differs substantially among the MTB clades (9) (Figure 4.2). The choice of iron mineral utilized by a MTB
has been proposed to be dictated by the location of the organism along the O
2
/H
2
S gradient of sediment
ecosystems, with greigite-forming MTB preferring more reduced environments (10).
As the magnetosome minerals consist of pure, single domain, biomineralized magnetic crystals, MTBs
have attracted the interest of many different scholars in addition to physicists and microbiologists: e.g.,
those working in material science, biotechnology, and some aspects of medicine. In the microbial world,
the interests and advances have been mostly with regard to the genetics and molecular biology of
magnetosome formation and operation (of which many reviews have been published, see (14, 15)).
107

Figure 4.2 Magnetotactic bacteria with visible magnetosomes. A.) Assorted MTB collected from Baldwin Lake, CA.
B.) Magnetospirillum from (11). C.) MTB with ‘bullet shaped’ magnetosome minerology (16). D.) Rod shaped MTB
reported by Blakemore (12). E.) Multicellular magnetotactic prokaryote (MMP) from (13).

These advances have led to the elucidation and identification of conserved gene clusters involved in
both the synthesis of the magnetite mineral (Fe
3
O
4
), and the synthesis of the magnetosome membrane
and larger organelle structure, collectively termed as a magnetosome gene island (MAI). Even as much
of the fundamental genomic research has been conducted using the aforementioned model organisms,
additional variations and motifs of the MAI model have been identified in the genomes of many
uncultivated MTB (Figure 4.1B). Although the genetic underpinnings of the magnetosome organelles
that provide the ability to perform magnetotaxis has been the focus of scientific investigation for
decades, the ecophysiological role of the behavior itself has been more difficult to ascertain.
Theorized evolution and function(s) of magnetotaxis:
Functionally, a chain of magnetosomes is generally accepted to act as an internal compass needle,
allowing a bacterium to orientate along the Earth's magnetic field and provides a detection sensor for
the magnetotaxis response. This is attributed to the fact that chains of magnetosome minerals within
MTB are oriented with parallel axes of magnetization, giving the entire chain the behavior of a single
magnetic dipole. In the early days of MTB studies, it was concluded that in the northern hemisphere,
108

bacteria swam towards the northern pole, thus always moving downwards and away from oxygen;
whereas, in the southern hemisphere the MTB were south seeking (16) (Figure 4.3). Given that most of
these microbes prefer very low levels of oxygen, this was hypothesized to be a mechanism that allowed
them to swim downwards in mixed sediments (6). However, with ensuing reports of isolates that “don’t
obey these rules”, and the unexplained existence of about equal numbers of north and south seeking
MTB in equatorial regions (17), the raison de’etre of mangetotaxis remains a subject of debate. Many of
the early observations of model MTB organisms have led to widespread assumptions that have needed
revision as more MTB are discovered, brought into culture, or sequenced through genomics. With the
discovery of larger, more complex MTB such as the Candidatus Magnetobacterium bavaricum, or
Candidatus Magnetoglobus multicellularis, many previous assumptions have had to be rewritten and
many discoveries are yet to be made (5)(11). As the number and diversity of organisms that are capable
of magnetotaxis has increased, it has become more difficult to come up with a convincing argument for
the “why” of magnetotaxis.
The evolution of the magnetosome organelle has been proposed to have occurred around the time of
the Archean Eon divergence of the Proteobacteria and Nitrospirae (18). Phylogenetic comparisons of
cultured and uncultured MTB through both 16S rRNA gene sequences and Mam protein amino acid
sequencing show strong homology, implying a monophyletic origin of the trait prior to the divergence
~3.0 Ga, additionally suggesting that the progenitor of all Proteobacteria lineages may have been a MTB
(19).
109

Figure 4.3 General conception of MTB behavior in aquatic environments. It is important to recall that the global
dipole of the Earth’s geomagnetic field is inverted with regard to the nomenclature of the geographical pole. The
magnetic pole located in the Northern hemisphere is functionally determined to be a southern pole (e.g. the pole
in which the magnetic field is headed and the direction the north needle of a compass is attracted towards). The
geomagnetic field exits the Earth in the southern hemisphere (thus it is functionally viewed as a northern magnetic
pole). For this reason, MTB in the Northern hemisphere (that exhibit northward polarity) are attracted towards the
southern pole of a bar magnet. In the environment, MTB are thought to use the geomagnetic field as a method of
locating the oxic/anoxic transition zone. Based on Uebe and Schüler (6).

Although horizontal gene transfer is an appealing mechanism through which the magnetotaxis genes
could be widespread among the various MTB clades, recent metagenomic analysis of 28 uncultured MTB
suggests that the behavior may have been a property of the progenitor of the various clades that
contain extant MTB. To this end, it has been proposed that the loss of the magnetosome genes and
functions has occurred many times throughout the evolution of the various clades giving rise to lineages
of non-magnetotactic bacteria (20). Potentially of interest to the experiments reported in this chapter,
the abundances of magnetofossils (the preserved magnetic minerals of MTB in the fossil record) are
notably diminished during times of low magnetic field intensity during periods associated with
geomagnetic field reversals. As the intensity of the Earth’s magnetic field and orientation has changed
110

throughout at least the last 3.5 Ga, the adaptation and evolution of MTB throughout geological history
remains a tantalizing question to ponder (21). As the evolution of the magnetosome organelle in
prokaryotes appears to be ancient, the variety of extant MTB organisms and their clade-specific nuances
is not surprising. A commonly attributed factor to the rise of the now ubiquitous and abundant
Proteobacteria lineages was the introduction of low atmospheric oxygen levels, which required
prokaryotic life to survive and adapt to a gradient of oxygen concentrations ranging from anaerobic to
microaerophilic (22). The presence of “oxygen oases” have been proposed to have existed in the early
anoxic ocean around 2.8-3.2 Ga and has been hypothesized to facilitate the redox gradients and
stratified systems in which MTB (and the Proteobacteria in general) may have adapted to exploit (18).
Although magnetoreception abilities have been observed in protists and more complex eukaryotes
(bees, sea turtles, etc), there are no known representatives of the Archaeal domain in which the
magnetotactic ability has been identified. As the sensing mechanism of the magnetotaxis response relies
on the accumulation and coordination of multiple magnetosomes, it is most likely that the original
function of the magnetosome was not involved with magnetism in any form.
Exaptation is the evolutionary process in which a trait or mechanism evolves to produce a function
distinct from its original purpose. The magnetotaxis exaptation hypothesis put forward by Wei Lin (2020)
proposes that the first intracellular biomineralization of iron-oxide nanominerals functioned primarily as
a method of removing reactive oxygen species (ROS) produced through ultraviolet radiation, the Fenton
reaction or physical mineral alterations (such as impact shocks). ROS production via ultraviolet radiation
was believed to be a dominant selection pressure on early microbial life in shallow marine systems, as
the lack of a substantial ozone layer in the atmosphere permitted much higher UV penetration. Under
near neutral pH (such as in the cytoplasmic interior of cells), iron-oxide particles have been shown to
catalyze hydrogen peroxide to water and elemental oxygen (23). The production of intracellular iron
oxide minerals (proto-magnetosomes) would allow microorganisms an advantage in overcoming
111

internal ROS accumulation, prior to the development of magnetotaxis. As cells produced increasing
numbers of iron-oxides (to mitigate ROS), the production and accumulation of these ferric minerals
would facilitate the single dipole nature of magnetosome chains in extant MTB. This unintended
development would give the bacteria an additional advantage: the ability to utilize geomagnetic field
lines to provide directionality away from the highly radiated and relatively more oxygenated surface
waters (24). In short, the exaptation hypothesis theorizes that proto-magnetosome minerals were used
primarily as a way of mitigating damage produced through ROS toxicity but was later utilized (and
adapted to be used) as a sensor, allowing the cells to navigate away from harmful surface water
environments.
The concept of magnetotactic prokaryotes utilizing magnetosome minerals as a redox ‘geobattery’ was
proposed by Vali and Kirschvink in 1991 (3). Although little evidence of this ability has been borne out,
the concept is still worth consideration. Just as many sulfur bacteria have been observed storing
intracellular sulfur globules for metabolic use at a later time, Vali and Kirschvink envision the partial
oxidation and reduction of the magnetosome minerals in conjunction with the reduction potential of the
cell’s environment (25). According to this hypothesis, when a MTB cell is in more oxidizing waters, the
magnetic mineral is partially oxidized, giving the cell a source of metabolically available electrons.
Alternatively, when the cell migrates to a more reducing environment, the minerals would be reduced
back to steady state mineral stoichiometry (10). Following this line of thought, the authors raise the
possibility that intracellular greigite minerals could have evolved first as an iron-sulfide based redox
energy storage system and later evolved into the extant magnetotaxis system as the minerals were
aligned within the cell to produce the dipole effect. However, this conclusion is contentious, as Abreu et
al. (26) used genomics to compare the mam genes of a range of MTB and proposed that the bullet
shaped magnetite magnetosome produced in the deepest-rooted magnetotactic clades (Nitrospirae and
more ancient Proteobacteria lineages) represent the earliest magnetosome organelles.
112

Another hypothetical alternative function of the magnetosome complex couples motility and flagellar
motion to the magnetosome. Philippe and Wu (27) propose that a magnetic torque is generated on the
magnetosome chain when the alignment is not with the geomagnetic field. The generated torque
affects the MamK cytoskeletal filaments and its interaction with the MCP-like protein Amb0994, which
in turn elicits a response form the CheA-CheY signal transduction system to control flagellar rotation
(17). Although this work proposes a novel function of magnetosomes in Magnetospirillum AMB-1, it may
actually provide clues as to how the detection of the geomagnetic field by the magnetosome chain is
translated to active motility.
What is known about the magnetotaxis motile response
The magnetotaxis response of MTB often display a ‘polarity’ in conjunction with the latitudinal location
of the bacteria on Earth. In the northern hemisphere, MTB typically display a northward polarity and
swim toward geomagnetic North; likewise, in the southern hemisphere MTB predominately swim
towards the direction of the geomagnetic southern pole. Polarity has been proposed to aid the bacteria
in navigating a three-dimensional environment, by focusing motility in a more linear (up and down)
fashion allowing for a more efficient method of navigating vertical chemical gradients of stratified redox
environments.
The magnetotaxis response is deeply integrated into the general chemotaxis signaling/sensory pathway
of MTB, evident as both magneto-aerotaxis and magneto-phototaxis responses have been observed in
MTB. MTB that display magneto-aerotaxis behaviors can be differentiated into two types, polar and
axial. Microbes that undergo polar magneto-aerotaxis swim in a unilateral direction in conjunction to
the magnetic field, dependent on oxygen concentrations (high O
2
concentration = increased movement
in the preferential B-Field direction that traditionally corresponds with less O
2
). Axial magneto-aerotaxis
capable MTB switch between magnetic field orientations (northward vs southward) dependent on local
113

O
2
concentrations to find zones of optimal O
2
gradients. Lefevre et al. (28) observed six different
variations of the polar and axial magneto-aerotaxis in cultured MTB strains, demonstrating that
although the oxygen gradient and magnetic field sensing can be utilized together, the navigational
inputs for MTB motile responses are more complex than simply regulating these two environmental
factors (8). Another line of evidence of the close coupling between general chemotaxis and magnetic
field sensing is the high number of chemotaxis and signal transduction proteins found in MTB. Multiple
copies of the chemotaxis operon and dozens of putative chemoreceptor genes have been found in MTB
genomes when compared to more traditionally studied bacteria (6).
Although much has been discovered through the use of model MTB organisms and more recently
through the various tools of the -omics revolution, it is important to discuss some of the shortcomings of
these techniques in the study of MTB and their behavior/role in the natural environment. The in depth
study of model organisms like the Magnetospirillum allows us to discern the biological mechanisms and
pathways utilized by certain clades of MTB, but as the range of MTB containing clades increases the
impact of these discoveries diminishes as a whole. Given that the majority of MTB are not in culture, the
impacts of the laboratory environment on pure culture strains skews our understanding of the
microorganisms in the natural environment. Complicating this is the fact the magnetosome gene island
(MAI) genes are easily lost in MTB cultures when grown under rich media or oxic conditions, and
multiple growth cultures must be inoculated to ensure that at least some of the MTB produce
magnetosomes.
The magnetotactic response of MTB studied in the laboratory is more often a reflection and
manifestation of the bacterium’s environment and growth conditions than their natural behavior. In
order to understand the larger picture and artifact-free behavior of MTB, in situ experiments are an
important piece of MTB research that has been lacking substantially in the last few decades.
114

When the dynamics of the magnetotaxis response of MTB have been studied in the laboratory, high
powered magnets (10’s- 100’s Gauss) are often used. These are orders of magnitude higher than the
Earth’s nominal geomagnetic field which is approximately 0.5 G. Observations with iron containing
microscopes introduce magnetic field distortions and many of the early (and quickly canonized) research
explored MTB response to imposed magnetic fields while ignoring the ever present geomagnetic field.
Although the pioneering work by Blakemore and Kirschvink, among others, demonstrated the sensitivity
of detection for MTB in changing magnetic fields, the work does not represent an accurate picture of the
bacterial behavior in the natural environment. In light of these technological and experimental hurdles,
much of the modern MTB research has not focused on magnetotaxis of the various MTB organisms, but
on the genomic, functional and signal transduction pathways utilized by MTB.
The following issues stand out when one reads the literature as an outsider: 1) very few, perhaps no
quantitative studies of magnetotaxis have been carried out under conditions simulating the average
magnetic field strength of the Earth (about 50 µT) (i.e., for laboratory studies, selecting mutants, etc.,
magnetic field strengths used are typically 6-10X of the Earth’s field strength (or more)) (16, 28, 29), 2)
no convincing arguments have been made regarding the function(s) of the magnetosomes (e.g., is there
one function for all, or are there many functions, with different bacteria adapting this ability to their
own needs), and 3) there is no convincing explanation for why magnetotaxis is so easily lost in
laboratory cultures of MTB. These of course are questions relating to the ecophysiology of the microbes.
Simply put, the ecophysiology of magnetotactic bacteria in the natural environment is relatively
unknown. MTB have long been considered important in the iron cycle of aquatic systems, as they
remove dissolved Fe from the biologically available Fe budget through biomineralization and
sedimentation (14). Although biologically important for all lifeforms, approximations of between 1-50%
of all Fe input into ecosystems may be removed through MTB activity (30). The lack of magnetotactic
Archaea is another looming question into the initial and extant function(s) of the magnetotaxis response
115

in the environment. It is plausible (and likely) that the invention of magnetosome biomineralization was
not initially used for sensing the external environment, but as a method of rapidly scavenging Fe. Once
enough magnetosomes were produced within the cell, the larger magnetic properties of the minerals
took effect and led to magnetic field sensing abilities. Although this does not adequately answer the
‘Why?’ of extant MTB microbes, the true ecophysiology of MTB will never be understood without
investigations of MTB under realistic field strengths.
Advantages of a DHM/magnetic coil-based experimental design
The pioneering work of Blakemore reported qualitative observations of the movement and
accumulation of MTB under natural conditions (12, 31), as well as when additional magnetic fields were
imposed in a single directional axis (32, 33). Given the technological constraints at the time, these latter
works documented the sensitivity of the magnetotaxis response (i.e. MTB will exhibit magnetotaxis
behavior with the addition of an imposed B-Field vector), but not the minimal vector intensity in which
the magnetotaxis behavior is elicited. It is therefore suggested that until the study of these organisms
can be done under conditions better mimicking that natural environment, a complete understanding of
the ecophysiology of this group will not be possible. As more genetic information arises, it will be
necessary to have a system for testing alternative hypotheses under “natural conditions”. To this end,
the implementation of digital holographic microscopy is used here to demonstrate the study of
environmental magnetotatic bacteria at low magnetic field strengths (≤ 50 µT).
The DHM optics system and magnetic coil design utilized in this research provides real-time observation,
tracking, and quantification of individual MTB as they respond to low magnetic field intensities. The
technique described here provides the ability to explore how the dynamics of the magnetotaxis
response change under environmentally relevant magnetic field regimes and how this behavior differs
among various MTB within a heterogenous environmental sample. Measuring and characterizing the
116

detection sensitivities and response times of the magnetotaxis response under environmental
conditions (and in situ) is an important first step towards understanding the fundamental biology and
physics of magnetotactic bacteria, as well as the potential function(s) of magnetotaxis. The artificial
selection pressures manifested in pure cultured MTB research may underscore the natural abilities,
diversity, and function of magnetotaxis in the environment. The work reported here attempts to remove
these selection pressures and biases of the laboratory by utilizing recently collected environmental
samples and describing the elicited behavior of MTB under magnetic field strengths comparable to that
experienced by the organism in the natural environment.
Materials and methods
The experimental design (Figure 4.4A), in brief, consisted of placing an off-axis DHM (constructed out of
non-ferric components) within a tri-axial magnetic field coil system based on the design of Alldred and
Scollar (34). Lacustrine environmental samples (Baldwin Lake, Arcadia, CA) consisting of both MTB and
non-magnetotactic bacteria were exposed to magnetic fields of 50, 40, 30, 20 and 10 µT during data
acquisition. Data was also acquired under a null-field condition in which the magnetic coil system was
used to reduce the ambient magnetic field to less than 0.1 µT. Data was processed, tracked, and
quantified through the use of Manual Tracking plugin of the open-source image processing software, FIJI
(35), and MATLAB (R2019a). A set of abiotic control experiments involved micron-sized magnetite
particles that were subjected to the same experimental conditions. Videos of processed holograms for
each of these experimental runs are shown in Videos 4.1-4.6.
117

Figure 4.4: Experimental system. A) Tri-axial magnetic coil system. B) Digital holographic microscope optical
pathway diagram and instrumentation (20). C) Hologram data reconstruction and TEM images of positively
identified magnetotactic bacteria used in this study. Black scale bar is 1 µm.

Collection and preparation of environmental sample
A slurry of lake sediment was collected from Baldwin Lake, a small body of freshwater located within the
Los Angeles Arboretum (Arcadia, CA, USA). The slurry was collected in a 1:1 ratio of lake sediment and
overlying waters into 2-liter bottles and were stored at room temperature, unsealed to allow gas
exchange, except during transport.
MTB were concentrated from these samples by vigorously shaking the sediment-water mix and placing
magnetic stir bars on either side of the bottle, one centimeter above the sediment-water interface. The
magnets were placed in line with the true magnetic orientation, with the southern pole of one magnet
on the northern facing edge of the glass, and the northern pole of the other magnet on the south facing
118

edge of the glass. After 1 hour, a visible mass of cells had accumulated on the glass near the magnets
and was removed via Pasteur pipette and added to 1mL of filter sterilized lake water. Further
enrichment was achieved by conducting an additional round of magnetic isolation. The presence of MTB
in these samples were confirmed via light microscopy and the hanging-drop technique (21).
No additional isolation or purification of the MTB from the concentrated samples was attempted, as the
experimental design required a variety of diverse microbes (both magnetotactic and non-magnetotatic
in nature). These samples were then diluted with additional filter sterilized lake water to an appropriate
density for digital holography.
The enriched samples were injected into the sample well of manufactured glass sample chambers (ALine
Microfluidics, Rancho Dominguez, California), while sterile lake water was used to match the refractive
index in the reference port. The chambers were then sealed with vacuum grease to limit drift and placed
on the microscope stage.
To ensure that any observed taxis in response to the magnetic field was not simply the migration of
magnetic particulate in the environmental samples, an abiotic control consisting of sterile micron size
magnetite minerals was created by suspending the minerals in the filtered lake water.
Digital Holographic Instrumentation
A DHM instrument was constructed similar in design to those previously used (20) (Figure 4.4B). The
instrument is a benchtop version of the orangeBox field instrument designed for laboratory
investigations. In brief, a coherent light source (405nm) is collimated and passed through two identical
but separate microfluidic wells. One well contained the sample of interest, while the other contained
sterile reference liquid in order to match optical path lengths. After interacting with both the sample
and reference liquids, the two beams are passed through separate but identical objective lenses (NA
=0.3). A relay lens is used to form an image of the two beams on a digital CCD camera. For these
119

experiments, all iron containing pieces were replaced with plastics or aluminum, to avoid any magnetic
interference. A magnetically shielded ethernet cable was incorporated to avoid any electrical
interference caused through data transfer. Data acquisition was accomplished using an open-source
software suite DHMx (https://github.com/dhm-org/dhm_suite).
Magnetic coil system
The coil system consisted of three different coils aligned orthogonally, with each of the three based on
the design of Alldred and Scollar (1967) (15). The two outer wraps of each Alldred–Scollar coil measured
2.1 m on a side and the two inner wraps measured 2.2 m on a side. Two of the coils generated
horizontal fields (one in the north-south direction, the other in the east-west direction) while the third
coil generated a vertical field. Electrical current to each coil was provided by a programmable power
supply controlled by a computer (22). In all experiments, the vertical coil was used to generate a field
equal in intensity but opposite in direction to the vertical component of the local magnetic field, so that
the vertical field component was eliminated. This was done so that the bacteria would move
horizontally within the focal plane of the microscope. Horizontal fields ranging in intensity from 0 µT
(null field) to 50 µT were used in tests (see below). Field measurements were confirmed with a tri-axial
magnetometer (Applied Physics, Inc., model 520A). In one additional field condition, a 50 µT field was
rotated 1080° (3 full rotations) in 60 seconds. This was achieved in 600 incremental steps; each step was
1.8°, so that the rate of directional change was 18° per second (Video 4.7).
Experimental design parameters
Both samples containing MTB or magnetite particles were first exposed to a null field (NF) in which the
Earth’s ambient magnetic field was not detectable. Data was acquired for 1 min at a rate of 15 frames
per second under the NF condition. For the field intensity experiments, holograms were acquired for 3-4
seconds under NF conditions prior to the establishment of a magnetic field of known intensity applied
120

solely in the xy-plane. Microbial responses were explored at magnetic field intensities of 50, 40, 30, 20,
and 10 µT. Between experiments, the magnetic field was reversed 180° at the same field strength for an
equal period of time. This was done to allow the magnetotatic bacteria to swim back towards the center
of the sample chamber and into the field of view of the instrument.
Tracking and data analysis
Tracking of all samples was performed through the use of the Manual Tracking plug-in in the open
source image analysis tool FIJI (16). Cell morphology was confirmed via holographic reconstruction
through the DHM Reconstruction plug-in in FIJI (23). Data analysis was conducted through MATLAB
(R2019a).
16S rRNA diversity
Samples were taken at various enrichment steps for taxonomic identification could be possible of non-
coccus MTB. Cells were collected near the poles of two bar magnets in the enrichment protocol
discussed above. Samples for 16S rRNA identification were collected from the initial cell pellet (N1,S1),
after an initial dilution into sterile media and secondary magnetic enrichment step (N2,S2) and for the
northward polarity sample, after 1 hour on the modified racetrack method (N3) described in (21).
Samples were sequenced using the Illumina Sequencing pipeline and annotated through MR DNA
(Shallowater, Texas).
Video construction
Image sequence videos were constructed through FIJI. These files were color corrected and annotated in
Flame v2020.1.1 (AutoDesk, San Rafael, California).
Results
121

All cells displaying non-random movement (i.e. non-Brownian movement) within the focal range were
tracked and analyzed to determine directionality and velocity for each B-field intensity. Collectively,
these microbes were designated as the “bulk community”. The overall trend in the turning angle of each
tracked cell was then categorized as either displaying positive northward polarity (north trending) or no
notable magnetotactic bias in direction of travel (no south trending cells were quantified, although the
presence of such MTB were observed in the rotating magnetic field experiment. Figure 4.5 illustrates
the tracks of positively identified MTB over time at 50 µT (0.5 G) and 10 µT (0.1 G).

Figure 4.5: Tracks of positively identified MTB exhibiting Northern polarity. A) At 50 µT and B) 10 µT over time.
The ‘N’ located in the bottom left corners of the plots denotes the direction of the imposed magnetic North.

122

The cells that displayed a magnetotactic response were subsampled from the bulk community
for further analysis. Micron-sized particles of magnetite were used as an abiotic control to ensure that
any observed northern polarity in the environmental samples was not simply the migration of magnetic
minerals through the fluid. Due to the length scale of Brownian motion observed in the abiotic
magnetite particle experiments, the apparent motion is subject to aliasing towards the cardinal
directions resulting in the quad-modal distribution. Figure 4.6 displays the directionality of the bulk
community, northward polarity group, and abiotic controls under each magnetic field intensity. Of note,
the observable trend that at lower magnetic field strengths, the northern polarity group has a broader
direction of travel (i.e. less fidelity in following the field lines directly in the northern direction).
Calculated mean speeds for each group are displayed in Table 4.1.

Table 4.1 Mean speeds of tracked cells and magnetite particles at different magnetic field intensities.

123

Figure 4.6 Polar histograms of the directional trajectories of tracked cells and magnetite particles under each
magnetic field regime. Bulk Community reflects the relative frequency of directional trajectory in all tracked cells,
Northern Polarity histograms contain the subset of tracked cells that were determined to exhibit a positive
magnetotaxis response in the northern direction.

Evidence for magnetotatic responses among non-coccoid cells was not anticipated, as previous work at
this site revealed only coccoid magnetotatic bacteria (1). Nevertheless, non-coccoid cells with northern
polarity were observed in reconstructed holograms and later confirmed through TEM (Figure 4.4C). 16S
rRNA diversity was determined through Illumina sequencing; results are shown in Figure 4.7 and
illustrate the diversity of bacterial sequences throughout the magnetic enrichment process at the order
and genus taxonomic levels. The identity of the non-coccoid MTB could not be determined.
Magnetosome alignment and cell morphology are consistent with the possibility that it is a member of
the Magnetovibrio group; however, sequence representation of the genera were not observed in any of
124

the samples. Although visually abundant in the environmental samples, traditional familial-level clades
of non-coccoid MTB (the nitrospiraceae, rhodospirllaceae and desulfovibrionaceae) were detected at
less than 0.4% of sequences in the bulk sediment, and even lower representation in the magnetically
enriched samples (<0.2%). The magnetococcaceae represent over 89% of sequences at the family level
taxa (Figure 4.7 C). Taxonomic classification of MTB sequences within communities are often dominated
by a single taxon, with species diversity correlated in part with the local strength of the geomagnetic
field (36). However, the visual observation of a diverse community of MTB in the Baldwin Lake
ecosystem, and the lack of sequences from traditional non-coccoid MTB clades in the 16S analysis
highlights the possibility that the MTB are a more diverse group of organisms that currently believed.
Tracked microbes in the positive taxa group were further characterized by morphology allowing for a
more nuanced description of the cells’ magnetotaxis responses. Upon clustering the tracked cells by
morphology (coccus and non-coccus), a few trends were distinguishable: 1) a correlation between rate
of swimming and field strength is suggested, but is not statistically significant, due to the limited number
of cells, 2) the cocci demonstrate a much higher variability in swimming speeds; whereas, the non-
coccus microbes have a much tighter range of swimming rates, and 3) the rate and variability of the
coccus cells appear to have a correlation with magnetic field intensity, while the non-coccoid cells have
a much more uniform response (Figure 4.8).

125

Family

Figure 4.7 Taxonomic classification of magnetically enriched samples via Illumina sequencing. 16S rRNA diversity
of bacteria that demonstrate Southern (S1-2) and Northern (N1-3) polarity at the order and genus level. Samples
x1 (S1, N1) represent taxonomic diversity of cells during the first magnetic enrichment step. Samples x2 and x3
reflect sequence diversity after each additional round of magnetic enrichments. Familial level taxonomic
composition (greater than 1.5% of sequence abundance) of the Northern polarity samples.

126

Figure 4.8 Mean speed vs B-Field intensity plot of bacteria (coccoid and non-coccoid) and abiotic magnetite
particles. Error bars reflect 95% confident interval. Determined r
2
values: coccoid (r
2
= 0.68), non-coccoid (r
2
=
0.90), magnetite (r
2
= 0.52).

As a qualitative proof-of-concept experimental design, a 50 µT field was rotated 1080° uniformly at
three revolutions per minute (Video 4.7). MTB were seen to follow the rotating field with varying fidelity
and, importantly, polarity allowing for the investigation into the dynamics of magnetotaxis. It is of note
that the northern polarity MTB (a non-coccoid MTB, Figure 4.9A, Video 4.8) followed the Northern
magnetic direction with a much more concise accuracy than the southern polarity MTB (a coccoid,
Figure 4.09B, Video 4.9) during field rotation, implying differences in the cells’ abilities to orientate and
respond to magnetic field changes and potentially direction.
127

Figure 4.9 Composite image of A) Northern polarity MTB and B) Southern polarity MTB during 1080° rotation of
a 50 µT magnetic field. Magnetic North is located in the bottom left of each hologram, T
start
and T
final
designate
starting and ending location of MTB. Inserts show morphology of cells in magnified holograms.
See Video 4.8 and 4.9.

Discussion
This work provides direct observation and quantification of the magnetotaxis response of single cells of
an environmental enrichment sample under earthly (or below) levels of magnetic field strengths. As the
DHM instrument was constructed in a way that prevents any localized magnetic field disturbances
around the bacteria, this technique provides the first view of MTB and the magnetotaxis response under
environmentally relevant conditions with both sub-micron spatial and high temporal resolution. The
experimental design utilized in this work allowed us to record, characterize and elucidate differences in
how MTB respond to the same stimuli, and how the MTB responses change as a function of field
strength. This is the first time that in situ identification, tracking and quantification of the magnetotaxis
response of individual MTB from within a mixed community has been accomplished below the Earth’s
nominal field strength (50 µT) with confirmed positive responses down to 10 µT.
128

Genomic analysis of the pure culture model organisms such as Magnetospirillum magneticum AMB-1,
have provided the basis of our understanding of magnetosome production, yet as additional MTB are
brought into pure culture and the laboratory setting, the assumptions of the past have had to be
reconsidered. Underscoring this fact, a recent study focused on identifying iron containing genes in
metagenomic assemblies found that the putative genes associated with magnetite biomineralization in
magnetosome production (mam) were only present in one of 27 environmental metagenomes analyzed,
although MTB are considered ubiquitous in aquatic systems, perhaps alluding to a larger undescribed
diversity of B-Field sensitive organisms (37) that is ubiquitous, but does not necessarily mean abundant.
These tactic microbes are likely easily visualized, but rather rare microbes, even in redox-interface
environments may be easily overlooked. By exploring the lower bounds of environmental MTBs’
magnetic field perceptions, sensitivities, and responses we can begin to understand and explain what
aspects of magnetotaxis appear universal among MTB and how diverse the response is among the MTB
clades. Using this approach, it may be possible to reassess the “why” of magnetotaxis, which has so far
been an area of some disagreement and confusion. Cryptic lines of evidence include: the presence of
MTB communities in equatorial locales (at which the geomagnetic field inclination is 0° and provides no
vertical directional cues), MTB that exhibit southern polarity in the northern hemisphere, and role of
magnetosomes in eliminating reactive oxygen species among others (17, 38, 39).
Examining behavior in situ, and under natural levels of field strength may provide a pathway for further
understanding, and a sense of whether the MTB may have a number of uses for this ability. For example,
discrepancies in the temporal scale of the sensing and response mechanisms related to magnetotaxis
between morphotypes is suggested, but not verified by the behaviors described here. Some cells appear
capable of instantaneously orienting and following the magnetic field lines continuously, while others
appear to take a magnetic bearing periodically and adjust their trajectories in accordance to the new
bearing. Our technique also provides a means of elucidating the impacts of regional magnetic field
129

intensities on the MTB community, as the magnitude of the local B-Field has been suggested to affect
both the magnetotaxis behavior and community composition of MTB-rich environments (36).
The observations and analyses conducted in this report provide evidence that recently harvested MTBs
are capable of responding to magnetic fields down to approximately 20% of the Earth’s nominal field
strength (and perhaps lower). The response (i.e. swimming rates) of the MTB may be correlated weakly
with field intensity, but this relationship may also vary among different bacterial morphotypes. As
previously identified, the magnetococcus ARB-1 group represents a diverse set of cocci that vary both in
size and magnetosome arrangement (1), which may explain the large range of swimming rates at higher
magnetic fields. As the B-field intensity is lowered, the swimming rates of the coccus cluster. One
possible interpretation is that differences in magnetosome organization impact the ability of ARB-1 cells
to sense lower field strengths. The high rates of swimming (often 70-100 body lengths/ second) typically
reported for magnetococus cells is notably diminished at lower magnetic field intensities, suggesting
that the observed behaviors may be an artifact of previous experimental designs, and not representative
of the natural behavior of the organisms in the environment.
The previously unknown, non-coccoid MTB show a slower rate of swimming under all magnetic field
regimes but are still capable of sensing field strengths down to 10 µT. The differences observed between
these two morphotypes may have underpinnings in unique chemical pathways in how the cells sense B-
field orientation, and how this information is translated to the flagella. As these magnetotactic non-
coccoid bacteria were not previously reported in the Baldwin Lake sediment ecosystem, this
experimental design allowed us to visually identify and later confirm the existence of additional MTB
through TEM. The discrepancy between the lack of representation of traditional MTB clades within the
community analysis and the presence of the non-coccoid bacteria suggests a larger diversity of bacteria
capable of sensing and responding to magnetic fields may exist in nature. Through the direct
observation of heterogeneous environmental samples, it is evident that MTB that are not identified
130

through sequencing are present in the environment and that magnetotaxis may be a more widespread
ability than previously thought. Additionally, evidence of MTB exhibiting southern polarity in the 1080°
rotating B-field, once again underscore the scarce insight we currently have into the ecophysiology of
MTB in natural conditions.
Using a combination of environmental samples, digital holography, and a magnetic coil system, this
work provides a promising and novel approach of exploring the dynamics of magnetotaxis of individual
cells under environmentally relevant conditions. An assortment of MTB were capable of sensing and
orienting to magnetic fields down to 20% of the Earth’s nominal field strength. In addition to this direct
observation, tracking of individual cells allowed for the quantification and analysis of cell velocities,
turning angles and directionality while providing visual confirmation of cell morphology which
elucidated differences in how the MTBs respond to the same stimuli. This work provides the foundation
of a new direction and methodology with which to study MTB and the dynamics of the magnetotaxis
response at environmentally relevant magnetic field strengths.

References:
1. B. Lea Cox, R. Popa, D. A. Bazylinski, B. Lanoil, S. Douglas, A. Belz, D. L. Engler, K. H. Nealson, Organization
and elemental analysis of P-, S-, and Fe-rich inclusions in a population of freshwater magnetococci.
Geomicrobiol. J. 19, 387–406 (2002).
2. A. Komeili, Molecular mechanisms of compartmentalization and biomineralization in magnetotactic
bacteria. FEMS Microbiol. Rev. 36, 232–255 (2012).
3. R. B. Frankel, D. A. Bazylinski, Magnetosomes and magneto-aerotaxis. Contrib. Microbiol. 16, 182–193
(2009).
4. S. Bellini, On a unique behavior of freshwater bacteria. Chinese J. Oceanol. Limnol. 27, 3–5 (2009).
5. C. Jogler, M. Niebler, W. Lin, M. Kube, G. Wanner, S. Kolinko, P. Stief, A. J. Beck, D. de Beer, N. Petersen, Y.
Pan, R. Amann, R. Reinhardt, D. Schüler, Cultivation-independent characterization of “Candidatus
Magnetobacterium bavaricum” via ultrastructural, geochemical, ecological and metagenomic methods.
Environ. Microbiol. 12, 2466–2478 (2010).
131

6. R. Uebe, D. Schüler, Magnetosome biogenesis in magnetotactic bacteria. Nat. Rev. Microbiol. 14, 621–637
(2016).
7. C. T. Lefèvre, L. F. Wu, Evolution of the bacterial organelle responsible for magnetotaxis. Trends Microbiol.
21, 534–543 (2013).
8. S. Barber-Zucker, N. Keren-Khadmy, R. Zarivach, From invagination to navigation: The story of
magnetosome-associated proteins in magnetotactic bacteria. Protein Sci. 25, 338–351 (2016).
9. D. L. Balkwill, D. Maratea, R. P. Blakemore, Ultrastructure of a magnetotactic spirillum. J. Bacteriol. 141,
1399–1408 (1980).
10. R. E. Kopp, J. L. Kirschvink, The identification and biogeochemical interpretation of fossil magnetotactic
bacteria. Earth-Science Rev. 86, 42–61 (2008).
11. C. T. Lefevre, D. A. Bazylinski, Ecology, Diversity, and Evolution of Magnetotactic Bacteria. Microbiol. Mol.
Biol. Rev. 77, 497–526 (2013).
12. R. P. Blakemore, Magnetotactic Bacteria. Annu. Rev. Microbiol. 36, 217–238 (1982).
13. E. Rosenberg, F. DeLong, E.F., Stackebrandt, E., Lory, S., Thompson, The Prokaryotes - Prokaryotic Biology
and Symbiotic | Eugene Rosenberg | Springer (2013; http://www.springer.com/us/book/9783642301933).
14. H. C. McCausland, A. Komeili, Magnetic genes: Studying the genetics of biomineralization in magnetotactic
bacteria. PLoS Genet. 16, e1008499 (2020).
15. T. Islam, C. Peng, I. Ali, Morphological and cellular diversity of magnetotactic bacteria: A review. J. Basic
Microbiol. 58, 378–389 (2018).
16. A. Komeili, H. Vali, T. J. Beveridge, D. K. Newman, Magnetosome vesicles are present before magnetite
formation, and MamA is required for their activation. Proc. Natl. Acad. Sci. U. S. A. 101, 3839–3844 (2004).
17. R. B. Frankel, R. P. Blakemore, F. F. T. De Araujo, D. M. S. Esquivel, J. Danon, Magnetotactic Bacteria at the
Geomagnetic Equator. Science (80-. ). 212, 1269–1270 (1981).
18. W. Lin, G. A. Paterson, Q. Zhu, Y. Wang, E. Kopylova, Y. Li, R. Knight, D. A. Bazylinski, R. Zhu, J. L. Kirschvink,
Y. Pan, Origin of microbial biomineralization and magnetotaxis during the Archean. Proc. Natl. Acad. Sci. U.
S. A. 114, 2171–2176 (2017).
19. C. T. Lefèvre, D. Trubitsyn, F. Abreu, S. Kolinko, L. G. P. de Almeida, A. T. R. de Vasconcelos, U. Lins, D.
Schüler, N. Ginet, D. Pignol, D. A. Bazylinski, Monophyletic origin of magnetotaxis and the first
magnetosomes. Environ. Microbiol. 15, 2267–2274 (2013).
20. W. Lin, W. Zhang, X. Zhao, A. P. Roberts, G. A. Paterson, D. A. Bazylinski, Y. Pan, Genomic expansion of
magnetotactic bacteria reveals an early common origin of magnetotaxis with lineage-specific evolution.
ISME J. 12, 1508–1519 (2018).
21. A. J. Biggin, M. J. de Wit, C. G. Langereis, T. E. Zegers, S. Voûte, M. J. Dekkers, K. Drost, Palaeomagnetism of
Archaean rocks of the Onverwacht Group, Barberton Greenstone Belt (southern Africa): Evidence for a
stable and potentially reversing geomagnetic field at ca. 3.5Ga. Earth Planet. Sci. Lett. 302, 314–328 (2011).
132

22. F. U. Battistuzzi, A. Feijao, S. B. Hedges, A genomic timescale of prokaryote evolution: Insights into the
origin of methanogenesis, phototrophy, and the colonization of land. BMC Evol. Biol. 4, 1–14 (2004).
23. Z. Chen, J. J. Yin, Y. T. Zhou, Y. Zhang, L. Song, M. Song, S. Hu, N. Gu, Dual enzyme-like activities of iron
oxide nanoparticles and their implication for diminishing cytotoxicity. ACS Nano. 6, 4001–4012 (2012).
24. W. Lin, J. L. Kirschvink, G. A. Paterson, D. A. Bazylinski, Y. Pan, On the origin of microbial magnetoreception.
Natl. Sci. Rev. 7, 472–479 (2020).
25. H. Vali, J. L. Kirschvink, in Iron Biominerals, R. B. Frankel, R. P. Blakemore, Eds. (Springer, Boston, MA,
1991).
26. F. Abreu, M. E. Cantão, M. F. Nicolás, F. G. Barcellos, V. Morillo, L. G. P. Almeida, F. F. Do Nascimento, C. T.
Lefèvre, D. A. Bazylinski, A. T. R. De Vasconcelos, U. Lins, Common ancestry of iron oxide- and iron-sulfide-
based biomineralization in magnetotactic bacteria. ISME J. 5, 1634–1640 (2011).
27. N. Philippe, L. F. Wu, An MCP-Like Protein Interacts with the MamK Cytoskeleton and Is Involved in
Magnetotaxis in Magnetospirillum magneticum AMB-1. J. Mol. Biol. 400, 309–322 (2010).
28. C. T. Lefèvre, M. Bennet, L. Landau, P. Vach, D. Pignol, D. A. Bazylinski, R. B. Frankel, S. Klumpp, D. Faivre,
Diversity of magneto-aerotactic behaviors and oxygen sensing mechanisms in cultured magnetotactic
bacteria. Biophys. J. 107, 527–538 (2014).
29. C. Chen, Q. Ma, W. Jiang, T. Song, Phototaxis in the magnetotactic bacterium Magnetospirillum
magneticum strain AMB-1 is independent of magnetic fields. Appl. Microbiol. Biotechnol. 90, 269–275
(2011).
30. M. Amor, M. Tharaud, A. Gélabert, A. Komeili, Single-cell determination of iron content in magnetotactic
bacteria: implications for the iron biogeochemical cycle. Environ. Microbiol. 22, 823–831 (2020).
31. R. P. Blakemore, Magnetotactic Bacteria. Science (80-. ). 190, 377–379 (1975).
32. R. P. Blakemore, R. B. Frankel, A. J. Kalmijn, South-seeking magnetotactic bacteria in the southern
hemisphere. Nature. 286, 384–385 (1980).
33. R. B. Frankel, R. P. Blakemore, Navigational Compass in Magnetic Bacteria. J. Magn. Magn. Miner. 15,
1562–1564 (1980).
34. J. C. Alldred, I. Scollar, Square cross section coils for the production of uniform magnetic fields. J. Sci.
Instrum. 44, 755–760 (1967).
35. J. Schindelin, I. Arganda-Carreras, E. Frise, V. Kaynig, M. Longair, T. Pietzsch, S. Preibisch, C. Rueden, S.
Saalfeld, B. Schmid, J. Y. Tinevez, D. J. White, V. Hartenstein, K. Eliceiri, P. Tomancak, A. Cardona, Fiji: An
open-source platform for biological-image analysis. Nat. Methods. 9, 676–682 (2012).
36. W. Lin, Y. Wang, Y. Gorby, K. Nealson, Y. Pan, Integrating niche-based process and spatial process in
biogeography of magnetotactic bacteria. Sci. Rep. 3, 1–9 (2013).
37. A. I. Garber, K. H. Nealson, A. Okamoto, S. M. McAllister, C. S. Chan, R. A. Barco, N. Merino, FeGenie: A
Comprehensive Tool for the Identification of Iron Genes and Iron Gene Neighborhoods in Genome and
133

Metagenome Assemblies. Front. Microbiol. 11, 1–23 (2020).
38. S. L. Simmons, D. A. Bazylinski, K. J. Edwards, South-seeking magnetotactic bacteria in the Northern
Hemisphere. Science (80-. ). 311, 371–374 (2006).
39. F. F. Guo, W. Yang, W. Jiang, S. Geng, T. Peng, J. L. Li, Magnetosomes eliminate intracellular reactive oxygen
species in Magnetospirillum gryphiswaldense MSR-1. Environ. Microbiol. 14, 1722–1729 (2012).
Closing remarks on MTB work:
The experiments described in the previous section provide clear evidence of the non-uniform
magnetotaxis response of environmental MTB under realistic magnetic field intensities. Although the
limits of stimuli detection was not ascertained in the first round of experiments, this work provides a
proof of concept in the technological application of DHM and a tri-axial magnetic coil system in the
study of ‘in situ’ MTB dynamics. Not reported above due to the constraints of the report, the inability of
Magnetospirillum AMB-1 wildtype cells to respond to field strengths as high as 0.5 Gauss dramatically
underscores our rationale in pushing environmental investigations of MTB. Another observation that
was not reported in the previous manuscript was the observation of MTB that exhibited a southern
polarity magnetotaxis response. At least two Baldwin Lake MTB were observed in the 1080° rotational
0.5G magnetic field experiments that swam following the southern pole of the B-Field (see Video 4.9).
The northern polarity bugs were noted completing three full rotations in the clockwise direction, while
the southern polarity MTB completed the rotations in a counterclockwise fashion.
Kirschvink (1982) suggested that the minimal magnetic field intensity required for magnetotaxis to be an
advantageous adaptation would be~ 6 µT (0.06 G). This number was approximated using a ‘prototypical’
cell size and the approximate size of magnetite magnetosomes known at that time. As the number and
diversity of MTB has expanded in the past 40 years, this theoretical number (based on multiple gross
assumptions) is still a valuable target range to explore. Our experiments did not definitively show
evidence of positive magnetotaxis responses from environmental samples at a 0.05 G B-Field intensity.
That said, as larger more complex MTB are now known to exist, determining the limits of magnetic field
134

perception among the different MTB clades and how the physical aspects of MTB (cell size, number of
magnetosome chains, magnetosome morphology, total surface area of magnetosome, etc.) affects the
sensitivity of B-Field orientation is a ripe field for future exploration.
The presence of MTB is the ultimate example of Geobiology in my mind. The coupling of the behavior of
microscale organisms with the geomagnetic field is unparalleled. Of all the research directions I have
pursued in six years of graduate studies, the exploration of MTB has been the most rewarding. The stark
difference between microorganisms in the natural setting versus those that have been cultured speaks
again to my personal thoughts on microbiological field studies and the importance of experimental
setting on the observations of life.
The meeting of real-time, sub-micron resolution volumetric imaging, tri-axial magnetic field control and
the study of magnetotactic bacteria provides an unprecedented ability. This system facilitates the
possibility to both observe bacterial motility and manipulate the motile responses of the cells in three-
dimensional space. As an external magnetic field may be the simplest, least invasive method to
influence bacterial behavior, future work that focuses on biosensors or the application of artificially
constructed organisms can be vetted and tested using MTB and the DHM/coil system.
Proposed future work:
Some of the future work I envision to sprout from these findings involve the development of a magnetic
coil/combined DHM field instrument for the in-situ isolation and experimentation of MTB from diverse
settings. Given the diversity of experiments that a DHM instrument is capable of being utilized in, I want
to highlight some of the areas of research that a DHM-magnetic coil field instrument could provide new
insight:
1) In situ identification and isolation of MTB in the field
2) Determining minimal B-Field required for a response at various geographical locations
135

3) Mutant and model organism studies
4) Magnetotaxis dynamics during magnetosome formation

In situ identification and isolation of MTB in the field:
One of the most obvious directions of inquiry that would be facilitated by a DHM/magnetic coil field
instrument would be to screen, identify and isolate MTB in the field. Similar to more traditional methods
of MTB isolation, the novel isolation technique I envision would utilize a specific sample chamber
designed for the concentration and collection of MTB (Figure 4.10). As the tri-axial magnetic coil allows
us to establish and impose complex magnetic field regimes in three dimensions, controlling an
established B-Field to isolate MTB is fairly straightforward. Using an alternating magnetic field
orientation (rotating the B-field 90° back and forth) to concentrate and isolate MTB, this technique and
chamber design would allow for both follow up experiments with the enriched MTB and also allow for a
transect of bacterial sequences to be obtained throughout the experiment.

Figure 4.10: Sample chamber design for use in MTB-DHM enrichment experiments. Using an alternating B field
orientation (90° rotations) controlled by a magnetic coil system, the enrichment of both northern and southern
polarity MTB could be done in situ.

136

The field instrument design would allow for rapid screening and confirmation of MTB from
environmental samples, allowing us to test the hypothesis that MTB are ubiquitous in oxic/anoxic zones.
As the magnetic field coil system software would allow for a predetermined scripted run of magnetic
field regimes, the isolation of MTB from environmental samples could be completed simply by loading
the sample chamber and initiating the sequence protocol. Presence and concentration of MTB could be
completed in less than 20 minutes with a higher fidelity than traditional isolation techniques. Through
this isolation technique, even low abundance MTB would be capable of isolation as the only cells located
within the final sample chamber well would be the bacteria that successfully navigate the magnetic field
‘maze’ that the sample chamber provides. Utilizing the same sample chamber design but differing the
magnetic field intensity used during magnetic enrichment may illuminate differences in the
magnetotaxis response of MTB clades within an ecosystem. Using the design envisioned here, low-
abundance MTB that may not be identified through sequencing would be easily isolated and
concentrated for additional study of culture attempts.
Determining minimal B-Field required for a response at various geographical locations:
As the ambient magnetic field varies with latitude (from 0.24 µT near the equator to 0.64 µT at the
poles) and MTB are found across the globe, understanding how the external B-Field impacts MTB
biology is a fundamental question in understanding the magnetotaxis response (Finlay, 2010). The work
above demonstrates that among the different MTB clades, B-Field intensity elicits different responses.
The sensitivity, directionality and motile responses of the different MTB morphotypes observed from a
single locale was apparent and statistically differentiable. Comparing similar MTB organisms (such as
magnetospirillum) from various latitudes and determining the correlation between ambient field
strength and the minimal B-Field necessary to trigger a magnetotaxis response would allow further
insight into how magnetic orientation is translated into motility. By comparing the genomes of these
geographically distant, closely related species and understanding how their magnetotaxis sensitivities
137

differ, inference may be gained through explorations of up-regulated genes or magnetosome
morphologies. Science is relatively strong in the comparison of two microorganisms from a genomic,
biochemical and transcriptional basis (among other techniques), using a DHM we can then see the
physical manifestation of these changes on the bacterium’s motile response. This work could benefit the
much anticipated and developing biosensor field immensely.
Another question this work may provide insight into is how magneto-sensitive microorganisms respond
during times of geomagnetic polarity reversals. MTB make up an important sedimentary component in
the fossil record, as magnetosome minerals are more easily preserved than the biological tissues of
bacteria. As the preservation, identification and observation of bacterial fossils in the rock record is an
extremely difficult task at a minimum, the non-proteinaceous component of the magnetosome provides
a valuable marker. Understanding any adaptations of MTB at low latitude and subjecting high latitude
microbes to low magnetic fields may provide some answers into the evolution of MTB during
geomagnetic field fluctuations.
Mutant and model organism studies:
Although our initial work did not observe obvious magnetotaxis behavior of AMB-1 (wt) cells under a
0.5G B-Field intensity, stronger magnetic fields could be tested using cultured MTB. Given the number of
model organism MTB mutants that currently exist (like those used in the Komeili group’s publications),
determining which genes impact magnetic sensing and response could be accomplished using the
proposed instrumentation. Screening an array of mutants under the same low magnetic field intensities
may allow us to determine and quantify any differences in the magnetotaxis response between mutants
of interest and the wildtype organism.
Comparing the magnitude and mechanisms underlying observational changes in the magnetotaxis
dynamics of model organism MTB may provide direct evidence of how the magnetic field orientation is
138

translated to flagellar motility. As mutants have been used to successfully illuminate the genetic basis of
magnetosome formation, our understanding of the signal transduction system between magnetic field
orientation and flagellar response has been limited. Quantifying the dynamics of the magnetotaxis
response of mutants compared to the wild type organism may provide valuable insight into the signaling
processes of MTB. Using both knock out and upregulated gene mutants, the genetic underpinnings and
chemical response of MTB to environmentally relevant B-Fields may be unveiled.
Additional insight may be gained through transcriptomics of model MTB when exposed to changing
magnetic field strengths. The transcriptional response of MTB under a null field vs an imposed field
condition may highlight how MTB regulate magnetotaxis. As evident from the previous work, the
magnetotaxis response can differ with magnetic field intensity. The behavior of MTB under relevant
magnetic fields may show that magnetotaxis is a variable motile response, useful when needed, but not
overwhelming the cell’s swimming trajectories as seen in bar magnet studies. The use of strong magnets
in magnetotaxis experiments may artificially inflate how important the response is in the organisms in
the environment.
Magnetotaxis dynamics during magnetosome formation:
Using the variable control of the tri-axial magnetic coil system would facilitate experiments in which the
growth of MTB and the onset of magnetotaxis responses can be quantified. Controlling the B-Field
intensity during critical growth stages of the magnetosome organelle would allow us to determine what
variables (number of magnetosomes, average size of magnetosome minerals, etc.) impact the onset of
magnetotaxis. Another potentially insightful experiment suite would be comparing magnetotaxis
dynamics and magnetosome biomineralization of a pure culture organism when grown in static and
rotational 0.1G and 0.5G B-Fields. Using model organisms to determine the impacts of the external B-
139

Field on magnetosome biomineralization and mineralogy could provide future advancements in the
fields of material sciences, biosensor production, applied microbiology and general MTB biology.

140

Bibliography
Abreu, Fernanda et al. 2011. “Common Ancestry of Iron Oxide- and Iron-Sulfide-Based Biomineralization
in Magnetotactic Bacteria.” ISME Journal 5(10): 1634–40.
Alldred, J. C., and I. Scollar. 1967. “Square Cross Section Coils for the Production of Uniform Magnetic
Fields.” Journal of Scientific Instruments 44(9): 755–60.
Amor, Matthieu, Mickaël Tharaud, Alexandre Gélabert, and Arash Komeili. 2020. “Single-Cell
Determination of Iron Content in Magnetotactic Bacteria: Implications for the Iron Biogeochemical
Cycle.” Environmental Microbiology 22(3): 823–31.
Bailey, E.H., Jr. Blake, M.C., and D.L. Jones. 1970. “On-Land Mesozoic Oceanic Crust in California Coast
Ranges.” U.S. Geological Survey Professional Paper 700: C70–C81.
Balkwill, D. L., D. Maratea, and R. P. Blakemore. 1980. “Ultrastructure of a Magnetotactic Spirillum.”
Journal of Bacteriology 141(3): 1399–1408.
Barnes, Ivan, LaMarche, V.C., Himmelberg, G. 1967. “Geochemical Evidence of Present-Day
Serpentinization.” Science (New York, N.Y) 156(3776): 830–32.
Battistuzzi, Fabia U., Andreia Feijao, and S. Blair Hedges. 2004. “A Genomic Timescale of Prokaryote
Evolution: Insights into the Origin of Methanogenesis, Phototrophy, and the Colonization of Land.”
BMC Evolutionary Biology 4: 1–14.
Bedrossian, Manuel et al. 2017. “Quantifying Microorganisms at Low Concentrations Using Digital
Holographic Microscopy (DHM).” Journal of Visualized Experiments 2017(129): 1–11.
Bedrossian, Manuel M. 2020. “A Novel Digital Holographic Microscope (DHM) to Investigate and
Characterize Microbial Motility in Extreme Aquatic Environments.” California Institute of
Technology.
Beese-Vasbender, Pascal F. et al. 2015. “Electrochemical Characterization of Direct Electron Uptake in
Electrical Microbially Influenced Corrosion of Iron by the Lithoautotrophic SRB Desulfopila
Corrodens Strain IS4.” Electrochimica Acta 167: 321–29.
https://linkinghub.elsevier.com/retrieve/pii/S001346861500818X.
Bellini, Salvatore. 2009. “On a Unique Behavior of Freshwater Bacteria.” Chinese Journal of Oceanology
and Limnology 27(1): 3–5. https://doi.org/10.1007/s00343-009-0003-5.
Biggin, Andrew J. et al. 2011. “Palaeomagnetism of Archaean Rocks of the Onverwacht Group, Barberton
Greenstone Belt (Southern Africa): Evidence for a Stable and Potentially Reversing Geomagnetic
Field at ca. 3.5Ga.” Earth and Planetary Science Letters 302(3–4): 314–28.
http://dx.doi.org/10.1016/j.epsl.2010.12.024.
Bird, Lina et al. “Serpentinomonas raichei and S. mccroryi, gen. nov., sp. nov., two hyperalkaliphilic,
hydrogen-oxidizing microaerobic bacteria isolated from terrestrial serpentinizing springs" (in Prep).
Blake, M Clark, Edgar H Bailey, and Carl M Wentworth. 2012. “The Cedars Ultramafic Mass, Sonoma
County, California.” USGS Open-File Report 2012–1164(April): 1–16.
Blakemore, R. P., R. B. Frankel, and Ad J. Kalmijn. 1980. “South-Seeking Magnetotactic Bacteria in the
Southern Hemisphere.” Nature 286(5771): 384–85.
141

Blakemore, R P. 1975. “Magnetotactic Bacteria.” Science 190(4212): 377–79.
https://www.jstor.org/stable/1741081.
Blakemore, R P. 1982. “Magnetotactic Bacteria.” Annual Review of Microbiology 36(1): 217–38.
Burdige, David J., Surya P. Dhakar, and Kenneth H. Nealson. 1992. “Effects of Manganese Oxide
Mineralogy on Microbial and Chemical Manganese Reduction.” Geomicrobiology Journal 10(1): 27–
48. http://www.tandfonline.com/doi/abs/10.1080/01490459209377902.
Callahan, Benjamin J. et al. 2016. “DADA2: High-Resolution Sample Inference from Illumina Amplicon
Data.” Nature Methods 13(7): 581–83.
Callahan, Benjamin J., Paul J. McMurdie, and Susan P. Holmes. 2017. “Exact Sequence Variants Should
Replace Operational Taxonomic Units in Marker-Gene Data Analysis.” ISME Journal 11(12): 2639–
43. http://dx.doi.org/10.1038/ismej.2017.119.
Carbajosa, Sofía et al. 2010. “Electrochemical Growth of Acidithiobacillus Ferrooxidans on a Graphite
Electrode for Obtaining a Biocathode for Direct Electrocatalytic Reduction of Oxygen.” Biosensors
and Bioelectronics 26(2): 877–80.
https://linkinghub.elsevier.com/retrieve/pii/S0956566310004124.
Chang, Rachel et al. 2018. “Thioclava Electrotropha Sp. Nov., a Versatile Electrode and Sulfur-Oxidizing
Bacterium from Marine Sediments.” International Journal of Systematic and Evolutionary
Microbiology 68(5): 1652–58.
Chen, Chuanfang, Qiufeng Ma, Wei Jiang, and Tao Song. 2011. “Phototaxis in the Magnetotactic
Bacterium Magnetospirillum Magneticum Strain AMB-1 Is Independent of Magnetic Fields.”
Applied Microbiology and Biotechnology 90(1): 269–75.
Chen, Zhongwen et al. 2012. “Dual Enzyme-like Activities of Iron Oxide Nanoparticles and Their
Implication for Diminishing Cytotoxicity.” ACS Nano 6(5): 4001–12.
Chong, Grace W., Amruta A. Karbelkar, and Mohamed Y. El-Naggar. 2018. “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.
https://doi.org/10.1016/j.cbpa.2018.06.007.
Cohoe, David, Iulia Hanczarek, J. Kent Wallace, and Jay Nadeau. 2019. “Multiwavelength Digital
Holographic Imaging and Phase Unwrapping of Protozoa Using Custom Fiji Plug-Ins.” Frontiers in
Physics 7(JULY): 1–14. https://www.frontiersin.org/article/10.3389/fphy.2019.00094/full.
Coleman, R. G., and D. E. Lee. 1962. “Metamorphic Aragonite in the Glaucophane Schists of Cazadero,
California.” American Journal of Science 260(8): 577–95.
http://www.ajsonline.org/cgi/doi/10.2475/ajs.260.8.577.
Coleman, R.G. “Prospecting for Ophiolites along the California Continental Margin.” In In Ophiolites and
Oceanic Crust: New Insights from Field Studies and the Ocean Drilling Program., ed. and A. Nicolas
Y. Dilek, E. Moores, D. Elthon. Boulder, CO: Geological Society of America, 351–64.
Cox, L.B. et al. 2002. “Organization and Elemental Analysis of P-, S-, and Fe-Rich Inclusions in a
Population of Freshwater Magnetococci.” Geomicrobiology Journal 19(4): 387–406.
Dean, W E, W S Moore, and K H Nealson. 1981. “Manganese Cycles and the Origin of Manganese
142

Nodules, Oneida Lake, New York, USA.” Chemical Geology 34: 53–64.
El-Naggar, M.Y., and S.E. Finkel. 2013. “Live Wires.” Scientist: 38–43.
El-Naggar, Mohamed Y. et al. 2010. “Electrical Transport along Bacterial Nanowires from Shewanella
Oneidensis MR-1.” Proceedings of the National Academy of Sciences of the United States of
America 107(42): 18127–31.
Ester, Martin, Hans-Peter Kriegel, Jörg Sander, and Xiaowei Xu. 1996. “A Density-Based Algorithm for
Discovering Clusters in Large Spatial Databases with Noise.” Proceedings of the 2nd International
Conference on Knowledge Discovery and Data Mining: 226–31.
Frankel, R. B. et al. 1981. “Magnetotactic Bacteria at the Geomagnetic Equator.” Science 212(4500):
1269–70. https://www.sciencemag.org/lookup/doi/10.1126/science.212.4500.1269.
Frankel, Richard B., and Dennis A. Bazylinski. 2009. “Magnetosomes and Magneto-Aerotaxis.”
Contributions to Microbiology 16: 182–93.
Frankel, Richard B., and R. P. Blakemore. 1980. “Navigational Compass in Magnetic Bacteria.” Journal of
Magnetism and Magnetic minerals 15: 1562–64.
https://digitalcommons.calpoly.edu/cgi/viewcontent.cgi?article=1108&context=phy_fac.
Fredrickson, James K. et al. 2008. “Towards Environmental Systems Biology of Shewanella.” Nature
Reviews Microbiology 6(8): 592–603.
Frost, Ronald B., and James S. Beard. 2007. “On Silica Activity and Serpentinization.” Journal of Petrology
48(7): 1351–68.
Garber, Arkadiy I. et al. 2020. “FeGenie: A Comprehensive Tool for the Identification of Iron Genes and
Iron Gene Neighborhoods in Genome and Metagenome Assemblies.” Frontiers in Microbiology
11(January): 1–23.
Gralnick, Jeffrey A. 2012. “On Conducting Electron Traffic across the Periplasm.” Biochemical Society
Transactions 40(6): 1178–80.
Gralnick, Jeffrey A., and Dianne K. Newman. 2007. “Extracellular Respiration.” Molecular Microbiology
65(1): 1–11.
Guo, Fang Fang et al. 2012. “Magnetosomes Eliminate Intracellular Reactive Oxygen Species in
Magnetospirillum Gryphiswaldense MSR-1.” Environmental Microbiology 14(7): 1722–29.
Hand, K et al. 2017. February, 2017 Report of the Europa Lander Science Definition Team.
Harris, H. W. et al. 2010. “Electrokinesis Is a Microbial Behavior That Requires Extracellular Electron
Transport.” Proceedings of the National Academy of Sciences of the United States of America
107(1): 326–31.
Harris, H. Wayne, Mohamed Y. El-Naggar, and Kenneth H. Nealson. 2012. “Shewanella Oneidensis MR-1
Chemotaxis Proteins and Electron-Transport Chain Components Essential for Congregation near
Insoluble Electron Acceptors.” Biochemical Society Transactions 40(6): 1167–77.
http://biochemsoctrans.org/lookup/doi/10.1042/BST20120232.
Harris, Howard W. et al. 2018. “Redox Sensing within the Genus Shewanella.” Frontiers in Microbiology
8(JAN): 1–11.
143

Hau, Heidi H., and Jeffrey A. Gralnick. 2007. “ Ecology and Biotechnology of the Genus Shewanella .”
Annual Review of Microbiology 61(1): 237–58.
Heidelberg, John F. et al. 2002. “Genome Sequence of the Dissimilatory Metal Ion-Reducing Bacterium
Shewanella Oneidensis.” Nature Biotechnology 20(11): 1118–23.
Jogler, C. et al. 2010. “Cultivation-Independent Characterization of ‘Candidatus Magnetobacterium
Bavaricum’ via Ultrastructural, Geochemical, Ecological and Metagenomic Methods.”
Environmental Microbiology 12(9): 2466–78.
Karbelkar, Amruta A., Annette R. Rowe, and Mohamed Y. El-Naggar. 2019. “An Electrochemical
Investigation of Interfacial Electron Uptake by the Sulfur Oxidizing Bacterium Thioclava
Electrotropha ElOx9.” Electrochimica Acta 324(2).
Kato, Souichiro et al. 2010. “Respiratory Interactions of Soil Bacteria with (Semi)Conductive Iron-Oxide
Minerals.” Environmental Microbiology 12(12): 3114–23.
Kim, Byung Hong, Hyung Joo Kim, Moon Sik Hyun, and Doo Hyun Park. 1999. “Direct Electrode Reaction
of Fe(III)-Reducing Bacterium, Shewanella Putrefaciens.” Journal of Microbiology and
Biotechnology 9(2): 127–31.
Kohl, Lukas. Cumming, Emily. 2016. “Exploring the Metabolic Potential of Microbial Communities in
Ultra-Basic, Reducing Springs at The Cedars, CA, USA....” : 1203–20.
Komeili, Arash. 2012. “Molecular Mechanisms of Compartmentalization and Biomineralization in
Magnetotactic Bacteria.” FEMS Microbiology Reviews 36(1): 232–55.
Komeili, Arash, Zhuo Li, Dianne K. Newman, and Grant J. Jensen. 2006. “Magnetosomes Are Cell
Membrane Invaginations Organized by the Actin-Like Protein MamK.” Science 311(5758): 242–45.
https://www.sciencemag.org/lookup/doi/10.1126/science.1123231.
Komeili, Arash, Hojatollah Vali, Terrance J. Beveridge, and Dianne K. Newman. 2004. “Magnetosome
Vesicles Are Present before Magnetite Formation, and MamA Is Required for Their Activation.”
Proceedings of the National Academy of Sciences of the United States of America 101(11): 3839–
44.
Kopp, Robert E., and Joseph L. Kirschvink. 2008. “The Identification and Biogeochemical Interpretation of
Fossil Magnetotactic Bacteria.” Earth-Science Reviews 86(1–4): 42–61.
Kühn, Jonas et al. 2014. “A Mach-Zender Digital Holographic Microscope with Sub-Micrometer
Resolution for Imaging and Tracking of Marine Micro-Organisms.” Review of Scientific Instruments
85(12). http://dx.doi.org/10.1063/1.4904449.
Lam, Bonita R. 2018. “Extracellular Electron Transport: Investigating the Diversity and Mechanisms
Behind an Understudied Microbial Process with Global Implications.” University of Southern
California.
Lam, Bonita R., Casey R. Barr, Annette R. Rowe, and Kenneth H. Nealson. 2019. “Differences in Applied
Redox Potential on Cathodes Enrich for Diverse Electrochemically Active Microbial Isolates From a
Marine Sediment.” Frontiers in Microbiology 10(August): 1–17.
Lefèvre, C. T., and D. A. Bazylinski. 2013. “Ecology, Diversity, and Evolution of Magnetotactic Bacteria.”
Microbiology and Molecular Biology Reviews 77(3): 497–526.
144

Lefèvre, Christopher T. et al. 2013. “Monophyletic Origin of Magnetotaxis and the First Magnetosomes.”
Environmental Microbiology 15(8): 2267–74.
Lefèvre, Christopher T. et al. 2014. “Diversity of Magneto-Aerotactic Behaviors and Oxygen Sensing
Mechanisms in Cultured Magnetotactic Bacteria.” Biophysical Journal 107(2): 527–38.
Lefèvre, Christopher T., and Long Fei Wu. 2013. “Evolution of the Bacterial Organelle Responsible for
Magnetotaxis.” Trends in Microbiology 21(10): 534–43.
Lin, W. C., M. V. Coppi, and D. R. Lovley. 2004. “Geobacter Sulfurreducens Can Grow with Oxygen as a
Terminal Electron Acceptor.” Applied and Environmental Microbiology 70(4): 2525–28.
http://www.ncbi.nlm.nih.gov/pubmed/21861844.
Lin, Wei et al. 2013. “Integrating Niche-Based Process and Spatial Process in Biogeography of
Magnetotactic Bacteria.” Scientific Reports 3: 1–9.
Lin, Wei et al. 2017. “Origin of Microbial Biomineralization and Magnetotaxis during the Archean.”
Proceedings of the National Academy of Sciences of the United States of America 114(9): 2171–76.
Lin, Wei et al. 2018. “Genomic Expansion of Magnetotactic Bacteria Reveals an Early Common Origin of
Magnetotaxis with Lineage-Specific Evolution.” ISME Journal 12(6): 1508–19.
http://dx.doi.org/10.1038/s41396-018-0098-9.
Lin, Wei et al. 2020. “On the Origin of Microbial Magnetoreception.” National Science Review 7(2): 472–
79.
Lin, Wei, Yongxin Pan, and Dennis A. Bazylinski. 2017. “Diversity and Ecology of and Biomineralization by
Magnetotactic Bacteria.” Environmental Microbiology Reports 9(4): 345–56.
Lindensmith, Christian A. et al. 2016. “A Submersible, off-Axis Holographic Microscope for Detection of
Liu, Yimo et al. 2014. “A Trans-Outer Membrane Porin-Cytochrome Protein Complex for Extracellular
Electron Transfer by Geobacter Sulfurreducens PCA.” Environmental Microbiology Reports 6(6):
776–85.
Lohmann, K. J., S. D. Cain, S. A. Dodge, and C. M.F. Lohmann. 2001. “Regional Magnetic Fields as
Navigational Markers for Sea Turtles.” Science 294(5541): 364–66.
Lovley, D R, and E J Phillips. 1987. “Rapid Assay for Microbially Reducible Ferric Iron in Aquatic
Sediments.” Applied and environmental microbiology 53(7): 1536–40.
Mahadevan, Radhakrishnan, Bernhard Palsson, and Derek R. Lovley. 2011. “In Situ to in Silico and Back:
Elucidating the Physiology and Ecology of Geobacter Spp. Using Genome-Scale Modelling.” Nature
Reviews Microbiology 9(1): 39–50.
Marques, José M. et al. 2008. “Origins of High PH Mineral Waters from Ultramafic Rocks, Central
Portugal.” Applied Geochemistry 23(12): 3278–89.
http://dx.doi.org/10.1016/j.apgeochem.2008.06.029.
McCausland, Hayley C., and Arash Komeili. 2020. “Magnetic Genes: Studying the Genetics of
Biomineralization in Magnetotactic Bacteria.” PLoS genetics 16(2): e1008499.
McGlynn, Shawn E., Grayson L. Chadwick, Christopher P. Kempes, and Victoria J. Orphan. 2015. “Single
Cell Activity Reveals Direct Electron Transfer in Methanotrophic Consortia.” Nature 526(7574):
145

531–35.
McMurdie, Paul J., and Susan Holmes. 2013. “Phyloseq: An R Package for Reproducible Interactive
Analysis and Graphics of Microbiome Census Data.” PLoS ONE 8(4).
Meysman, Filip J.R. 2018. “Cable Bacteria Take a New Breath Using Long-Distance Electricity.” Trends in
Microbiology 26(5): 411–22. http://dx.doi.org/10.1016/j.tim.2017.10.011.
Morrill, Penny L. et al. 2013. “Geochemistry and Geobiology of a Present-Day Serpentinization Site in
California: The Cedars.” Geochimica et Cosmochimica Acta 109(May): 222–40.
http://dx.doi.org/10.1016/j.gca.2013.01.043.
Myers, Charles R., and Kenneth H. Nealson. 1988. “Microbial Reduction of Manganese Oxides:
Interactions with Iron and Sulfur.” Geochimica et Cosmochimica Acta 52(11): 2727–32.
Nealson, K. H., C. R. Myers, and B. B. Wimpee. 1991. “Isolation and Identification of Manganese-
Reducing Bacteria and Estimates of Microbial Mn(IV)-Reducing Potential in the Black Sea.” Deep-
Sea Research, Part A 38(Suppl. 2A): S907–20. http://dx.doi.org/10.1016/S0198-0149(10)80016-0.
Nelson, William C., and James C. Stegen. 2015. “The Reduced Genomes of Parcubacteria (OD1) Contain
Signatures of a Symbiotic Lifestyle.” Frontiers in Microbiology 6(JUN): 1–14.
Nicholas L. Taliaferro. 1943. “Franciscan-Knoxville Problem.” AAPG Bulletin 27(1514): 12–12.
https://www.sciencemag.org/lookup/doi/10.1126/science.59.1514.12-a.
Pfeffer, Christian et al. 2012. “Filamentous Bacteria Transport Electrons over Centimetre Distances.”
Nature 491(7423): 218–21. http://dx.doi.org/10.1038/nature11586.
Philippe, Nadège, and Long Fei Wu. 2010. “An MCP-Like Protein Interacts with the MamK Cytoskeleton
and Is Involved in Magnetotaxis in Magnetospirillum Magneticum AMB-1.” Journal of Molecular
Biology 400(3): 309–22. http://dx.doi.org/10.1016/j.jmb.2010.05.011.
Raiche, R. 2009. “The Cedars: Sonoma County’s Hidden Treasure.” Fr 37(2): 3–15.
Richardson, David J. et al. 2012. “The ‘porin-Cytochrome’ Model for Microbe-to-Mineral Electron
Transfer.” Molecular Microbiology 85(2): 201–12.
Rosenberg, Eugene, and F. DeLong, E.F., Stackebrandt, E., Lory, S., Thompson. 2013. The Prokaryotes -
Prokaryotic Biology and Symbiotic | Eugene Rosenberg | Springer.
http://www.springer.com/us/book/9783642301933.
Rowe, Annette R. et al. 2015. “Marine Sediments Microbes Capable of Electrode Oxidation as a
Surrogate for Lithotrophic Insoluble Substrate Metabolism.” Frontiers in Microbiology 6(JAN): 1–
15.
Rowe, Annette R. et al. 2017. “In Situ Electrochemical Enrichment and Isolation of a Magnetite-Reducing
Bacterium from a High PH Serpentinizing Spring.” Environmental Microbiology 19(6): 2272–85.
Rowe, Annette R. et al. 2018. “Tracking Electron Uptake from a Cathode into Shewanella Cells:
Implications for Energy Acquisition from Solid-Substrate Electron Donors.” mBio 9(1): 1–19.
Schindelin, Johannes et al. 2012. “Fiji: An Open-Source Platform for Biological-Image Analysis.” Nature
Methods 9(7): 676–82.
Shi, Liang et al. 2016. “Extracellular Electron Transfer Mechanisms between Microorganisms and
146

Minerals.” Nature Reviews Microbiology 14(10): 651–62.
Simmons, Sheri L., Dennis A. Bazylinski, and Katrina J. Edwards. 2006. “South-Seeking Magnetotactic
Bacteria in the Northern Hemisphere.” Science 311(5759): 371–74.
Sleep, Norman H., Dennis K. Bird, and Emily C. Pope. 2011. “Serpentinite and the Dawn of Life.”
Philosophical Transactions of the Royal Society B: Biological Sciences 366(1580): 2857–69.
Summers, Z. M. et al. 2010. “Direct Exchange of Electrons Within Aggregates of an Evolved Syntrophic
Coculture of Anaerobic Bacteria.” Science 330(6009): 1413–15.
https://www.sciencemag.org/lookup/doi/10.1126/science.1196526.
Suzuki, Shino et al. 2013. “Microbial Diversity in the Cedars, an Ultrabasic, Ultrareducing, and Low
Salinity Serpentinizing Ecosystem.” Proceedings of the National Academy of Sciences of the United
States of America 110(38): 15336–41.
Suzuki, Shino et al. 2014. “Physiological and Genomic Features of Highly Alkaliphilic Hydrogen-Utilizing
Betaproteobacteria from a Continental Serpentinizing Site.” Nature Communications 5(May).
Suzuki, Shino et al. 2017. “Unusual Metabolic Diversity of Hyperalkaliphilic Microbial Communities
Associated with Subterranean Serpentinization at the Cedars.” ISME Journal 11(11): 2584–98.
http://dx.doi.org/10.1038/ismej.2017.111.
Tebo, B, R Rosson, and K Nealson. 1991. “Potential for Manganese(II) Oxidation and Manganese(IV)
Reduction to Co-Occur in the Suboxic Zone of the Black Sea.” In Black Sea Oceanography. NATO ASI
Series, ed. Murray J.W İzdar E. Dordrecht: Springer.
Tiago, Igor, and António Veríssimo. 2013. “Microbial and Functional Diversity of a Subterrestrial High PH
Groundwater Associated to Serpentinization.” Environmental Microbiology 15(6): 1687–1706.
Trojan, Daniela et al. 2016. “A Taxonomic Framework for Cable Bacteria and Proposal of the Candidate
Genera Electrothrix and Electronema.” Systematic and Applied Microbiology 39(5): 297–306.
http://dx.doi.org/10.1016/j.syapm.2016.05.006.
Uebe, René, and Dirk Schüler. 2016. “Magnetosome Biogenesis in Magnetotactic Bacteria.” Nature
Reviews Microbiology 14(10): 621–37.
Vali, H., and Joseph L. Kirschvink. 1991. “Observations of Magnetosome Organization, Surface Structure,
and Iron Biomineralization of Undescribed Magnetic Bacteria: Evolutionary Speculations.” In Iron
Biominerals, eds. R. B. Frankel and R. P. Blakemore. Boston, MA: Springer.
Wallace, J. Kent et al. 2015. “Robust, Compact Implementation of an off-Axis Digital Holographic
Microscope.” Optics Express 23(13): 17367. https://www.osapublishing.org/abstract.cfm?URI=oe-
23-13-17367.
Wegener, Gunter et al. 2015. “Intercellular Wiring Enables Electron Transfer between Methanotrophic
Archaea and Bacteria.” Nature 526(7574): 587–90.
Wrighton, K. C. et al. 2011. “Evidence for Direct Electron Transfer by a Gram-Positive Bacterium Isolated
from a Microbial Fuel Cell.” Applied and Environmental Microbiology 77(21): 7633–39.

147

Other Publications
Bedrossian, Manuel et al. 2017. “Quantifying Microorganisms at Low Concentrations Using
Digital Holographic Microscopy (DHM).” Journal of Visualized Experiments 2017(129): 1–
11.
Bedrossian, Manuel et al.(in Prep). “Prevalence and Stimulation of Motility in Fresh Field
Samples from Diverse and Extreme Environments across North America.”
Bird, Lina et al. (in Prep) “Serpentinomonas Raichei and S. Mccroryi, Gen. Nov., Sp. Nov., Two
Hyperalkaliphilic, Hydrogen-Oxidizing Microaerobic Bacteria Isolated from Terrestrial
Serpentinizing Springs.”
Chang, Rachel et al. 2018. “Thioclava Electrotropha Sp. Nov., a Versatile Electrode and Sulfur-
Oxidizing Bacterium from Marine Sediments.” International Journal of Systematic and
Evolutionary Microbiology 68(5): 1652–58.
Lee, C et al. (in review). “MOTS-c is an Interferon-Responsive Immune Factor Encoded in the
Mitochondrial Genome.”
Lam, Bonita R., Casey R. Barr, Annette R. Rowe, and Kenneth H. Nealson. 2019. “Differences in
Applied Redox Potential on Cathodes Enrich for Diverse Electrochemically Active
Microbial Isolates From a Marine Sediment.” Frontiers in Microbiology 10(August): 1–17.
Sarbu, Serban et al. 2018. “Sulfur Cave (Romania), an Extreme Environment with Microbial
Mats in a CO2-H2S/O2 Gas Chemocline Dominated by Mycobacteria.” International Journal
of Speleology 47(2): 173–87. http://scholarcommons.usf.edu/ijs/vol47/iss2/7/.
Thompson, J. et al. (in Prep). “Characterization of Two Novel Alkaliphiles Isolated from The
Cedars, a Terrestrial Serpentinizing System in Northern California.”

Abstract (if available)

Abstract This thesis explores the use of Digital Holographic Microscopy (DHM) to observe and characterize some of the most fundamental behaviors of various geobiologically relevant bacteria. DHM is a non-invasive volumetric interferometry-based technique with vast potential in the study of single celled microorganisms both in the laboratory and in the field. DHM instruments provide the ability to reconstruct three-dimensional space (while maintaining sub-micron resolution) over time, allowing microbiological researchers the opportunity to investigate the motile response of individual cells as they navigate the environment under natural ambient conditions or during experimental runs in which specific behaviors are elicited and described. As the technological innovation of DHM capabilities rapidly approaches a lower cost threshold and more user-friendly state, this thesis exists primarily as a collection of research that explores the development and application of DHM in the field of geobiology and environmental microbiology. ❧ Chapter I provides a background of the DHM system: in theory, development and practice. Highlighting both the advantages and hurdles of this up-and-coming technique as of the year 2020 and describing a pathway to building an inexpensive DHM instrument. Chapter II presents an experimental DHM-electrochemical cell system developed to study the microscale interactions of bacteria capable of extracellular electron transport (EET) using poised electrode surfaces, providing insight into the blossoming field of electromicrobiology. Chapter III demonstrates how the DHM platform can be used to quantify microbial cell density in an extreme environment (pH ≥ 11) where cell numbers are very low (< 10³ cells/mL) and laboratory methods for the study of the unique organisms endemic to these systems. Chapter IV investigates the response of magnetotactic bacteria (MTB) from a mixed environmental sample at realistic magnetic field strengths through the incorporation of DHM and a tri-axial coil system, allowing for the identification, quantification and characterization of the magnetotaxis response of individual cells at different environmentally-relevant field strengths for the first time. ❧ In all of these studies, it is clear that DHM provides a novel view of the microbial world, allowing researchers to view the three-dimensional environment that these simple lifeforms inhabit, navigate and interact in. As the artwork of Farooq Azam helped illustrate the nature of microbial ecology and reality of the microscopic environment, DHM is a tool that allows us to view this world directly and in real-time (Figure A.1). As much of the microbial sciences has turned toward molecular and -omics techniques, direct observations of individual cells are fundamental in grounding our perceived and theorized assumptions of the organisms in nature. This thesis attempts to explore exactly that and provide a launch pad for DHM technology in the microbial sciences.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Investigating microbial biofilm community mediated processes on surfaces: from single cell genomics to community meta-omics

PDF

Genomic and physiological characterization of Escherichia coli evolving in long-term batch culture

PDF

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

PDF

Non invasive detection of microbes in natural environments using deep UV native fluorescence spectroscopy and hyperspectral imaging

PDF

Bacterial survival under low nutrient conditions in the laboratory and the deep biosphere

PDF

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

PDF

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

PDF

Geomicrobiology in the shallow-sea hydrothermal system at Milos Island, Greece

PDF

Habitability of Saturn's moon Titan for acetylenotrophy: laboratory culturing and energetics as tools for the search for life

PDF

Aerobic respiration rates of bacteria: the influence of nutrient-limitation, cell density, and nutrient pulsing

PDF

The geobiological role of bioturbating ecosystem engineers during key evolutionary intervals in Earth history

PDF

Electronic, electrochemical, and spintronic characterization of bacterial electron transport

PDF

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

PDF

Nutritional competence in Escherichia coli: the use of double-stranded DNA as a nutrient

PDF

Quorum sensing interaction networks in bacterial communities

PDF

Exploring bacteria-mineral interactions

PDF

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

PDF

An investigation into low current density and microbially mediated arsenic electrocoagulation kinetics

PDF

Microbial metabolism in deep subsurface sediments of Guaymas Basin (Gulf of California): methanogenesis, methylotrophy, and asgardarchaeota

PDF

Genetic diversity and bacterial death in the context of adaptive evolution

Asset Metadata

Creator Barr, Casey R. (author)

Core Title Investigations in the field of geobiology through the use of digital holographic microscopy

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Geological Sciences

Publication Date 10/28/2020

Defense Date 08/10/2020

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

Tag astrobiology,digital holographic microscopy,EET,extracellular electron transport,Geobiology,magnetotactic bacteria,microscopy,OAI-PMH Harvest

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Nealson, Kenneth H. (committee chair), Amend, Jan (committee member), Finkel, Steven (committee member)

Creator Email caseyrbarr@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-386965

Unique identifier UC11665811

Identifier etd-BarrCaseyR-9078.pdf (filename),usctheses-c89-386965 (legacy record id)

Legacy Identifier etd-BarrCaseyR-9078.pdf

Dmrecord 386965

Document Type Dissertation

Rights Barr, Casey R.

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