Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Dps contributes to typical growth, survival, and genome organization in E. coli
(USC Thesis Other)
Dps contributes to typical growth, survival, and genome organization in E. coli
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
Dps contributes to typical growth, survival, and genome organization in E. coli
By
Katie Orban
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
(MOLECULAR BIOLOGY)
August 2022
Copyright 2022 Katie Orban
ii
Acknowledgments
The work presented in this dissertation would not have been possible without Ray R. Irani
Hall and my home, the latter of which served as my workspace for much of the time between
March 2020 and the present. These spaces were built on stolen Gabrielino-Tongva land. I
acknowledge them as the past, present, and future caretakers of this land.
I have many people (and one animal) to thank for their support over the past six years.
First, to Patrick Alcerro: thank you for loving me as I am. To my parents: thank you for raising me
to believe in myself and for supporting me through all my endeavors. To Nikki: thank you for
lifting me up when I feel down. To the Alcerros: thank you for accepting me into your family.
Thank you also to the many amazing friends I have made through the MCB program at USC and
beyond – you have been my rock over the past six years. Last, but not least, I thank my dog, Bosco,
for his emotional support and endless energy, which have kept me grounded for the past year.
I have several people to thank for their support of my development as a scientist. To Carl
Urbinati, my undergraduate research advisor at Loyola Marymount University: thank you for
taking a chance on me and being a wonderful mentor and friend. To Steve Finkel, my Ph.D.
advisor: thank you for allowing me to follow my curiosity. To the members of my committee, the
Finkel lab, Maria Pellegrini, and Molly Schmid: thank you for your critical eye and honest
feedback, which have helped me grow as both a critical thinker and a communicator.
I also need to acknowledge two groups that have kept me sane during my time at USC.
First, to the Molecular Biology Graduate Student Association (MBGSA): thank you for fostering
a supportive environment in MCB. Second, to Women in Molecular Biology (WiM): words cannot
express the pride and gratitude I feel watching women lift each other up even under the toughest
of circumstances. I hope you can continue this important work for years to come.
iii
Table of Contents
Acknowledgments ........................................................................................................................... ii
List of Tables .................................................................................................................................. v
List of Figures ................................................................................................................................ vi
Abstract ......................................................................................................................................... vii
Preface .......................................................................................................................................... viii
Chapter 1: Dps is a universally conserved dual-action DNA-binding and ferritin protein ............ 1
Introduction ................................................................................................................................. 2
Monomer Structure ..................................................................................................................... 4
Dodecamer Structure .................................................................................................................. 5
Other Quaternary Structures ....................................................................................................... 6
Dps as a DNA binding protein .................................................................................................... 7
Interactions with other NAPs .................................................................................................... 11
Dps as a ferritin ......................................................................................................................... 12
Dps Expression ......................................................................................................................... 15
Dps as a Regulator of Gene Expression .................................................................................... 17
Macromolecular sequestration and phase separation ................................................................ 19
Stress Response ......................................................................................................................... 21
Other functions ......................................................................................................................... 23
Dps in nanotechnology ............................................................................................................. 25
Future directions ....................................................................................................................... 26
Conclusion ................................................................................................................................ 29
Figures ...................................................................................................................................... 30
Tables ........................................................................................................................................ 35
Chapter 2: Differential contributions of ferritins to long-term survival in Escherichia coli ........ 40
Introduction ............................................................................................................................... 41
Results ....................................................................................................................................... 44
Contribution of ferritins to wild-type growth phenotype in long-term batch monoculture .. 44
Competition of mutant strains with wild-type ....................................................................... 46
Competition of mutant strains against one another .............................................................. 47
Varying iron availability impacts ferritin mutant monoculture growth and survival .......... 48
Varying iron availability impacts ferritin growth and survival in competition .................... 49
Ferritins are differentially expressed by growth phase ........................................................ 51
Discussion ................................................................................................................................. 52
Conclusion ................................................................................................................................ 57
Materials and methods .............................................................................................................. 58
Figures ...................................................................................................................................... 60
Tables ........................................................................................................................................ 67
iv
Chapter 3: E. coli Dps organizes the stationary phase nucleoid in a non-random manner ........... 70
Introduction ............................................................................................................................... 71
Results ....................................................................................................................................... 73
Experimental design .............................................................................................................. 73
DNA is more accessible in dps-null nucleoids during stationary phase .............................. 74
A putative Dps binding motif ................................................................................................ 75
Discussion ................................................................................................................................. 77
Conclusion ................................................................................................................................ 81
Methods .................................................................................................................................... 82
Figures ...................................................................................................................................... 85
Tables ........................................................................................................................................ 89
Chapter 4: MNase-Seq reveals E. coli stationary phase genome accessibility ............................. 97
Introduction ............................................................................................................................... 98
Results ..................................................................................................................................... 100
Experimental Design ........................................................................................................... 100
Calling peaks ...................................................................................................................... 101
Motif analysis ...................................................................................................................... 102
Discussion ............................................................................................................................... 104
Conclusion .............................................................................................................................. 108
Figures .................................................................................................................................... 109
Tables ...................................................................................................................................... 115
Methods .................................................................................................................................. 120
Chapter 5: Dps contributes to wild-type population dynamics and genome organization ......... 123
Introduction ............................................................................................................................. 124
A ferritin hierarchy in E. coil .................................................................................................. 126
Investigating the nature of the Dps-DNA biocrystal .............................................................. 129
Dps-dependent chromatin accessibility revealed .................................................................... 131
Discussion ............................................................................................................................... 133
Conclusion .............................................................................................................................. 135
Figures .................................................................................................................................... 136
Tables ...................................................................................................................................... 137
References ................................................................................................................................... 138
v
List of Tables
Table 1.1: Key Dps properties in several bacterial species .......................................................... 35
Table 1.2: Dps as a stress response protein ................................................................................... 38
Table 2.1: Lag time and log growth rate by strain ........................................................................ 67
Table 2.2: Strains used in this study ............................................................................................. 68
Table 2.3: Competition schema .................................................................................................... 69
Table 3.1: Consensus sequence in every stationary phase-specific differentially protected
region ............................................................................................................................................ 89
Table 3.2: Some E. coli NAP DNA binding motifs and their occurrence in the wild-type
genome .......................................................................................................................................... 90
Table 3.3: Putative Dps motif is overrepresented in stationary phase-specific significant sites .. 91
Table 3.4: Strains used in this study ............................................................................................. 92
Table 3.5: Primers used in this study ............................................................................................ 93
Table 4.1: Peak length varies by sample ..................................................................................... 115
Table 4.2: Putative motif occurrences in the wild-type E. coli genome ..................................... 116
Table 4.3: Putative motif occurrences in peak sets ..................................................................... 117
Table 4.4: Putative motif occurrences in raw data ...................................................................... 118
Table 5.1: Motif frequency in alternate datasets ......................................................................... 137
vi
List of Figures
Figure 1.1: Dps monomer structure displays homology to ferritins ............................................. 30
Figure 1.2: Quaternary structure of E. coli Dps dodecamer and ferritin 24-mers ........................ 31
Figure 1.3: Model of E. coli Dps-DNA binding ........................................................................... 32
Figure 1.4: E. coli Dps induces formation of the stationary phase-specific biocrystal ................ 33
Figure 1.5: Major E. coli NAP abundance by growth phase ........................................................ 34
Figure 2.1: Monoculture of wild-type and ferritin mutants .......................................................... 60
Figure 2.2: Monoculture outgrowth of wild-type and ferritin mutants ......................................... 61
Figure 2.3: Ferritin mutant coculture with wild-type populations ................................................ 62
Figure 2.4: Ferritin mutant pairwise coculture ............................................................................. 63
Figure 2.5: Monoculture with varied iron availability .................................................................. 64
Figure 2.6: Competitions among ferritin-null populations with varied iron availability .............. 65
Figure 2.7: Ferritin gene expression profiles ................................................................................ 66
Figure 3.1: A PCR-based approach to assess genome accessibility ............................................. 85
Figure 3.2: Restriction enzyme loci for AvrII, NotI, and XbaI along the E. coli chromosome ... 86
Figure 3.3: Nucleoid accessibility changes with growth phase and presence of dps ................... 87
Figure 3.4: Putative Dps DNA-binding motif ............................................................................... 88
Figure 4.1: MNase-Seq experimental design .............................................................................. 109
Figure 4.2: Peaks vary by sample and are distributed across the genome .................................. 110
Figure 4.3: Populations reveal differential inclusion in peak sets .............................................. 111
Figure 4.4: Peak lengths vary by strain and timepoint ................................................................ 112
Figure 4.5: Putative Dps DNA-binding motif ............................................................................. 114
Figure 5.1: PCR- and MNase-Seq-derived motif alignment ...................................................... 136
vii
Abstract
The DNA-binding protein from starved cells, Dps, is a highly abundant protein during
stationary phase in Escherichia coli. It combines two distinct biological functions: as a ferritin it
detoxifies and stores iron, and as a DNA-binding protein it binds and structures the chromosome.
These activities, though biochemically separable, function jointly to protect the cell from a variety
of stresses, the most well-studied of which is oxidative stress. This dissertation focuses on both
major aspects of Dps function. First, in comparing the growth and survival phenotypes of ferritin-
null E. coli populations, growth phase-dependent importance of each ferritin gene is elucidated.
While Dps has the most dramatic impact on growth and survival of the E. coli ferritins, iron
availability experiments suggest this impact is only minimally due to its ferritin function, leading
to the hypothesis that its major role in growth and survival is as a DNA-binding protein. However,
Dps has been identified as a non-specific DNA-binding protein. As such, the next question
addressed in this dissertation is whether Dps binds the same genomic loci in a repeatable manner.
Using a novel assay to probe chromatin accessibility, Dps is shown here to bind at least a portion
of the chromosome repeatably. Finally, MNase treatment followed by DNA sequencing (MNase-
Seq) was used to determine chromosome accessibility on a more high-resolution scale. The two
different chromatin accessibility experiments described in this dissertation reveal two putative
Dps-DNA binding motifs. The first, determined by comparing Dps-dependent protected
chromosomal regions between mid-log phase and late stationary phase, may denote a primary Dps
binding motif. The second, determined by comparing Dps-dependent protected chromosomal
regions shared in early and late stationary phase, may denote a secondary Dps binding motif.
viii
Preface
The subject of this dissertation is the DNA-binding protein from starved cells (Dps) in
Escherichia coli. Dps has two main functions. One is DNA binding, in which capacity it acts as a
nucleoid-associated protein (NAP), a major structural protein component of the bacterial nucleoid,
the region of the bacterial cell that contains DNA in complex with protein. The other is iron
detoxification and storage, in which capacity it acts as a ferritin protein.
Chapter 1 describes the state of knowledge regarding Dps. Briefly, Dps is highly conserved
across bacteria, and acts universally as a ferritin protein, with fewer Dps species possessing DNA-
binding activity. Dps is necessary for a stationary phase-specific chromatin restructuring event into
a hexacrystalline lattice known as the biocrystal. While these characteristics are biochemically
separable in E. coli and most other dual-action Dps species, they function in concert to protect the
cell from oxidative stress. Several Dps proteins possess peripheral functionality, one of the most
debated of which is regulation of global gene expression. In addition to these functions, Chapter 1
also discusses Dps structure, interactions between Dps and other NAPs, and phase separation, as
well as the potential applications of Dps in nanotechnology.
Chapter 2 studies the ferritin functions of E. coli Dps. Growth and survival curves of wild-
type and ferritin mutant strains were used to evaluate the impact of Dps and the two other E. coli
ferritins (ferritin and bacterioferritin) on population dynamics. Monoculture and coculture
population dynamics were examined under typical laboratory conditions, as well as in the presence
of added iron or an iron chelator. These methods ultimately suggest a hierarchy among E. coli
ferritin proteins as they pertain to population growth and survival.
Chapter 3 investigates a key facet of Dps DNA-binding ability. Dps has been classified as
a non-specific DNA binding protein. Moreover, to date, no published work has addressed whether
ix
Dps binds specific genomic loci in a biologically programmed manner. Using restriction digestion
of crosslinked chromatin followed by PCR, this chapter explores Dps DNA-binding plasticity
across dozens of endogenous restriction sites in the E. coli chromosome. This work confirms how
Dps impacts chromosome accessibility during stationary phase and identifies novel loci that are
biologically programmed for Dps binding.
Chapter 4 probes Dps DNA-binding at higher resolution. Micrococcal nuclease followed
by sequencing (MNase-Seq) was performed during early stationary phase and late stationary phase
in wild-type and dps-null populations. In addition to demonstrating differential accessibility across
strains and timepoints, this data also provides a subset of chromosomal loci that are bound only
when Dps is present during early and late stationary phase. These loci may point to a Dps DNA-
binding consensus sequence, the first time in situ data has done so.
Chapter 5 discusses the strides made toward more fully understanding E. coli Dps in this
dissertation. In this chapter, results from each chapter are compared to better appreciate Dps action.
The hierarchy established among ferritins in Chapter 2 suggests that the growth and survival
advantage Dps confers is due, in large part, to its DNA-binding function. This led to the study of
Dps DNA-binding laid out in Chapter 3. Because Chapter 3 points to some amount of biologically
programmed Dps DNA-binding, a higher resolution method was utilized in Chapter 4. The MNase-
Seq experiment described in Chapter 4 showcases where Dps binds in the E. coli genome and may
identify points of nucleation for biocrystal formation during stationary phase.
1
Chapter 1: Dps is a universally conserved dual-action DNA-binding and ferritin protein
Chapter appears essentially as published in Orban K, Finkel SE. 2022. Dps is a universally
conserved dual-action DNA-binding and ferritin protein. Journal of Bacteriology
https://doi.org/10.1128/jb.00036-22
2
Introduction
During typical bacterial growth and survival in the laboratory, a bacterial population that
initially exists at low cell density transitions into a phase of rapid growth and cell division known
as logarithmic (log) or exponential phase. As the population approaches high cell density, growth
slows and population density levels out as cells enter stationary phase. The transition from log
phase to stationary phase includes a series of environmental and cellular changes that must be
managed, including lower nutrient availability, increased concentrations of metabolic waste
products, nucleoid remodeling, and managing intracellular concentrations of important cofactors,
including iron. To adapt to the changing environment and stresses of stationary phase, cells modify
their gene expression patterns and protein levels. Curious about this phenomenon, Almirón and
colleagues performed an SDS-PAGE experiment in which newly-synthesized proteins in batch
Escherichia coli cultures were labeled with radioactive methionine at several timepoints during
log phase and stationary phase (1). One of the most highly labeled proteins as cells transitioned
into stationary phase was Dps.
The DNA-binding protein from starved cells, Dps, is conserved across bacterial species
(1–5) (Table 1.1). To date, a UniProt search of genes annotated as “dps” returns 93,962
prokaryotic proteins. Only one bacterial phylum, Thermomicrobia, is not represented in this list;
of note, all gammaproteobacterial orders are represented. Most bacterial genomes contain one dps
gene, but some encode as many as five (5–7). Additionally, some archaeal species have
dodecameric, Dps-like ferritin proteins (8, 9).
In most species, Dps functions as a ferritin, which is an iron detoxifying and storage protein
with ferroxidase activity. However, in some species, Dps is also a double-stranded DNA-binding
protein (Table 1.1). The DNA-binding and ferroxidase activities of Dps, in species with both
3
functions, are biochemically discrete, but function jointly to protect DNA and mediate stress
tolerance (10–13).
Dps binding requires a minimum length of ~90bp of double-stranded DNA (dsDNA) in
species in which it binds DNA, but otherwise has no well-defined DNA sequence binding motifs
or structural specificity (1, 14–18). In many organisms in which Dps functions as a dsDNA-
binding protein, Dps becomes the major nucleoid-associated protein (NAP) during stationary
phase (19–22). Like other NAPs, Dps modulates nucleoid shape and compaction (23).
During stationary phase, Dps and DNA form a tightly-packed nucleoprotein complex
called the biocrystal (1, 2, 24–27). Biocrystal formation is stationary phase-specific, requires Dps
(11, 20, 22, 24, 25, 28–31), and, in E. coli, occurs gradually, beginning in early stationary phase
and continuing until late stationary phase, at which point the nucleoid is organized into the
hexacrystalline array (22, 25).
The ferritin properties of Dps are three-fold. First, it is proposed that Dps stores iron and
releases Fe(II) when needed (32–40). Second, Dps detoxifies excess iron in the cell using its
ferroxidase activity to oxidize soluble Fe(II) to insoluble Fe(III), which is unavailable to participate
in potentially damaging redox chemistry (see below). Third, as a ferroxidase, Dps uses H2O2 to
oxidize ferrous iron to the ferric form, making it unique among ferritins (4, 8, 9, 26, 34–38, 41–
44). Because using H2O2 as the oxidizing agent for iron results in its breakdown, Dps helps prevent
the synthesis of reactive oxidative species (ROS), through the Fenton reaction, capable of
damaging nucleic acids, proteins, and lipids (34).
4
Monomer Structure
Dps was originally discovered in E. coli as a DNA-binding protein (1). When the X-ray
crystal structure was determined, it became apparent that the Dps monomer, comprised primarily
of a 4-helix bundle, shows striking similarity to the ferritin monomer, despite a lack of sequence
homology (Figure 1.1) (18, 32). Ferritins are iron-sequestering proteins and are conserved
throughout all three domains of life. Like ferritin, the E. coli Dps (Ec-Dps) monomer’s A and B
helices are connected by a short loop; its C and D helices are similarly connected (18). The AB
and CD helix pairs are connected by a longer loop, also akin to ferritins (18).
There are some notable differences between Dps and ferritin monomers. The Ec-Dps
monomer has an additional, smaller N-terminal helix (Figure 1.1) (18). This helix is flexible,
making it difficult to characterize using typical X-ray crystallography techniques. Ferritins lack
this N-terminal helix, but they have an additional C-terminal helix (Figure 1.1), which is
postulated to contribute to the 24-mer organization in ferritins, compared to the Dps dodecamer
(36).
5
Dodecamer Structure
Dps monomers of molecular weight ~19 kDa come together to form a dodecamer (1). This
dodecamer, like the monomer, is structurally similar to ferritin oligomers (18). Both Dps and
ferritins oligomerize into a hollow sphere, although ferritins contain 24 subunits compared to 12
for Dps (2, 18, 27, 36, 45, 46). The Dps sphere, smaller than that formed by ferritins, is
approximately 90 Å in diameter with a 45 Å hollow core (9, 18, 27, 36, 45). Ferritins assemble
into 120 Å-diameter spheres with 80 Å hollow cores (Figure 1.2) (46). Symmetry also differs for
the Dps dodecamer compared to the ferritin 24-mer: the Dps dodecamer has tetrahedral symmetry,
in contrast to ferritin’s octahedral symmetry (18, 27, 36, 45–47).
Several assembly models for Dps have been proposed. In one model, the AB loop acts as
a switch for the number of subunits in an oligomer: a rigid AB loop (like that found in Dps) directs
assembly of a dodecamer with 2- and 3-fold symmetry, while a flexible AB loop (like that found
in bacterioferritin, Bfr, and ferritin, Ftn, proteins) directs assembly of a ferritin 24-mer with 2-, 3-
, and 4-fold symmetry (48). In another model, the N-terminus of Dps may modulate Dps
dodecamer formation and self-association during biocrystal formation (12, 49, 50). When
examining the step-wise dynamics of dodecamer assembly, it has been suggested that some Dps
species first form trimers and then dodecamers, whereas others first form dimers and then
dodecamers; the difference roughly correlates with the length of the N-terminal helix (51). It is
also notable that two arginine residues, R83 and R113 in Ec-Dps, have been found to be necessary
for Ec-Dps dodecamer assembly (52).
6
Other Quaternary Structures
While Dps proteins typically function as dodecamers, some Dps proteins can form smaller,
semi-functional oligomers. Both Lactococcus lactis DpsA (Ll-DpsA) and Mycobacterium
smegmatis Dps1 (Ms-Dps1) form stable dimers and trimers in addition to dodecamers (2, 53). The
Ms-Dps1 trimer has ferroxidase activity, though it cannot store iron, and does not bind DNA; the
dodecamer performs all three functions (2, 50, 54). In some cases, formation of non-dodecameric
oligomers is due to environmental conditions. The hexuronates D-glucuronate and D-
galacturonate have been found to destabilize the Ec-Dps dodecamer (55). Notably, in E. coli,
hexuronate concentrations are highest during log phase, when Dps is poorly expressed, and
decrease significantly during stationary phase, when Dps is highly expressed (56). Additionally,
Dr-Dps1 forms DNA-binding dimers at low salt concentration in vitro (12, 14), consistent with
models of Dps-DNA binding sensitivity to cation concentration (see ‘Dps as a DNA binding
protein’ section). Bacillus cereus Dps3 (Bc-Dps3) is found primarily as a dimer, but forms
dodecamers upon addition of Fe(II), suggesting an environmentally mediated mechanism for
dodecamer assembly, akin to Dr-Dps1 (57).
7
Dps as a DNA binding protein
During stationary phase in organisms whose Dps is a DNA-binding protein, Dps and DNA
assemble into a tight nucleoprotein complex called the biocrystal. The biocrystal forms a
hexagonally-packed assembly, with offset planar arrays stacked on top of one another, similar to
oranges packed in a crate (Figure 1.3) (1, 2, 24–27). While Ec-Dps can form hexacrystalline
biocrystal-type structures with itself in vitro, addition of DNA greatly accelerates the process (11,
49). However, the spacing of the Ec-Dps crystalline lattice is similar with or without DNA (11).
Initially during the formation of the biocrystal, Ec-Dps-DNA aggregates form toroidal (ring-
shaped) structures (Figure 1.4) (25). It has been hypothesized that Ec-Dps-DNA toroids are points
of initial nucleation from which biocrystallization spreads, until the nucleoid is largely restructured
by Ec-Dps. Spacing of nucleoprotein complexes in early stationary phase toroids and late
stationary phase biocrystal is similar, supporting this hypothesis (25).
Dps-DNA complexes are only formed during stationary phase in E. coli (Figure 1.4) (11,
24, 28, 58), where one study has found Ec-Dps occupies over 50% of the nucleoid (22). This
growth phase-specific phenomenon is not due simply to increased Dps abundance during
stationary phase, because overexpression of Ec-Dps during log phase does not result in biocrystal
formation (24). Whether this stationary phase specificity is due to unfavorable environmental
conditions for Dps binding during log phase or reflects the concentration of other competing NAPs
with higher DNA affinity during log phase is not well-understood.
Stationary phase-specific nucleoid compaction is dependent on Dps (11, 20, 22, 24, 25,
28–31). To date, research suggests Ec-Dps is unique among NAPs in that it is necessary for proper
nucleoid structure. Removal of any other NAP during log phase does not significantly change the
log phase nucleoid structure, but removal of Ec-Dps significantly changes the structure of the
8
stationary phase nucleoid (59, 60). Without Dps, the late stationary phase nucleoid in E. coli is
configured into a cholesteric (liquid crystalline) phase (24). While cholesteric organization has
been shown to reduce the accessibility of DNA molecules to a variety of damaging factors,
suggesting an overarching necessity for DNA protection during late stationary phase, this nucleoid
conformation also leads to a longer lag phase and other defects (61, 62). This is likely due to DNA
being in a conformation that is more resistant to the remodeling to log phase chromatin structure
that occurs upon inoculation into fresh medium.
Dps-dependent nucleoid compaction during the transition from log phase to stationary
phase is gradual (22, 25). Nucleoid restructuring by Ec-Dps lags behind Ec-Dps binding (63, 64).
This could explain the timing of the shift from toroids to mature biocrystal between early and late
stationary phase. Hysteresis, the phenomenon of a physical change lagging behind its inducing
effect, appears to be a feature of the stationary phase nucleoid, as nucleoid restructuring by
Integration host factor (IHF), the other major NAP during early stationary phase, also displays this
property (64).
Not all Dps proteins with DNA-binding activity create a biocrystal. Some form non-
crystalline aggregates, whereas others bind DNA without condensation (15, 53, 54). For example,
the DNA-binding activity of Helicobacter pylori neutrophil-activating protein (HP-NAP), which
is unique among Dps proteins for its positively charged exterior, is modulated by pH (15). Ms-
Dps1, which does not induce DNA condensation and does not protect DNA from nuclease-induced
cleavage, does protect DNA from hydroxyl radical-induced damage (54). However,
overexpression of either Ms-Dps protein in M. smegmatis results in nucleoprotein toroid formation
(31).
9
The DNA-binding activity of several Dps species is modulated by environmental factors
in vitro. While the DNA-binding and ferritin-like properties of Dps proteins are biochemically
separable, incorporated Fe
3+
enhances DNA-binding efficiency in some species (12, 65). In
Staphylococcus aureus, stationary phase alone is not sufficient for nucleoid compaction, but
oxidative stress or overexpression of MrgA, the S. aureus Dps protein, results in a highly
compacted nucleoid (30, 66). Additionally, treatment of the M. smegmatis stationary phase
nucleoid with RNase “loosens” nucleoid structure (31). RNase-induced loosening may be due to
decreased macromolecular crowding forces, with fewer macromolecules (RNA, DNA, and
proteins) in the cytoplasm to promote nucleoid condensation (67).
Dps has no currently identified sequence or structural specificity for its DNA binding
activity. Ec-Dps binds DNA with a Kd of approximately 2x10
-7
M, which is relatively low for a
NAP (Kd for specific binding of other major NAPs typically falls around 10
-9
M (68–72)) and may
explain the lack of observed sequence specificity (16, 18). It has been observed that Ec-Dps does
not discriminate among linear dsDNA, circular dsDNA, and ssRNA in vitro (11, 73). Furthermore,
Ec-Dps shows no preference between supercoiled and relaxed DNA (14). However, there appears
to be a minimum size for high-affinity DNA binding: dsDNA fragments smaller than ~90 bp do
not bind Ec-Dps efficiently (18). Interestingly, 90 bp is about the length required to encircle a Dps
dodecamer, a number that has been used to suggest a wrapping model for Dps-DNA binding (74).
While there is no universally accepted model for Dps DNA-binding specificity, several
studies have suggested DNA-binding motifs. One RNA-seq study suggested that Dps binding
regions are enriched for inverted repeats, overlap significantly with promoter islands, tend to show
increased structural flexibility, or overlap with binding sites of other NAPs, particularly the Factor
for inversion stimulation (Fis), one of the major log phase NAPs (17, 75). A SELEX-Seq
10
experiment suggested a consensus sequence for linear DNA binding by Dps (76), but the
universality of this sequence in vivo has yet to be confirmed.
There are 2 non-mutually exclusive models proposed for DNA-Dps interactions. The first
involves interaction mediated through divalent cation bridges. This is supported by the observation
that Dps will only bind DNA in a certain cationic range of ~1.0 mM Mg
2+
in vitro, which is
abolished when EDTA is added to sequester cations (24, 53, 64). The second is that the short,
lysine (K)-rich, N-terminal helices from three adjacent Dps dodecamers coalesce around a dsDNA
molecule (Figure 1.3) (10, 18, 49). Species with a K-rich N-terminal helix, including Ec-Dps, tend
to also have DNA-binding activity, and species that lack a K-rich N-terminal helix tend to lack
this activity (Table 1.1) (14, 26, 45, 65, 77, 78). This second model does have exceptions: Ms-
Dps1 and Dr-Dps1 appear to require both the N-terminal and C-terminal regions to bind DNA,
and HP-NAP is postulated to use its positively-charged exterior to bind DNA (15, 50, 79, 80).
Additionally, Agrobacterium tumefaciens Dps (At-Dps) has a positively-charged N-terminal helix,
though it is 11 amino acids shorter than the 20 amino acid long Ec-Dps N-terminal helix, and does
not bind DNA (36). The convergent evolution of different DNA-binding modes exhibited by
different Dps species suggests a biological demand for DNA protection during stationary phase.
11
Interactions with other NAPs
Log phase is characterized by rapid growth and cell division, which necessitates high levels
of tightly regulated gene expression and thus ready access to the chromosome. This is reflected by
the plurality of major NAPs during log phase and the resultant log phase nucleoid structure.
However, easy access to genetic material is not necessarily beneficial during stationary phase. Low
nutrient availability and high population density shifts the cell’s focus to maintenance and
protection; this requires Dps-dependent sequestration of DNA. The transition from log phase to
stationary phase nucleoid structure requires transition in NAP availability and perhaps particular
interactions between log phase and stationary phase NAPs.
Dps interacts in various ways with other NAPs. Ec-Dps acts antagonistically to Fis as a
nucleoid structural agent (58, 63, 75). Additionally, Dps and Fis expression has been found to
invert between log and stationary phase in E. coli: Fis is highly expressed during log phase and
below the limit of detection during stationary phase, while Dps is weakly expressed during log
phase and highly expressed during stationary phase (Figure 1.5) (19). Fis and H-NS each regulate
Ec-Dps expression (81). Both Dps and Curved DNA binding protein A (CbpA), the two major
NAPs present during late stationary phase in E. coli, self-aggregate and are postulated to cause
nucleoid compaction by clustering distal DNA loci (82).
The interactions between Dps and some of the major log phase NAPs may suggest a
mechanism for Dps accessing the chromosome throughout stationary phase. Perhaps Dps replaces
other NAPs, including Fis, as they dissociate from the chromosome. If log phase NAPs dissociate
in a concentration-dependent manner, as previously reported (83, 84), this may provide the
opportunity for Dps molecules to nucleate DNA locally before restructuring the entire
chromosome into the biocrystal.
12
Dps as a ferritin
Iron, a cofactor in many essential biological processes, can be critical in the generation of
reactive oxidative species (ROS), which are capable of damaging a broad range of macromolecules
(85). This is achieved primarily through the Fenton reaction:
Fe
2+
+ H2O2 → [FeO]
2+
+ H2O
[FeO]
2+
+ H
+
→ Fe
3+
+ HO•
Ferritins help the cell manage the dual nature of iron by converting it into its insoluble, less
reactive form, storing this detoxified Fe(III), and releasing it when needed. Bacteria have two
highly conserved ferritin proteins: ferritin (Ftn) and the heme-containing bacterioferritin (Bfr). Ftn
and Bfr have little sequence homology, except at their ferroxidase sites, which are highly
conserved (86). Dps, while ferritin-like, has a number of notable differences when compared to
both canonical ferritins (86). Only a few Dps proteins, including the two Dps proteins of L. lactis,
have been found not to possess ferritin activity; both of these proteins bind DNA (53). Dps uses
H2O2 to oxidize Fe(II) to Fe(III), while traditional ferritins use O2 as the oxidant (4, 6, 34, 41). One
H2O2 is metabolized for every two Fe(II) atoms oxidized, avoiding production of hydroxyl radicals
via Fenton chemistry (34). For some species with two or more Dps proteins, one may use O2 to
oxidize Fe(II) and the other may use H2O2 (39, 87–89).
Dps dodecamers have iron entry pores that are unique among ferritins. These pores are
negatively charged and located at the four 3-fold interfaces within the dodecamer (6, 18, 27).
Interestingly, Dr-Dps1 has distinct iron exit channels that constantly release Fe(II) and contribute
to DNA damage in vitro (79). When these iron exit channels are disrupted via mutation, Dr-Dps1
loses the ability to contribute to DNA damage, and instead DNA is partially protected from iron-
mediated cleavage (79).
13
The Dps dodecamer has 12 highly conserved ferroxidase centers, which can each oxidize
two iron atoms simultaneously (34, 79, 88–91). Unlike ferritins, the Dps ferroxidase site is
comprised of residues from two adjacent monomers: two histidine residues from one subunit and
an aspartate and glutamate from the other (6, 27, 41, 45, 51, 86, 92). This is distinct from ferritins,
whose active sites are formed solely within each of the 24 monomers.
Once oxidized, iron is stored in the hollow core of the Dps dodecamer, at which point it
organizes as microcrystals (92). While Dps can form a ferric core with O2 as oxidant similar in
size to that observed with H2O2, O2 core formation reaction is less cooperative and leads to
increased heterogeneity in ferric core size in a population of Dps dodecamers (34). A crystalline
iron core is also observed in ferritin proteins (93). Two steps in Fe(III) reduction and release from
Ec-Dps have been observed in biochemical experiments, which may indicate two populations of
iron in the protein – perhaps one representing the bulk iron in the core and the other attached to
the interior of the shell (92).
One Dps dodecamer typically contains up to 500 Fe(III) atoms in its internal cavity under
aerobic culture conditions (32–40). When grown anaerobically, the ferric core contains ~400
Fe(III); traditional ferritins can store up to ~4000 oxidized iron atoms per 24-mer (34, 46, 94).
Several Dps species store fewer iron atoms: Thermosynechococcus elongatus Dps (Te-Dps) and
Halobacterium salinarum DpsA (Hs-DpsA) each hold ~100 Fe(III)/dodecamer, and
Trichodesmium erythraeum Dps (Te-Dps) and Sulfolobus solfataricus Dps (Ss-Dps) each store
~300 Fe(III)/dodecamer (42, 95, 96) (Table 1.1). The biological mechanism(s) for these
discrepancies is unknown.
Some Dps proteins have been found to bind other metals, including zinc, calcium, cobalt,
copper, nickel, manganese, terbium; and one small charged molecule, phosphate (9, 14, 79, 87, 89,
14
95, 97–101). Each has been found to bind at the ferroxidase site and/or an allosteric (non-
ferroxidase) site, depending on the Dps species and non-iron substrate (Table 1.1). Dr-Dps-1
contains 2 allosteric Co(II) sites, one near the C-terminus and another near the N-terminus of the
protein, as well as an allosteric Zn(II) site in its longer-than-typical Dps N-terminal helix (79, 99).
This can alter Dps action: for example, Zn(II) inhibits ferroxidase activity when bound to the iron
site in Nostoc punctiforme Dps4 (Np-Dps4) (89). Additionally, it has been reported that phosphate
can affect the crystallinity and chemical reactivity of ferritin cores, which may be due to
interactions between the negatively-charged phosphate and positively-charged iron ions,
suggesting this molecule may serve to modulate these properties in the Dps core (102). Finally,
Dps has been shown to protect Anabaena PCC 7120 and E. coli from copper toxicity (Table 1.2),
contributing further to a physiological role for Dps of binding non-iron metals.
15
Dps Expression
Expression of Dps is primarily dependent on growth phase. During log phase, dps
expression is low; however, it is upregulated in response to oxidative stress (7, 66, 103–105). dps
expression is upregulated during stationary phase relative to log phase (1, 19, 21, 104, 106).
Dps protein concentration is relatively low during log phase. Several studies have
quantified Ec-Dps concentration as <1000 molecules per cell during log phase (19, 21, 106). Dps
expression is controlled at the transcriptional, post-transcriptional, translational, and post-
translational levels during log phase. Transcription is repressed by ferric uptake regulation protein
(Fur) (107–110). During log phase, Ec-Dps is induced in response to a variety of stresses (3). This
is not due to growth phase-specific changes in the transcription machinery: E. coli s
70
, the
housekeeping sigma factor, and s
s
, the stationary phase sigma factor, have similar affinities for
the dps promoter in vitro (81). Instead, various stress-activated transcription factors modulate dps
expression. OxyR, the oxidative stress response transcription factor, activates dps expression
during log phase by binding upstream of the dps promoter and recruiting s
70
(105, 111, 112). This
is modulated by the oxidative stress level encountered by the cell. Reduced OxyR has significantly
lower affinity for the dps promoter; i.e. oxidized OxyR induces dps expression (113). In addition
to OxyR, dps expression is regulated by PerR, the peroxide regulon repressor (7, 66, 103, 114).
dps is upregulated in response to iron depletion stress, iron excess stress, thermal stress, NaCl
stress, ethanol stress, gamma irradiation, and in the presence of acetyl phosphate (42, 57, 96, 108,
115–120). The log phase stressor concentration-dependent expression of Dps is similar to ferritin
expression, which is low unless stressors are added (121).
Dps transcription is controlled by other NAPs during log phase. In E. coli, Fis inhibits dps
promoter accessibility by RNAP formed with s
70
; Fis and s
70
are able to co-repress transcription
16
by s
s
(81). H-NS binds the -10 promoter region of dps, blocking s
70
from binding the promoter
(81). Dps, which is an N-end rule degradation pathway substrate in E. coli, is rapidly degraded by
ClpXP (122–125). During log phase, the Dps protein’s half-life is ~10 minutes; this increases to
~40 minutes with the addition of oxidative stress (122).
Dps is highly expressed during stationary phase, where several studies have quantified Dps
levels in the range of hundreds of thousands of molecules per cell (19, 21, 106). As in log phase,
Dps levels are controlled at the transcriptional, post-transcriptional, translational, and post-
translational levels during stationary phase. dps transcription is induced by s
s
, which directly
activates dps expression by binding the -10 promoter region (1, 3, 111, 126). dps transcription is
also controlled by other NAPs during stationary phase. IHF, the other major NAP during early
stationary phase, has been found to cooperate with s
s
in s
s
-mediated upregulation of Dps during
early stationary phase (58, 111). At the post-transcriptional level, Dps degradation is not detected
during stationary phase (122). In species with more than one dps gene, different dps loci are
differentially regulated (7, 57, 127–131). Additionally, if one dps locus is knocked out in species
with more than one dps gene, expression of other dps loci may compensate for its absence (128).
Because Dps is a large family of proteins across many species, there are exceptions to these
patterns. In Campylobacter jejuni, dps (Cj-dps) is constitutively expressed during log and
stationary phase, and is not upregulated in response to oxidative stress (37). Additionally, Borrelia
burgdorferi Dps (Bb-Dps) is constitutively synthesized in both log and stationary phase with no
change due to oxidative stress, but is differentially expressed when incubated in mice (low
expression) or ticks (high expression) (40). Finally, Porphyromonas gingivalis dps (Pg-dps)
expression is not modulated by oxidative stress (132).
17
Dps as a Regulator of Gene Expression
One mechanism by which NAPs can affect gene expression is through altering nucleoid
architecture (133). Like the other major nucleoid structural proteins, which affect gene expression
through altering nucleoid structure, Ec-Dps is distributed throughout the nucleoid (1, 63, 134).
Data also suggests Dps may be a regulator of gene expression. When first identified, radiolabeled
2D-PAGE showed many differences between newly-synthesized proteins in E. coli wild-type and
dps-null strains in late stationary phase, suggesting a role for Dps as a regulator of stationary phase-
specific gene expression (1). In addition, a series of promoter-less lacZ fusions made in an
arabinose-inducible dps background show differential expression depending on dps expression
status (SEF, unpublished results). Further, a SELEX-Seq experiment identified 624 Dps binding
sites throughout the E. coli chromosome (76); from the locations of these sites, regulatory targets
have been predicted (https://shigen.nig.ac.jp/ecoli/tec/). Dps in Salmonella enterica and Anabaena
sp. PCC7120 has been shown to affect global gene expression, though it is unknown whether the
modulation effect is direct or indirect (135, 136). However, there are some conflicting data with
respect to gene expression regulation. Antipov and colleagues (75) showed differential regulation
of genes by Dps between biological replicates in late stationary phase E. coli. These results may
suggest regulatory plasticity modulated by Dps. However, an extensive regulatory study on Ec-
Dps found no significant differences in expression via RNA-Seq or proteomics due to the presence
or absence of Dps in log, stationary, or early long-term stationary phase (137).
There are several potential explanations for differing Dps gene expression results,
particularly across E. coli studies. First, methodologies differ. Genetic experiments such as
promoter-less lacZ fusions, which examine larger-scale, population-level effects, may show
different phenomena than sequencing experiments such as ChIP-Seq and mRNA-Seq, which
18
examine changes at the molecular level. Similarly, biochemical experiments such as radiolabeled
2D-PAGE and mass spectrometry may show different phenomena than one another, as the former
examines newly-synthesized proteins and the latter probes global protein distribution. Second, the
strains being studied differ. The same mutation can produce a range of phenotypic changes, or no
change at all, depending on the genetic background (138). This may be the case with Dps gene
expression effects.
19
Macromolecular sequestration and phase separation
Formation of the Dps-DNA biocrystal may be one example of a larger biological
mechanism of protection. Ferritins also form crystalline assemblies, potentially indicating an
evolutionary pressure to promote biocrystallization processes as a stress response (139). Ferritin
crystals form when ferritin is overproduced in E. coli and cells are exposed to Fe(II), a potent
source of oxidative stress that can damage macromolecules like DNA and proteins (61). This may
indicate an evolutionary advantage in structural motifs that facilitate a transition of proteins into
crystalline structures that protect cellular components through the rapid sequestration of valuable
macromolecules from damaging agents. Additionally, biocrystallization has been suggested as a
means to maintain homeostasis in stressful environments. For example, the nucleoids of dormant
Bacillus spores are arranged into a SspC (small, acid-soluble spore protein C)-DNA crystalline
lattice via toroid-mediated condensation, similar to the early stationary phase-specific nucleoid
packaging mediated by Dps, though the packing of toroids and crystalline lattice formation in the
spore is different than that mediated by Dps (140, 141). Because sporulation is induced in response
to various stresses, the nucleoid repackaging occurring during this time supports the hypothesis of
biocrystallization as a stress-response mechanism.
The crystalline assembly of Dps-DNA complexes has been posited to create a distinct
phase within the heterogeneous mixture of the nucleoid, a potentially important example of the
role of phase separation in biological systems (142, 143). Early studies of Dps-DNA complexes in
E. coli show crystalline assembly (11, 24, 25), though the mechanism could be due to either liquid-
or solid-phase separation. Additionally, Dps exhibits highly cooperative binding, which is
emblematic of phase separation (143, 144). Finally, RNA polymerase can access DNA when
bound by Dps (137), but nucleases and other DNA damaging agents cannot as efficiently (1, 137).
20
This observation is consistent with other phase separated complexes, which can selectively
concentrate enzymes and other factors (143). Just as Dps is highly expressed in stationary phase,
phase separation is dependent on high concentrations of the proteins involved in the phase (143).
Phase separation has been shown to play an important role in managing stress response, including
thermal and pH stress (145). This is notable when considering Dps-mediated phase separation, as
Dps is a known contributor to stress responses (see below).
21
Stress Response
Dps confers resistance to several environmental stresses (Table 1.2), the most extensively
studied of which is oxidative stress (1, 11, 30, 57, 62, 146, 147). It does so in three notable ways:
(1) physical protection of DNA, (2) sequestering of iron, and (3) neutralization of H2O2. First, Dps
specifically protects DNA from oxidative stress-induced damage (10, 11, 57, 91, 147). This is akin
to eukaryotic histone proteins, which physically protect DNA from oxidative stress (148).
Moreover, Ec-Dps has been found to interact with DnaA to impede DNA replication initiation
during log phase in periods of oxidative stress, suggesting a secondary regulatory role that is
targeted at protecting DNA (149). Second, in its role as a ferritin, Dps protects the cell from
oxidative stress by sequestering iron and thus preventing formation of ROS. In addition to its
physical sequestration of iron, Ec-Dps has also been found to interact with the iron-sulfur cluster
protein YtfE to diminish YtfE-induced oxidative stress (150). This demonstrates a secondary
regulatory role that is targeted toward reducing oxidative stress, which is a similar theme to how
Dps protects DNA. The third way in which Dps protects the cell from oxidative stress is through
detoxification of H2O2. This is inherent in its ferritin function, as this protein’s preferred oxidant
for iron is H2O2. Dps also confers resistance to other stressors, though the protective effects differ
depending on growth phase and species (Table 1.2).
In species with more than one dps gene, each gene may confer differential resistance to
different stresses (7, 128, 129). In Bacillus cereus, two of its three Dps proteins (Bc-Dps1 and Bc-
Dps2) act cooperatively to confer resistance to oxidative stress (7). In E. coli, both the ferritin and
DNA-binding properties of Dps are required for full Dps-dependent DNA protection (10, 11). Ec-
Dps significantly reduces the number of DNA strand breaks, abasic sites and ruptured/oxidized
22
guanine, and GC->TA + TA->AT base mutations (147). This is due to Dps DNA protection, as
the protein is not involved in the repair of oxidatively damaged DNA (147).
23
Other functions
In addition to its ferritin and DNA-binding abilities, Dps has been identified in other
important cellular functions. Dps has been identified in a screen for genes involved in biofilm
formation, though no specific role has been classified, and one study of spontaneously occurring
phage-tolerant E. coli identified Dps at the outer membrane, which may implicate Dps in an as-yet
undetermined role in each of these processes (3, 151–154). Dps has also been implicated in
virulence (40, 77, 109, 126, 152, 155–163). This makes particular sense since iron acquisition can
play a vital role in determining pathogenicity (164). Salt sensitivity, which is modulated by Dps
(Table 1.2), is also highly correlated with virulence. The Dps protein from Helicobacter pylori,
NAP, has been shown to impair human iron absorption and target iron to H. pylori during infection
(165). Another Dps protein, Microbacterium arborescens amino acid hydrolase (AAH), catalyzes
the cleavage and formation of amide bonds (78). Additionally, overexpression of Ec-Dps has been
found to impede colony growth on agar plates by two- to three-fold during log phase (149).
In some species, Dps has been found at the outer membrane, though the function of this
localization is currently unclear (3, 151, 153, 154). In Synechococcus sp. Strain PCC 7942, more
DpsA is observed at the inner cell membrane during lag phase and log phase than during stationary
phase (166). This makes sense in the context of stationary phase-specific DNA binding, as more
Dps should be observed in the nucleoid at that time, leaving fewer proteins available to participate
in their outer membrane function(s). This, of course, assumes Dps is able to move between the
inner membrane and the nucleoid with relative freedom. The same study found DpsA localized at
the cell membrane and the nucleoid (166). The authors of that study proposed that there are two
“pools” of DpsA that function in Synechococcus: an insoluble, DNA-binding fraction at the
nucleoid and a soluble, ferritin-active fraction at the membrane (166). Two “pools” of a Dps
24
protein have been observed in D. radiodurans: Dr-Dps2, which can function either as the full-
length gene product or as a truncated form (lacking the nonpolar portion of the N-terminus that
protrudes past the positively-charged portion), is found full-length at the membrane and in the
truncated form in the nucleoid (80). Currently, the functional differences of the full-length and
truncated forms of Dr-Dps2 are unclear; localization to different cellular components suggest
broad functional differences.
25
Dps in nanotechnology
Recently, Dps proteins, like ferritins, have been used for a range of nanotechnology
applications. The “hollow ball” structure of Dps and ferritin proteins make them excellent
candidates for nanotechnologies that require protein cages. It is advantageous to use Dps instead
of canonical, 24-mer ferritins for several reasons. First, Dps is smaller, making it a better option
when smaller size is desirable (167, 168). Second, Dps is highly thermostable, making it easy to
purify and often more durable (168).
Dps and ferritins have primarily been used in materials science and drug development and
delivery. In materials science, horse spleen ferritin has been used as nanoreactors (169), Ec-Dps
has been used as a scaffold for nanodevice assembly (168), and Bacillus subtilis Dps (Bs-Dps) and
Listeria innocua Dps (Li-Dps) have been used as catalysts for the formation of carbon nanotubes
with limited diameter distribution (170) and for platinum nanocluster formation for hydrogen
production (171), respectively. Ec-Dps has also been used as a platform to experimentally
reconstitute protein-protein interfaces (172), and Li-Dps has been used to synthesize CdSe
nanoparticles with nanometric gaps (173) and to fabricate a “high-density, periodic silicon-
nanodisc (Si-ND) array” for use in silicon quantum dot solar cells (174). In drug development and
delivery, horse spleen and human ferritins have been used as platforms for antigen presentation
(175), vaccine development (175), cancer immunotherapy development (176), drug delivery (177),
and MRI contrast agent (177). In recent years, there has been increasing interest in nanotechologies
for use in a range of applications, such as tissue repair, drug delivery, and immunoassays (178).
Combined with its smaller size and high thermostability, these demonstrated uses make Dps
increasingly valuable in nanotechnology.
26
Future directions
While much progress has been made in the understanding of the ferritin properties of Dps,
there is much more to be learned. The dynamics of Dps ferric core organization are not well
understood. If two iron sub-populations are present in the protein, it is important to understand the
division between, dynamism within, and biological relevance of those populations. Further, the
determining factors behind which Dps species bind other ions, metals, or small charged molecules
awaits clarification.
In addition to ferritin activity, the dynamics of Dps dodecamer formation are still poorly
understood. While data suggest that certain Dps species form dimers and/or trimers before the
dodecamer forms, a more thorough inquiry is required to better understand these dynamics. The
driving force behind the stable dimers/trimers and dodecamers formed by some Dps proteins may
be due to pH or salt concentration, as previously suggested (12, 14, 54). If that is the case, however,
a comparative study of those Dps proteins that form stable dimers/trimers and those that do not is
warranted to distinguish mechanisms of assembly.
The additional functions of Dps present interesting avenues of experimentation. Several
studies have found Dps at the outer membrane. However, it is still unclear exactly why Dps is
localized there. If Dps exists in substantial quantity in the membrane, any movement or changes
in quantity or concentration may indicate its function there, whether ferritin, stress response, DNA-
binding, or an additional as yet undetermined activity. Examining membrane composition and
permeability in dps-null strains, dps overexpressing strains, as well as mutants for oligomerization,
DNA-binding, and ferritin activity could shed further light on why Dps exists at the outer
membrane.
27
When compared to log phase, little is known about the stationary phase nucleoid. This
includes its structure, dynamics, and protein composition. More specifically, Dps-dependent
nucleoid compaction during stationary phase may be due to cellular environmental conditions.
Another model for this phenomenon is that Dps binds DNA when DNA is available to it, which is
more likely after log phase when there are lower concentrations of other NAPs with which to
compete. Further, it is important to explore how the biocrystal forms. It is unknown what stimulates
the formation of biocrystal precursor toroids – might this result from log phase NAPs dissociating
from DNA? It is also unknown whether the biocrystal forms at programmed chromosomal loci or
if the process is more stochastic. Perhaps there is an intermediate mechanism, by which
preliminary, local nucleoid restructuring by Dps occurs when log phase NAPs dissociate from the
chromosome, and the secondary, global restructuring occurs in a programmed manner. If Dps truly
has no discernable sequence or structural specificity, but the biocrystal forms reproducibly at
certain loci, how does Dps get directed to the sites it binds?
There is conflicting evidence with regard to Dps as a regulator of gene expression. Perhaps
Dps does affect gene expression as suggested by Almirón and colleagues’ 2D gel electrophoresis
studies (1), but the stationary phase intracellular environment is such that these changes cannot be
detected by transcriptomic or proteomic techniques that focus on the global mRNA/protein
population. If both gene expression, as well as RNA and protein degradation, slow during
stationary phase, a larger relative shift in expression profiles may be necessary to outweigh the
baseline from log phase and detect these phenomena during stationary phase. Alternatively, a post-
transcriptional mechanism of gene expression regulation may be yet-undiscovered; direct
interaction between Dps and mRNA may explain the biological relevance of the ssRNA-binding
ability of Dps.
28
Until recently, the role of post-translational modifications (PTMs) on NAPs had not been
studied in bacteria (179). This is still largely the case for Dps proteins and stationary phase. While
it has been shown that S. enterica Dps can be glycosylated (180) and Dr-Dps2 can have its N-
terminus cleaved in vivo (80), it would be interesting to further study how and where Dps acquires
PTMs and what the effects are, if any. The activity of Fis, to which Dps acts antagonistically,
seems to be less subject to alterations by PTM than other log phase NAPs. It has been hypothesized
that this is due to Fis activity being more dependent on growth phase than those other NAPs, so
PTMs would potentially be a redundant signal here (179). Because Dps activity is also highly
regulated by growth phase, PTMs may not act as frequently on Dps as other NAPs. The results of
a study of Dps PTMs may help bolster the Fis hypothesis or shed light on another factor in play.
29
Conclusion
During stationary phase, the cell encounters an environment in which nutrient availability
is more limited than during log phase. The cell has a biological imperative during this time to
protect its genetic information from damaging agents, including ROS-inducing ferrous iron ions.
Dps provides an elegant solution to this problem, both sequestering iron in its inner cavity and
creating a phase separated nucleoid that is less accessible to DNA damaging agents. While
stationary phase-specific nucleoid compaction is surely impacted by the action of other NAPs, it
requires Dps. This is likely because the cell needs a rapid switch to adapt its nucleoid to the
pressures of stationary phase. In its role as a major NAP throughout stationary phase, Dps offers
DNA protection against damaging agents and accessibility to “trusted” DNA-binding proteins such
as RNA polymerase.
Dps is a highly conserved bacterial ferritin and NAP. It has unique ferritin and DNA
binding properties that make it not only interesting to study from a basic biological standpoint, but
also increasingly important in the development of nanotechnologies and in drug delivery. Dps is
involved in conferring resistance to a myriad of stresses. Whether Dps functions as a direct
transcriptional regulator is not clearly understood; however, it is involved in regulating gene
expression, even if indirectly through nucleoid restructuring. Additional studies will prove useful
to understand the dynamics of the ferritin core of Dps; the process of Dps dodecamer formation;
the structure, dynamics, and composition of the stationary phase nucleoid; gene expression
regulation by Dps; PTMs on Dps and other NAPs during stationary phase; and other functions of
Dps, particularly as they pertain to virulence, phage resistance, biofilm formation, and presence in
the cell membrane.
30
Figures
Figure 1.1: Dps monomer structure displays homology to ferritins
A) tertiary structure of Ec-Dps (PDB 1DPS), B) Ec-FtnA (PDB 1EUM), and C) Ec-Bfr (PDB
3E1J) (18, 181, 182). Homologous alpha helices are displayed in the same color: helix A, red;
helix B, blue; helix C, yellow; and helix D, orange. Non-homologous helices shown in white. N-
termini and C-termini for each molecule are labeled with white ‘N’ and white ‘C’, respectively.
Nipsum Nipsum Nipsum
Cipsum
Cipsum
Cipsum
Aipsum Bipsum C
31
Figure 1.2: Quaternary structure of E. coli Dps dodecamer and ferritin 24-mers
A) Ec-Dps dodecamer (PDB 1DPS), B) Ec-FtnA 24-mer (PDB 1EUM), and C) Ec-Bfr 24-mer
(PDB 3E1J) (18, 181, 182). Monomer subunits shown in distinct colors. Ec-FtnA is a partial
structure; mirrored subunits are shown in the same color. Images scaled to approximate size
difference between the 90 Å-diameter Dps dodecamer and the 180 Å-diameter ferritin 24-mers.
Created with BioRender.com.
32
Figure 1.3: Model of E. coli Dps-DNA binding
Model of Ec-Dps DNA binding, where blue shapes represent Dps dodecamers, and yellow double
helices represent dsDNA molecules. A) Single Dps dodecamer is separate from dsDNA, B) triad
of Dps dodecamers coalesce around a single dsDNA molecule, C) multiple Dps dodecamer triads
coalesce around dsDNA, D) 3-dimensional Dps-DNA hexacrystalline array. Model is inspired by
Grant and colleagues (18). Not to scale. Created with BioRender.com.
33
Figure 1.4: E. coli Dps induces formation of the stationary phase-specific biocrystal
E. coli nucleoid structure bounded within cell membranes, where brown circles represent
ribosomes, yellow shapes represent DNA, and blue circles represent Dps dodecamers. A) During
log phase, chromatin is interspersed with translation machinery. B) During the transition between
log phase and stationary phase, toroids composed of regularly spaced Dps-DNA nucleoprotein
complexes form, which are segregated from ribosomes. C) By late stationary phase, the nucleoid
has been restructured to a regularly spaced “biocrystal” nucleoprotein complex which is segregated
from ribosomes. Double helices in C represent locally parallel DNA within the crystalline
nucleoid. Model is inspired by Frenkiel-Krispin and colleagues (24, 25). Not to scale. Created with
BioRender.com.
34
Figure 1.5: Major E. coli NAP abundance by growth phase
NAPs are differentially expressed during different growth phases. Dps (red) and IHF (purple)
protein abundance is low during log phase and high during stationary phase; Fis (black), H-NS
(dark gray), HU (light gray), and Hfq (medium gray) are highly expressed during log phase and
lowly expressed during stationary phase; CbpA (blue) is lowly expressed until mid- to late
stationary phase. Graph is inspired by Ali Azam et al. (19). X-axis is not linear with time.
0
25
50
75
100
log early
stationary
late
stationary
growth phase
percent of maximum abundance
Fis
Dps
IHF
CbpA
H-NS
HU
Hfq
35
Other ion binding site
Ferroxidase center (14)
C-term, N-term Co(II)
site (79)
N-term Zn(II) site (99)
Other ions
bound
Ca(II) (14)
Co(II) (79)
Mn(II) (184)
Zn(II) (99)
Mn(II) (184)
K-rich N
term
Y
N
N
N
N
N
N
Y
N
Y
N
Y
N
Y
Y
N
DNA-
binding
N (36)
N (183)
N (183)
Y (57)
Y (146)
N (40)
Y (91)
Y (14)
Y (14)
Y (184)
Y (1)
N (33)
Y (53)
Y (53)
N (118)
Iron core
(aerobic)
500 (36)
500 (39)
500 (39)
500 (40)
500 (37)
250 (184)
400 (184)
500 (34)
500 (33)
Preferred
oxidant
O 2 (39)
H 2O 2 (39)
H 2O 2 (37)
H 2O 2 (184)
H 2O 2 (184)
H 2O 2 (34)
H 2O 2 (38)
Ferroxi-dase
Y (36)
Y (39)
Y (39)
Y (40)
Y (37)
Y (79)
Y (14)
Y (185)
Y (34)
Y (33)
N (53)
N (53)
N (53)
N (53)
Y (186)
Iron
storage
Y (36)
Y (183)
Y (183)
Y (7)
Y (7)
Y (57)
Y (40)
Y (37)
Y (79)
Y (185)
Y (34)
Y (33)
Y (186)
Oligomer
Dodecamer (36)
Dodecamer (183)
Dodecamer (183)
Dodecamer (7)
Dodecamer (7)
Dodecamer (57)
Dimer (57)
Dodecamer (146)
Dodecamer (40)
Dodecamer (37)
Dodecamer (79)
Trimer (184)
Dimer (14)
Dodecamer (185)
Dodecamer (1)
Dodecamer (157)
Dodecamer (53)
Trimer (53)
Dimer (53)
Dodecamer (53)
Dodecamer (186)
Gene
dps
dps1
dps2
dps1
dps2
dps3
mrgA
dps
dps
dps1
dps2
dps
nap
dpsA
dpsB
dps
Species
Agrobacterium
tumefaciens
Bacillus anthracis
Bacillus cereus
Bacillus subtilis
Borrelia burgdorferi
Campylobacter jejuni
Deinococcus
radiodurans
Escherichia coli
Helicobacter pylori
Lactococcus lactis
Listeria innocua
Tables
Table 1.1: Key Dps properties in several bacterial species
36
Other ion binding site
Ferroxidase center (89)
Ferroxidase center (100)
Ferroxidase center (100)
Ferroxidase center (100)
Ferroxidase center (100)
Ferroxidase center (97)
Novel Zn(II) site (97)
Ferroxidase center (97)
Ferroxidase center (87)
Other ions
bound
Zn(II) (89)
Co(II) (100)
Cu(II) (100)
Mn(II) (100)
Ni(II) (100)
Zn(II) (97)
Tb(II) (97)
Zn(II) (9)
Zn(II) (87)
K-rich N
term
N
N
N
N
N
Y
Y
N
N
Y
N
N
N
N
Y
Y
N
N
DNA-
binding
N (117)
Y (187)
N (2)
Y (27)
N (189)
Y (189)
Y (189)
Y (190)
Y (190)
Y (40)
Y (66)
N (193)
N (193)
N (4)
Iron core
(aerobic)
300 (42)
100 (96)
Preferred
oxidant
H 2O 2 (117)
H 2O 2 (189)
H 2O 2 (189)
H 2O 2 (189)
O 2 (89)
H 2O 2 (193)
H 2O 2 (193)
H 2O 2 (42)
H 2O 2 (4)
Ferroxi-dase
Y (118)
Y (117)
Y (2)
Y (2)
Y (188)
Y (189)
Y (189)
Y(189)
Y (89)
Y (191)
Y (26)
Y (193)
Y (193)
Y (42)
Y (4)
Y (96)
Iron
storage
Y (118)
Y (117)
Y (2)
N (2)
Y (188)
Y (189)
Y (189)
Y (189)
Y (89)
Y (192)
Y (193)
Y (193)
Y (42)
Y (4)
Y (96)
Oligomer
Dodecamer (118)
Dodecamer (117)
Dodecamer (187)
Trimer (2)
Dodecamer (27)
Dodecamer (189)
Trimer (189)
Dodecamer (189)
Trimer (189)
Dodecamer (189)
Dodecamer (89)
Dodecamer (190)
Dodecamer (77)
Dodecamer (191)
Dodecamer (26)
Dodecamer (193)
Dimer (193)
Dodecamer (193)
Dodecamer (42)
Dodecamer (96)
Gene
fri
aah
dpsA
dpsB
dps1
dps2
dps3
dps4
dps5
dps
mrgA
dpr
dpsA
dpsB
dpsC
dps
dpsA
dpsA
Species
Listeria monocytogenes
Microbacterium
arborescens
Mycobacterium
smegmatis
Nostoc punctiforme
Porphyromonas
gingivalis
Staphylococcus aureus
Staphylococcus suis
Streptomyces coelicolor
Sulfolobus solfataricus
Synechococcus sp.
Strain PCC 7942
Thermosynechococcus
elongatus
37
Other ion binding site
Other ions
bound
Phosphate (95)
K-rich N
term
N
DNA-
binding
Y (95)
Iron core
(aerobic)
300 (95)
Preferred
oxidant
H 2O 2 (95)
Ferroxi-dase
Y (95)
Iron
storage
Y (95)
Oligomer
Dodecamer (95)
Gene
dps
Species
Trichodesmium
erythraeum
38
Table 1.2: Dps as a stress response protein
Stress Species Growth phase*
Acid stress Escherichia coli (153)
Log (194, 195)
Stationary (62, 194)
Streptococcus pyogenes Log (196)
Base stress Escherichia coli Log (62)
Streptococcus pyogenes Log (196)
Carbon limitation Anabaena PCC 7120 Stationary (197)
Cold shock Listeria monocytogenes (198)
Streptococcus thermophilus (199, 200)
Copper stress Anabaena PCC 7120 Stationary (136, 197)
Escherichia coli
Log (201)
Stationary (62)
Endonucleases Campylobacter jejuni (91)
Helicobacter pylori (15)
Trichodesmium erythraeum (95)
Ethanol stress Bacillus cereus (57)
High NaCl Anabaena PCC 7120 Stationary (136, 197)
Bacillus cereus (57)
Escherichia coli Log (10)
Legionella pneumophila Stationary (116)
High pressure Escherichia coli Stationary (202)
Heat stress Anabaena PCC 7120 Stationary (136, 197)
Bacillus cereus Log (7, 57)
Escherichia coli Log (10)
Stationary (62)
Legionella pneumophila Stationary (116)
Iron excess Escherichia coli Log (10, 62)
Stationary (62)
Vibrio cholerae (126)
Iron limitation Anabaena PCC 7120 Stationary (136, 197)
Escherichia coli Stationary (KO, unpublished)
Nitrogen limitation Anabaena PCC 7120 Stationary (197)
Oxidative stress Agrobacterium tumefaciens (36)
Bacillus anthracis Stationary (44)
Bacillus cereus Log (57)
Bacillus subtilis Log (104, 146)
Stationary (146)
Campylobacter jejuni Log (91)
Escherichia coli Log (1, 10, 195)
Stationary (62, 147)
Helicobacter hepaticus (65)
Legionella pneumophila Log (116)
Listeria innocua (38)
Listeria monocytogenes Log (160)
Stationary (160)
Microbacterium arborescens Log (117)
Nostoc punctiforme (203)
Porphyromonas gingivalis (77)
Salmonella enterica sv. Typhimurium (109, 204) Log (158)
Staphylococcus aureus Log (30)
Stationary (13, 30)
Streptococcus mutans (205)
Streptococcus pyogenes Log (196, 206)
Streptococcus suis Stationary (41, 207)
Thermosynechoccus elongatus (43, 87)
Vibrio cholerae Log (126)
Stationary (126)
39
Stress Species Growth phase*
Phosphorus limitation Anabaena PCC 7120 Stationary (136, 197)
UV and gamma
irradiation
Anabaena PCC 7120 Stationary (136, 197)
Escherichia coli Stationary (62)
Staphylococcus aureus Log (30)
Visible light stress Nostoc punctiforme (128)
Zinc excess Escherichia coli Transition (62)
Streptococcus pyogenes Stationary (196)
*For species that are not assigned a growth phase for Dps-mediated stress response, the work was
done either in vitro or on plates.
40
Chapter 2: Differential contributions of ferritins to long-term survival in Escherichia coli
41
Introduction
Ferritins are iron detoxifying and storage proteins conserved across all domains of life
(208). They manage a major biological paradox: iron is both an important cofactor in many cellular
functions and a potent source of oxidative stress via formation of reactive oxidative species (ROS).
The latter is achieved through a chemical reaction with peroxide known as the Fenton reaction:
Fe
2+
+ H2O2 → [FeO]
2+
+ H2O
[FeO]
2+
+ H
+
→ Fe
3+
+ HO•
Fe(II) is one of two biologically relevant forms of iron. While this water-soluble form is capable
of producing ROS, Fe(III) is not. This principle is fundamental to ferritin function. First, iron enters
the cell in its water-soluble, more dangerous, ferrous (Fe(II)) form. Next, when a ferritin
encounters Fe(II), it can oxidize Fe(II) to Fe(III) at one of its ferroxidase sites. Once iron has been
detoxified to its ferric (Fe(III)) form, it enters the core of the ferritin. In this sense, ferritins act as
cellular iron sinks. Ferritins are hypothesized to release stored iron when the cell needs it; in this
sense, ferritins act as iron sources for the cell.
There are three highly conserved families of ferritins found in bacteria: ferritin (Ftn),
bacterioferritin (Bfr), and the DNA-binding protein from starved cells (Dps). An overarching
question about bacterial ferritins persists: why do bacterial genomes encode three ferritins?
Perhaps some hierarchy exists among them. Alternatively, redundancy may allude to the
importance of iron detoxification and storage. This chapter attempts to answer this question for
the three E. coli ferritins, which are among the most extensively studied bacterial ferritins.
Though all three E. coli ferritins function as iron detoxifying and storage proteins, there
are some notable differences among FtnA (the E. coli Ftn protein), Bfr, and Dps. First, they have
different iron storage capacities. FtnA and Bfr 24-mers can each hold up to 4000 Fe(III) atoms,
42
while the Dps dodecamer can only hold up to 500 Fe(III) (18, 34, 46). Considering the E. coli cell
holds ~10
5
-10
6
iron atoms (209, 210), this can theoretically be achieved by 25-250 fully loaded
FtnA or Bfr molecules or 8-fold more fully loaded Dps molecules. These are not excessive levels
of expression, which calls into question the idea that cells “need” all three of these proteins to deal
with the iron paradox. A second difference is the location of ferroxidase sites: the 24 ferroxidase
sites for FtnA and Bfr are located within each monomer, compared to the 12 ferroxidase sites for
Dps, which are each located between two monomers (86, 92). Third, the preferred oxidants for
Fe(II) at ferroxidase sites differ: FtnA and Bfr prefer O2, whereas Dps prefers H2O2 (34, 46). The
preference for H2O2 as oxidizing agent allows Dps to greatly reduce the amount of ROS in the cell
by diminishing the levels of both initial reagents for Fenton chemistry (Fe
2+
and H2O2). The final
and perhaps most surprising difference among E. coli ferritins is that Dps is also a DNA-binding
protein (1). During stationary phase, Dps is the major nucleoid-associated protein (NAP) (19–22),
during which time it is responsible for restructuring the nucleoid into 3-dimensional crystalline
lattice known as the biocrystal (24, 141). Dps may also function as a regulator of gene expression,
though there is conflicting evidence regarding this assertion (1, 75, 76, 137).
E. coli population dynamics change over time and can be distinguished as one of five
phases of growth and survival (211). During lag phase, the population adjusts to its new, nutrient-
rich environment, preparing to grow and divide. This phase lasts for about 2 hours and is
characterized by little population growth. Log phase denotes the next ~2 hours of rapid growth and
division. At this point, the population doubles approximately every 20 minutes. Population density
then levels out at high density during stationary phase in LB medium, which persists for one or
more days, depending on the strain and environment. After stationary phase, the population
undergoes death phase, during which ~99% of the population dies over the course of one to two
43
days. From there, the population enters long-term stationary phase (LTSP), during which time
population density remains relatively constant. LTSP can persist for years without the addition of
new nutrients in batch culture.
Growth and survival curves are used to measure population fitness. Compared to
examining individual timepoints, this method allows us to get a holistic picture of how population
dynamics change over time. The growth and survival curve is particularly powerful in establishing
time-sensitive trends in fitness when certain biologically relevant components are added or
subtracted from the laboratory environment to which populations are accustomed.
This study focuses on the effects of each of the E. coli ferritins on population survival and
dynamics. Over the course of 14 days, population density was measured for strains lacking one
ferritin gene under a range of environments. When compared to wild-type E. coli, each ferritin-
null strain (Table 2.2) behaves differently in monoculture, coculture, and under conditions of iron
stress, suggesting different roles and temporal importance of each ferritin.
44
Results
Contribution of ferritins to wild-type growth phenotype in long-term batch monoculture
The first question investigated here is the extent to which each of the three E. coli ferritins
contribute to the observed wild-type growth phenotype. To address this question, wild-type, ftnA-
null, bfr-null, and dps-null populations were cultured individually over a 14-day period, and
population density was measured daily. Wild-type cultures (Figure 2.1) reach high density (~10
9
cfu/ml) by day 1, remain at high density through day 2, then undergo death phase before entering
LTSP by day 4. dps-null cultures (Figure 2.1) experience a more drastic death phase and enter
later into long-term stationary phase. More specifically, dps-null population density decreases by
four orders-of-magnitude over the course of two days before regaining 100-fold cfu/ml over the
next three days. The ftnA-null strain (Figure 2.1) acts differently than either of the previously
discussed strains: it has a 1-day stationary phase, compared to the wild-type and dps-null 2-day
stationary phase, and experiences a two-step death (step 1 between days 1 and 2, step 2 between
days 3 and 4). bfr-null cultures (Figure 2.1) look relatively unperturbed compared to wild-type,
with a 2-day stationary phase and entrance into LTSP by day 4.
In a typical growth and survival curve experiment, population density is sampled daily.
However, early population dynamics set the stage for long-term population survival. To better
understand how ferritins impact growth during lag phase and log phase, monoculture outgrowth
into stationary phase was monitored for wild-type and ferritin mutants by sampling every 30
minutes over 5 hours (Figure 2.2). The wild-type strain has a 1.5-hour lag time before entering
log phase, and doubles, on average, every 21.4 minutes (Table 2.1). The bfr-null and ftnA-null
populations have a 1.5-hour lag time before entering log phase, similarly to the wild-type, but have
slightly longer doubling times of 22.9 and 24.3 minutes, respectively (Table 2.1). The most
45
different ferritin mutant outgrowth is that of the dps-null strain, with a longer lag time of 2 hours
and faster log phase doubling time of 18.7 minutes, though this strain does enter stationary phase
at the same time as the other three strains (Table 2.1, Figure 2.2).
46
Competition of mutant strains with wild-type
Coculture can add another pressure to populations by making strains compete for resources.
To better understand the magnitude of growth advantage conferred by Bfr, FtnA, and Dps, each
ferritin-null strain was cocultured with wild-type cells. When cocultured with wild-type, the dps-
null strain experiences a consistently lower population density (Figure 2.3). Here the dps-null
strain experiences a 1-day stationary phase. Like its monoculture, the dps-null strain experiences
a drastic death phase. However, in coculture, the dps-null population density does not recover to
wild-type levels during LTSP like it does in monoculture, instead remaining 10- to 100-fold lower
in population density than its wild-type counterpart.
The ftnA-null strain exhibits temporal advantage over the wild-type strain in these
experimental conditions (Figure 2.3). As in monoculture, the ftnA-null strain has a 1-day
stationary phase; in coculture with wild-type, the two-step death seen in monoculture is abolished.
The wild-type maintains higher population density between days 1 and 4. By day 4, population
density is roughly even for the wild-type and ftnA-null populations, which continues through day
9 or 10. At that point, the ftnA-null strain outcompetes the wild-type by about 10-fold.
When cocultured with wild-type, the bfr-null strain exhibits an altered survival phenotype
(Figure 2.3). In death phase, the bfr-null strain exhibits a lower-than-wild-type population density.
Between days 3 and 7, the bfr-null strain manifests a modest disadvantage (<10-fold lower
population density). By day 8, the gap between wild-type and bfr-null population density widens
to ~10-fold, which continues through the end of the experiment.
47
Competition of mutant strains against one another
To parse apart the temporal importance of the E. coli ferritins, ferritin mutants were
cocultured with one another. When the dps-null and ftnA-null strains are cocultured, both strains
exhibit a 1-day stationary phase (Figure 2.4). The ftnA-null strain’s two-step death is observed
again in this coculture. The dps-null strain experiences a more drastic death than during its
monoculture, and is driven to around the limit of detection by day 9.
During bfr-null and dps-null coculture, the dps-null strain exhibits a 1-day stationary phase
and a more drastic death than that of the bfr-null population (Figure 2.4). However, the dps-null
strain experiences a less severe dip in population density between days 2 and 7 than in monoculture
or coculture with wild-type. Between days 7 and 9, the dps-null strain displays only marginally
lower population density than the bfr-null strain, before the former is driven to just above the limit
of detection by the end of the experiment.
When cocultured with the bfr-null strain, the ftnA-null strain exhibits its typical 2-step
death phase (Figure 2.4). The bfr-null strain behaves normally through day 7. After day 7, the bfr-
null strain population density drops, and ultimately ends with population density near or below the
limit of detection.
48
Varying iron availability impacts ferritin mutant monoculture growth and survival
Because ferritins act as both sources and sinks of iron, the effect of adding or sequestering
iron in the cultures was tested. Iron was added as 100 µM Fe2SO4 and depleted using 200 µM 2,2’-
dipyridyl. To ensure that the level of iron addition or sequestration would not add pressure beyond
that which could be handled by typical ferritin levels, iron was first added or sequestered in wild-
type cultures (Figure 2.5). Compared to the typical growth and survival curve, the iron
supplemented and depleted populations demonstrate a 2-day stationary phase, followed by death
of approximately 99% of the population, and a relatively stable population density following death
and through the end of the experiment. This is also the case for the bfr-null population in either of
the iron-altered environments (Figure 2.5). Neither treatment affected growth in either strain.
When supplemented with 100 µM Fe2SO4, the dps-null population (Figure 2.5) follows
the same general trend as the population grown in LB, with a slightly (~5-fold) higher population
density in the supplemented environment on day 4. The curve becomes more atypical compared
to wild-type in the iron sequestered environment. First, stationary phase shifts to only one day in
this environment. The population in the iron depleted environment undergoes a similar, drastic
death phase, but notably does not reach density typical of LTSP by the end of the experiment.
Addition or depletion of iron similarly alters the ftnA-null growth and survival curve
(Figure 2.5). In the case of iron addition, the population exhibits a small (~3-fold) increase in
population density on day 2 than the unsupplemented population, which is more similar to wild-
type population dynamics at this time. In both the iron supplemented and depleted environments,
the 2-step death seen in the unsupplemented ftnA-null population is abolished in favor of a 1-step
death where 99% of the population dies. Both the iron supplemented and depleted populations
enter LTSP by day 4, a similar timeline to that of the wild-type.
49
Varying iron availability impacts ferritin growth and survival in competition
To study the temporal importance of each E. coli ferritin in context of their functions as
iron source and sink, cocultures of all three single ferritin-null strains were performed in LB, iron
supplemented (100 µM Fe2SO4), or iron depleted (200 µM 2,2’-dipyridyl) conditions.
Under normal laboratory conditions, a temporal hierarchy of survival emerges. All three
strains outgrow to high density (in the range of 10
9
cfu/ml) by day 1 (Figure 2.6). The bfr-null
population undergoes a typical 2-day stationary phase followed by a death phase in which ~99%
of the population dies off, and enters LTSP by day 3. The ftnA-null population exhibits a 1-day
stationary phase followed by a hastened (compared to monoculture) 2-step death, entering LTSP
by day 4. During this time, the dps-null population experiences a 1-day stationary phase followed
by a decrease in population density of ~2.5 orders-of-magnitude by day 2, and an additional 10-
fold decrease by day 3. Over the remainder of the experiment, the ftnA-null population retains
relatively consistent population density, the dps-null population is driven below the limit of
detection by day 11, and bfr-null population density falls to below 10
4
cfu/ml by the end of the
experiment.
To test the contribution of each ferritin as an iron source or sink, parallel competitions
among the three mutant populations supplemented with Fe2SO4 or 2,2’-dipyridyl were performed.
When supplemented with 100 µM Fe2SO4, the bfr-null population displays a 2-day stationary
phase followed by a death phase during which ~99% of the population dies, and enters stationary
phase by day 3 (Figure 2.6). The bfr-null population density begins to drop around day 6, but in
this case the endpoint population density is in the range of 10
5
cfu/ml (compared to 10
3
or below)
or, in one case, 10
7
cfu/ml. Similarly to the unsupplemented competition, the dps-null population
experiences a 1-day stationary phase followed by a drastic death phase, and by day 8 remains at or
50
below the limit of detection for the remainder of the experiment. The ftnA-null growth and survival
curve under these conditions looks similar to the unsupplemented environment, with a shift in
death phase from a 2-step death to a 1-step death in the supplemented competition.
When iron is chelated with 200 µM 2,2’-dipyridyl, the bfr-null population undergoes a 2-
day stationary phase followed by a 1-day death phase, and enters LTSP by day 3 (Figure 2.6). As
in the iron supplemented culture, the ftnA-null population exhibits a 1-day stationary phase and a
1-step death phase, here entering LTSP by day 2. Both the bfr-null and ftnA-null populations have
a 10-fold dip in density between days 4 and 6. After that dip, the ftnA-null population remains
above 10
7
cfu/ml for the remainder of the experiment, as in the unsupplemented competition, and
bfr-null population density decreases in the final 4 days of the experiment, ending in one case
around 10
7
cfu/ml, in another case around 10
5
cfu/ml, and in the final case around 10
3
cfu/ml. The
dps-null population, as in the previous two competitions with both other ferritin-null populations,
undergoes a 1-day stationary phase followed by a drastic death phase. Unlike the other 3-way
competitions, however, here the dps-null population gradually drops to below the limit of detection
by day 13.
51
Ferritins are differentially expressed by growth phase
One explanation for the growth phase-dependent variation of altered ferritin-null strain
survival phenotypes is differences in gene expression patterns. To assess whether this might be the
case, we examined published mRNA expression patterns for each of the E. coli ferritins across 8
days of growth (Figure 2.7) (212). This data shows high expression of dps, particularly during the
transition from log phase into stationary phase (here, the 8-hour timepoint), as has previously been
described (19). Further, dps expression remains relatively high (at least thousands of transcripts
per million (tpm)), even higher than the highest bfr or ftn expression, through LTSP. Interestingly,
while ftnA and bfr were expressed at comparable numbers (~1000 tpm per gene) in mid-log phase
(4 hours), by stationary phase ftnA transcripts per million had dropped to tens, while bfr dropped
to hundreds of transcripts per million. These numbers persisted through the end of the experiment.
52
Discussion
This study asks why the E. coli genome encodes three proteins to perform one task. A bfr-
null strain of E. coli shows no survival defect in long-term batch monoculture (Figure 2.1, Figure
2.2, Table 2.1), suggesting that there is sufficient activity to maintain viability in the absence of
Bfr in monoculture under these laboratory conditions. However, when grown in coculture with
wild-type, the bfr-null strain begins to show a modest decrease in population density after
stationary phase compared to its wild-type competitor, which becomes greater (10- to 100-fold
lower than wild-type) by day 14 (Figure 2.3). This suggests that, post-stationary phase, Bfr begins
to confer a growth advantage, which becomes increasingly vital as a population advances
throughout LTSP.
An ftnA-null strain undergoes a 1-day stationary phase and 2-step death phase in long-term
batch monoculture (Figure 2.1). This suggests ftnA confers an advantage during stationary and
death phase. In coculture with wild-type, the ftnA-null strain displays a 1-step death phase and
lower population density than its wild-type competitor by day 11 (Figure 2.3). This suggests FtnA
is important during stationary phase, but presents a disadvantage during LTSP. Perhaps FtnA helps
to stabilize iron levels during stationary phase, but more cellular energy is put into maintaining
FtnA levels during LTSP than is necessary to provide a growth advantage. This may have to do
with how iron is trafficked, perhaps via stability of ferritin iron cores, which may in turn impact
the ability to resolubilize Fe(III) to Fe(II) from FtnA.
A dps-null strain has a longer lag time (Figure 2.2, Table 2.1), undergoes a drastic death
phase, and enters LTSP later than wild-type in monoculture (Figure 2.1), suggesting Dps confers
a growth advantage during lag, log, death, and potentially LTSP. Perhaps stress response is more
difficult to regulate during stationary phase when a strain lacks Dps. When grown in coculture
53
with wild-type, the dps-null strain undergoes a 1-day stationary phase, 1,000-fold death phase, and
has a consistently lower population density than wild-type through day 14 (Figure 2.3). This
suggests Dps is important for growth and survival in stationary, death, and LTSP. Taken together,
perhaps FtnA is a preferred iron source/sink during log and stationary phase, Bfr is a preferred iron
source/sink during death and LTSP, and Dps is a preferred iron source/iron sink/DNA-binding
protein across the growth curve.
Pairwise competition between ferritin mutants reveals a survival hierarchy among the
ferritins. When grown in coculture with an ftnA-null strain, the dps-null strain is increasingly
outcompeted starting by day 2 (Figure 2.4), suggesting dps is more important than ftnA for
conferring a growth advantage during stationary, death, and LTSP. When grown in coculture with
a bfr-null strain, the dps-null strain is outcompeted starting by day 2 and lasting through the end
of the experiment (Figure 2.4). This suggests that dps is more important than bfr for conferring a
growth advantage during stationary, death, and LTSP. When a bfr-null strain is grown in coculture
with an ftnA-null strain, the bfr-null strain outcompetes the ftnA-null strain on day 2, and the ftnA-
null strain outcompetes the bfr-null strain between days 3 and 4, then again after day 7 (Figure
2.4). This suggests FtnA is more important than Bfr in conferring a growth advantage during
stationary phase, and Bfr is more important than FtnA in conferring a growth advantage during
LTSP. Taken together, Dps seems to be the most important of these proteins in conferring a growth
advantage during stationary, death, and LTSP. The data collected in these experiments suggests a
temporal hierarchy of ferritins in conferring a competitive advantage under these experimental
conditions: Dps is the most crucial during stationary phase, death phase, and LTSP; FtnA is the
second-most important during log phase, stationary phase, and death phase; and Bfr is the second-
most important during LTSP.
54
In a competition among the ferritin mutants, this hierarchy is further exposed (Figure 2.6):
the bfr-null strain outcompetes both other mutants through day 2, the ftnA-null strain outcompetes
the others on day 3, and the bfr-null and ftnA-null strains maintain relatively even population
densities between days 4 and 10, at which point the bfr-null population density begins to drop,
which continues through the end of the experiment. Between day 2 and the end of the experiment,
the dps-null strain has dramatically lower population density than its competitors, ultimately
dropping below the limit of detection by day 11. This accentuates the relative temporal
contributions of each ferritin to survival: Dps is the most important of these proteins during
stationary, death, and LTSP; FtnA is second-most important during stationary phase; and Bfr is
the second-most important during LTSP.
Because Dps functions both as a ferritin and a DNA-binding protein, either or both of its
functions could be important in conferring the growth advantage described above. When grown in
long-term batch monoculture supplemented with iron, the dps-null growth curve is essentially
unchanged; when iron is sequestered, the dps-null strain exhibits lower population density during
LTSP (Figure 2.5). This suggests that, for Dps, a small part of the growth advantage conferred is
due to its ferritin iron reservoir function, but much of this advantage is due to forces outside its
ferritin abilities. The ftnA-null curve, on the other hand, shifts from a 2-step death to a 1-step death
when supplemented with iron or when iron is sequestered (Figure 2.5), which makes its curve
look more like wild-type, suggesting this ferritin is important as both an iron source and sink in
conferring the observed wild-type growth advantage. The wild-type and bfr-null growth curves are
unchanged by the supplementation or sequestration of iron in monoculture (Figure 2.5),
suggesting both strains have enough preexisting iron sources and sinks in monoculture.
55
Competition among ferritin-null strains in iron supplemented and depleted environments
further exposes the temporal hierarchy of ferritins. These competitions further clarify that the
function of Dps as a ferritin is not the most important facet of this protein in conferring its growth
advantage. When the competition is supplemented with iron, the ftnA-null strain shifts from a 2-
step death to a 1-step death, the bfr-null strain loses viability more slowly and with lower
magnitude than in the unsupplemented competition, and the dps-null strain still loses viability
quickly and remains around the limit of detection through LTSP (Figure 2.6). This suggests the
capacity of FtnA and Bfr, but not Dps, as iron providers is important in the growth advantages
each of these proteins provides. When the competition is treated with an iron chelator, the ftnA-
null strain shifts from a 2-step to a 1-step death, the bfr-null strain starts losing viability later than
in the unsupplemented environment, and the dps-null strain still loses viability quickly and drops
below the limit of detection by the end of the experiment (Figure 2.6). This similarly suggests the
capacity of FtnA and Bfr, but not Dps, to sequester iron is important in the growth advantages each
of these proteins provide. Taken together, these results suggest that, while FtnA and Bfr provide
growth advantages rooted in their ferritin functions, the advantage Dps confers is largely not.
One explanation for the ferritin hierarchy presented here is differences in gene expression.
To this end, we examined RNA-Seq data for the three E. coli ferritins. Two caveats must be noted
for this data. First, the strain used in the published RNA-Seq experiment is PFM2 (MG1655 rph
+
);
the strains used in the original research presented here are ZK126 (W3110 DlacU169 tna-2)
derivatives (Table 2.2). These genotypes have different growth and survival phenotypes, the most
conspicuous of which is a 1-day stationary phase in PFM2 and a 2-day stationary phase in ZK126
in LB medium. Second, while this gene expression data is unique in the length of time post-
inoculation RNA-Seq was performed, the latest timepoint is 8 days post-inoculation, which is the
56
earliest timepoint in which the bfr-null strain displays a growth defect in LTSP. This methodology
thus may miss larger differences in gene expression later in LTSP. Keeping these caveats in mind,
this RNA-Seq data is informative in our model of ferritin hierarchy. dps is the most highly
expressed of the E. coli ferritin genes at any given timepoint, which has been previously
demonstrated; its highest expression during stationary phase is assigned to its DNA-binding ability
(19, 20). It is also possible that this is its purpose during LTSP. Turning to the other two ferritins,
bfr is more highly expressed than ftnA at every timepoint, which is intriguing given the phenotypic
ambivalence of the cell when bfr is absent. Though it is less highly expressed, FtnA has been
shown to greatly increase the amount of iron stored in the cell between log phase and stationary
phase, while Bfr has no meaningful impact on cellular iron storage during this time (5) (22). RNA-
Seq data demonstrates modest changes in dps, bfr, and ftnA mRNA levels from stationary phase
into LTSP (Figure 2.7). This is not unexpected, as populations in LTSP may deal with increased
oxidative stress, which ferritins can help alleviate. If FtnA is the major iron storage protein and
Bfr is the major iron detoxification protein in E. coli, as has been previously suggested (213), our
model that Bfr is the preferred ferritin during LTSP is supported by the cell’s preference for Bfr
as an antioxidant.
The hierarchy we present may alternatively be explained by differences in protein stability,
mRNA stability, or translation dynamics among dps, bfr, and ftnA. While there is no published
data for Bfr or FtnA protein turnover, Dps has been shown to be degraded rapidly during log phase
and degradation slows significantly during stationary phase (122). Additionally, no information
about dps, bfr, or ftnA mRNA stability or translation dynamics exists. Further studies into Bfr and
FtnA protein degradation; dps, bfr, and ftnA mRNA stability; and ferritin gene translation
dynamics would help validate these possibilities.
57
Conclusion
Through the work presented in this study, it is clear that, with respect to growth advantage
conferred, a hierarchy exists among ferritins in E. coli. While FtnA is important during stationary
and death phase and Bfr is important during LTSP, Dps is most important throughout the growth
curve. However, while some small part of this is due to its capacity as a ferritin, some other
function may confer a larger portion of this advantage. The most intuitive other function of Dps
that might explain its conferred fitness benefit is its DNA-binding ability, but that is comprised of
multiple functions, some of which may not have yet been elucidated. Thus, it would be beneficial
to further study the different DNA-binding properties of Dps, including its DNA-binding locations
(and whether these are predetermined or random), interactions with other DNA-binding proteins,
and its potential impact on gene expression.
58
Materials and methods
Strains and growth conditions
The characteristics of the strains used in this study are detailed in Table 2.2. The genetic
background of all strains is ZK126, which is E. coli K-12 W3110 DlacU169 tna-2 (214). For
competition experiments, either ZK1142 (ZK126 nalR) or ZK1143 (ZK126 strR) were used,
depending on the antibiotic resistance marker of its competitor (Table 2.3). Overnight cultures
were inoculated from 20% glycerol stocks stored at -80°C. Cultures were grown in 23 ml
borosilicate test tubes (Thermo Fischer) in 5 ml LB liquid medium (Difco). These were incubated
at 37°C with 70% humidity on a TC-7 roller drum (New Brunswick Scientific).
Growth curves
All cultures were inoculated at a total final concentration of 1:1000 (v:v) from one or more
overnight cultures. For monocultures, this meant addition of 5 µl of an overnight to 5 ml of LB.
For 2-strain cocultures, this meant addition of 2.5 µl of each of 2 overnights to 5 ml of LB. For 3-
strain cocultures, this meant addition of 1.67 µl of each of 3 overnights to 5 ml of LB. To count
colony forming units per ml (cfu/ml), 10 µl of culture was diluted 1:10 in serial dilution and 10 µl
of each dilution was plated on LB (in monoculture) or LB with appropriate antibiotic (in coculture)
either 1) daily for 14 days for long-term growth curves or 2) every 30 mins for 5 hours for
outgrowth curves. Antibiotics (Sigma-Aldrich) were used at the following concentrations:
nalidixic acid, 20 µg/ml; streptomycin, 25 µg/ml; kanamycin, 50 µg/ml. Titer plates were
incubated at 37°C with 70% humidity overnight to develop visible colonies and number of colonies
was counted.
59
Iron supplementation and depletion
Growth curves were performed as described above, with one change. To supplement a
culture with iron, 100 µM Fe2SO4 (Ward’s Science) was added before inoculation. To deplete iron
from a culture, 200 µM 2,2’-dipyridyl (Sigma) was added before inoculation. These concentrations
were determined using minimum inhibitory concentration (MIC) curves, where varying amounts
of either chemical were added to culture tubes before inoculation with wild-type (ZK126) E. coli
cells. The highest concentration of each that did not perturb wild-type growth was selected for
further study.
60
Figures
Figure 2.1: Monoculture of wild-type and ferritin mutants
Monoculture long-term growth curves of wild-type and ferritin mutant populations. Black, wild-
type; blue, dps-null; gold, ftnA-null; purple, bfr-null. Circles denote replicate A; triangles denote
replicate B; squares denote replicate C.
●
●
●
●
●
● ●
● ●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ● ● ●
● ●
●
●
● ● ●
●
●
●
●
● ●
●
● ●
●
●
●
●
● ●
10
3
10
4
10
5
10
6
10
7
10
8
10
9
10
10
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
day
cfu/ml
wild-type
bfr-null
ftnA-null
dps-null
61
Figure 2.2: Monoculture outgrowth of wild-type and ferritin mutants
Monoculture outgrowth growth curves of wild-type and ferritin mutant populations. Black, wild-
type; blue, dps-null; gold, ftnA-null; purple, bfr-null. Circles denote replicate A; triangles denote
replicate B; squares denote replicate C.
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
10
6
10
7
10
8
10
9
10
10
0 1 2 3 4 5
hour
cfu/ml
wild-type
dps-null
bfr-null
ftnA-null
62
Figure 2.3: Ferritin mutant coculture with wild-type populations
Coculture of wild-type with A) dps-null, B) ftnA-null, and C) bfr-null strains. Black, wild-type;
blue, dps-null; gold, ftnA-null; purple, bfr-null. Circles denote replicate A; triangles denote
replicate B; squares denote replicate C.
10
3
10
4
10
5
10
6
10
7
10
8
10
9
10
10
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
day
cfu/ml
wild-type vs. dps-null
10
3
10
4
10
5
10
6
10
7
10
8
10
9
10
10
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
day
cfu/ml
wild-type vs. ftnA-null
10
3
10
4
10
5
10
6
10
7
10
8
10
9
10
10
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
day
cfu/ml
wild-type vs. bfr-null
wild-type
dps-null
wild-type
ftnA-null
wild-type
bfr-null
A B
C
63
Figure 2.4: Ferritin mutant pairwise coculture
Coculture of A) ftnA-null and dps-null, B) bfr-null and dps-null, and C) ftnA-null and bfr-null
strains. Blue, dps-null; gold, ftnA-null; purple, bfr-null. Circles denote replicate A; triangles denote
replicate B; squares denote replicate C.
10
3
10
4
10
5
10
6
10
7
10
8
10
9
10
10
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
day
cfu/ml
ftnA-null vs. dps-null
10
3
10
4
10
5
10
6
10
7
10
8
10
9
10
10
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
day
cfu/ml
bfr-null vs. dps-null
10
3
10
4
10
5
10
6
10
7
10
8
10
9
10
10
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
day
cfu/ml
bfr-null vs. ftnA-null
ftnA-null
dps-null
bfr-null
dps-null
ftnA-null
bfr-null
A B
C
64
Figure 2.5: Monoculture with varied iron availability
Monoculture of wild-type and mutant strains in A) LB, B) LB with 100 µM Fe2SO4, or C) LB with
200 µM 2,2’-dipyridyl. Black, wild-type; blue, dps-null; purple, bfr-null; and gold, ftnA-null.
Circles, triangles, and squares denote replicate A, B, and C, respectively.
10
3
10
4
10
5
10
6
10
7
10
8
10
9
10
10
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
day
cfu/ml
LB
10
3
10
4
10
5
10
6
10
7
10
8
10
9
10
10
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
day
cfu/ml
Iron added
10
3
10
4
10
5
10
6
10
7
10
8
10
9
10
10
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
day
cfu/ml
Iron depleted
bfr-null
dps-null
ftnA-null
wild-type
bfr-null
dps-null
ftnA-null
wild-type
bfr-null
dps-null
ftnA-null
wild-type
A B
C
65
Figure 2.6: Competitions among ferritin-null populations with varied iron availability
Coculture of dps-null, ftnA-null, and bfr-null populations in A) unsupplemented LB, B) LB with
100 µM Fe2SO4, and C) LB with 200 µM 2,2’-dipyridyl. Blue represents dps-null, yellow
represents ftnA-null, and purple represents bfr-null. Circles, triangles, and squares denote replicate
A, B, and C, respectively.
10
3
10
4
10
5
10
6
10
7
10
8
10
9
10
10
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
day
cfu/ml
LB
10
3
10
4
10
5
10
6
10
7
10
8
10
9
10
10
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
day
cfu/ml
Iron added
10
3
10
4
10
5
10
6
10
7
10
8
10
9
10
10
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
day
cfu/ml
Iron depleted
bfr-null
dps-null
ftnA-null
bfr-null
dps-null
ftnA-null
bfr-null
dps-null
ftnA-null
A B
C
66
Figure 2.7: Ferritin gene expression profiles
mRNA abundance for ferritins in E. coli: blue, dps; purple, bfr; and yellow, ftnA. Here, hour 4
corresponds to early log phase, hour 8 corresponds to the transition between log phase and
stationary phase, hour 24 corresponds to stationary phase, hour 72 corresponds to LTSP post-death
phase, and hours 144 and 192 correspond to LTSP. mRNA sequencing performed in triplicate:
distinct shape markers denote replicate. Data from Kram et. al 2020 (212).
10
100
1000
10000
144 144 144 192 192 192 24 24 24 4 4 4 72 72 72 8 8 8 144 144 144 192 192 192 24 24 24 4 4 4 72 72 72 8 8 8 144 144 144 192 192 192 24 24 24 4 4 4 72 72 72 8 8 8
hour
transcripts per million
gene
dps
bfr
ftnA
replicate
A
B
C
67
Tables
Table 2.1: Lag time and log growth rate by strain
Strain Lag time (hours) Average doubling time (minutes)*
Wild-type 1.5 hours 21.4
dps-null 2 hours 18.7
bfr-null 1.5 hours 22.9
ftnA-null 1.5 hours 24.3
68
Table 2.2: Strains used in this study
Strain Genotype Reference
ZK126* W3110 DlacU169 tna-2 (214)
ZK1142 ZK126 nalR (214)
ZK1143 ZK126 strR (214)
ZK1058 ZK126 dps::kan (1)
SF2426 ZK126 bfr::nal (215)
SF2433 ZK126 ftnA::str (215)
* ZK126 is the parental strain for the other strains used in this study.
69
Table 2.3: Competition schema
Competition scheme Strain A Strain B Strain C
Wild-type vs. dps-null ZK1142 ZK1058
Wild-type vs. bfr-null ZK1143 SF2426
Wild-type vs. ftnA-null ZK1142 SF2433
dps-null vs. bfr-null ZK1058 SF2426
dps-null vs. ftnA-null ZK1058 SF2433
bfr-null vs. ftnA-null SF2426 SF2433
dps-null vs. bfr-null vs. ftnA-null ZK1058 SF2426 SF2433
70
Chapter 3: E. coli Dps organizes the stationary phase nucleoid in a non-random manner
71
Introduction
Under laboratory conditions, Escherichia coli experiences five major phases of growth and
survival. Upon introduction into fresh, rich medium, such as LB, cells enter lag phase, which is
characterized by little population growth as the cells adjust to their new, nutrient-rich environment.
Log phase, a period of rapid growth and cell division during which the population can double in
number every 20 minutes, follows. In stationary phase, population density plateaus and can remain
high for several days. Then, approximately 99% of the population dies during death phase. After
death phase, population density remains relatively stable (though slowly decreasing) during long-
term stationary phase (LTSP), which can persist for years without the addition of new nutrients
(211).
In addition to population density, the structure of the nucleoid (the region of the prokaryotic
cell that contains the chromosome and its associated proteins) changes with growth phase. During
log phase, chromatin is intermixed with ribosomes (25). During the transition from log to
stationary phase, toroidal (ring-shaped) structures devoid of ribosomes appear, inside which the
chromatin is more regularly spaced than in log phase chromatin (25). After cells fully transition to
stationary phase, much of the nucleoid becomes regularly packed and is segregated from
ribosomes (25). This nucleoprotein complex is a crystalline lattice referred to as the biocrystal
(11).
This stationary phase-specific restructuring of the E. coli nucleoid requires the DNA
binding protein from starved cells, Dps (24). Dps is the major nucleoid-associated protein (NAP)
during stationary phase (11). In addition to its DNA binding properties, Dps is also a ferritin protein
in E. coli and many other species (see Chapter 2) (18). The protein uses both functions to confer
protection to DNA (10). Notably, Dps has no accepted sequence or structural specificity for DNA
72
binding, though a consensus sequence derived from in situ SELEX-Seq has been proposed (76).
Further, it is unknown whether Dps binds the same genomic loci during every iteration of
biocrystal formation.
This study begins to address the question of Dps binding specificity. Using wild-type and
dps-null populations, crosslinked nucleoids were digested using restriction enzymes. Following
decrosslinking, PCR with site-specific primers was used to measure relative amounts of digestion
at each of the genomic restriction sites probed. The results show that certain sites are bound
preferentially by Dps during stationary phase-specific nucleoid restructuring, suggesting some
DNA binding specificity. This specificity may be due to particular DNA sequences, DNA
structures, or some combination of the two. It may also reflect a more opportunistic binding pattern
based on which sites are vacated by log phase-specific NAPs.
73
Results
Experimental design
The goal of this study was to better understand the changing protein occupancy landscape
of the E. coli genome between mid-log phase and late stationary phase, and how Dps impacts
accessibility. To this end, a PCR-based method was developed (Figure 3.1). First, a restriction
enzyme digestion was performed on crosslinked chromatin from one of four conditions: wild-type
or dps-null cells harvested at mid-log phase or late stationary phase. Three restriction enzymes
were chosen: AvrII (C^CTAGG), which has 14 restriction sites in the E. coli genome, NotI
(GC^GGCCGC), which has 22 restriction sites, and XbaI (T^CTAGA), which has 38 restriction
sites (Figure 3.2). After restriction enzyme digestion, chromatin was de-crosslinked, and DNA
was collected using ethanol precipitation. PCR was performed to 25 cycles (so as not to saturate
the reactions with PCR product) using primers flanking each restriction site. Agarose gel
electrophoresis was performed on PCR primers, and gel band intensity for each PCR product was
quantified using ImageJ (216). Finally, gel band intensity was compared between wild-type and
dps-null nucleoids at each timepoint.
74
DNA is more accessible in dps-null nucleoids during stationary phase
During log phase, Dps presence does not explain the observed wild-type levels of DNA
protection from restriction enzymes, with most sites displaying increased accessibility when Dps
is present (Figure 3.3). However, during stationary phase, the dps-null nucleoid confers lower
levels of protection compared to the wild-type population (Figure 3.3). Moreover, the dps-null
nucleoid is more accessible at more restriction sites during stationary phase than the wild-type
nucleoid. Out of the 74 sites surveyed, 20 were differentially protected between wild-type and dps-
null populations solely during log phase, 26 were differentially protected uniquely during
stationary phase, and eight were differentially protected due to Dps presence during both log and
stationary phase (statistical significance determined using Student’s t test).
75
A putative Dps binding motif
The 26 restriction sites that were differentially protected in the presence or absence of Dps
in a stationary phase-specific manner presented an interesting subset of the genome that may point
to a Dps binding motif. To probe for such a DNA binding motif, the DNA sequences of each of
these 26 restriction sites, along with 1kb upstream and 1kb downstream, was analyzed with
MEME, a motif search program (217). One 21 bp motif was present in all 26 2kb regions of
interest: KGCMGMGAAASCGGCVGCWKM (Figure 3.4). There is some inverted dyad
symmetry (two regions of a strand of DNA whose nucleotides are inverted repeats of one another)
in this consensus motif (2-5 against 12-15), which is replicated in 17 of the significant sites (Figure
3.4, Table 3.1). Inverted dyad symmetry has been demonstrated in other NAPs, including Fis, the
NAP to which Dps is most often compared (218). There seems to be no correlation of motif-
restriction site distance with strength of protection by Dps – distance between restriction site and
motif ranges from as close as 2 bp to as far as 945 bp (Table 3.1). Interestingly, occurrences of this
motif are overrepresented compared to occurrences expected by, on average, about two orders-of-
magnitude, depending on the stringency (number of mismatches tolerated) of the alignment (Table
3.2). Additionally, this motif is not significantly similar (q < .05) to other known prokaryotic DNA-
protein binding motifs when compared using Tomtom (219). Moreover, this consensus motif is
overrepresented to a much greater degree in the 2kb regions centered on the stationary phase-
specific significant sites than in those centered on the remaining 48 sites at every tested alignment
stringency. The stationary phase-specific dps-dependent significant sites display one (exact
match), two (one mismatch), three (two mismatches), 15 (three mismatches), 31 (four
mismatches), and 74 (five mismatches) occurrences of KGCMGMGAAASCGGCVGCWKM,
compared to zero (exact match or one, two, or three mismatches), one (four mismatches), and five
76
(five mismatches) expected occurrences of this motif. There are zero (exact match or one, two, or
three mismatches), two (four mismatches), and eight (five mismatches) occurrences expected and
zero (exact match or one or two mismatches), one (three mismatches), six (four mismatches), and
39 (five mismatches) observed in the remaining 48 sites (Table 3.3). This results in a 10-fold
increase in overrepresentation in the stationary phase-specific sites compared to the non-stationary
phase-specific sites. Together, these data support a model where this motif is uniquely enriched in
the regions that are differentially protected uniquely during stationary phase.
77
Discussion
This study presents a novel method to explore genome accessibility using chromatin in its
native structure. Moreover, this method uses common lab techniques and free software to produce
reproducible results. Notably, the 74 restriction sites probed represent 488 nucleotides, less than
.01% of the genome. This is an important feature of these results: a small portion of the genome
was probed, essentially at random, and significant loci for potential genome condensation via Dps
action were established.
Of the 74 restriction sites probed, 20 (~27%) were reproducibly differentially protected
between wild-type and dps-null populations only during log phase. Additionally, eight restriction
sites (~11%) were reproducibly differentially protected between wild-type and dps-null
populations during both log and stationary phase. Structure of the dps-null nucleoid presents a
likely explanation for these results. When Dps is absent, the stationary phase nucleoid assembles
into a liquid crystalline cholesteric phase (24). The molecular structure of crystals (which liquid
crystalline phases possess) is necessarily sturdy, likely making this nucleoid structure take longer
to revert to log phase conformation than the wild-type nucleoid. Because cultures are inoculated
from stationary phase populations, the log phase timepoints may reflect this lagging stationary
phase effect; the sites found to be differentially protected during both log and stationary phase may
be a product of the additional time required for the dps-null stationary phase nucleoid to change
its structure. Similarly, the significant differences observed uniquely during log phase may fall
into regions of DNA that are intermediately differentially protected due to the slower unraveling
of the dps-null nucleoid.
26 restriction sites (~35%) were reproducibly differentially protected between wild-type
and dps-null populations only during stationary phase. While this is a substantial proportion of the
78
sites examined, 35% is likely an underrepresentation of the sites bound by Dps in a wild-type
population: previous research suggests as much as 50% of the genome is bound by Dps during
stationary phase (22). If Dps binds the same genomic regions during every biocrystal formation,
as is one potential Dps DNA-binding model, this disparity may reflect the small number of
nucleotides probed (<.01% of the genome) or a limitation of the method (quantifying bands on a
gel could produce more noise and thus more false negatives than a standard qPCR experiment). If
Dps binds regions based on low NAP occupancy, reflected in a more opportunistic Dps DNA
binding model which suggests a lack of both reproducibility and statistical significance among loci
bound by Dps, this disparity may reflect differences due to chance or protection induced by
environmental factors that are stochastically unbound by other NAPs. However, the most likely
scenario suggested by this data lies on a spectrum between the two aforementioned models: some
sites bound by Dps are bound consistently, while others exhibit more plasticity. This is borne out
especially when considering the overall increased accessibility in dps-null nucleoids during
stationary phase. While individual loci may be bound consistently, including those determined to
be statistically significant, others may be bound differentially, with Dps binding those loci in some
subsets of the population and not in others. In any case, the result is a decrease in accessibility in
both the number of sites and magnitude of restriction enzyme digestion during stationary phase
when Dps is present.
Dps DNA-binding activity may be modulated by a variety of factors, including vacancy of
log phase NAPs. For example, Dps and Fis expression levels invert between log and stationary
phase, and Fis has been shown to occlude Dps binding, which has led to some speculation about
antagonism between these two proteins (19, 58, 63). Further, Fis and H-NS have been shown to
regulate Dps expression during log phase (81). Taken together, these data make a compelling case
79
for log phase NAP interference with Dps that may extend to DNA binding. Perhaps when Dps first
binds DNA during early stationary phase it binds loci that are immediately available to it, including
regions that are first vacated by log phase-specific NAPs. In addition to occlusion by other NAPs,
DNA binding may also be affected by the motif proposed in this chapter
(KGCMGMGAAASCGGCVGCWKM) or 3D genome organization, as accessibility may play a
role in Dps binding.
The 26 stationary phase-specific protected restriction sites lent themselves to the discovery
of a putative Dps DNA-binding motif. This 21bp motif is overrepresented in the genome (Table
3.2) and similar in length to motifs published for other E. coli NAPs (218, 220). Other important
sites overrepresented in the E. coli genome include Chi sites, which are sites proximal to regions
of homologous recombination (221). As has been postulated about Chi sites, this suggests an
evolutionary pressure to acquiring and maintaining these sites; this pressure may be to promote
Dps binding for increased survival during stationary phase. Additionally, this motif displays some
inverted dyad symmetry, which has also been demonstrated for several NAPs, including Fis, the
NAP to which Dps is most frequently compared (222). However, this motif is both overrepresented
to a greater degree than the consensus sequences for Fis and IHF and occurs much less frequently
than the Fis and H-NS consensus sequences (Table 3.2). On average, the motif determined by this
study occurs more often than it is expected to by two orders-of-magnitude, compared to 1.4X for
Fis and 2.4X for H-NS. This putative motif also occurs less frequently by several orders of
magnitude compared to either the Fis or H-NS consensus sequences, a fact that clashes with the
sheer volume of Dps in the late stationary phase nucleoid, during which time it has been shown to
bind up to 50% of the E. coli genome (22). These data, taken together with the fact that the Dps
consensus sequence presented in this study is based on a sample size of 26, suggest it is unlikely
80
that the Dps consensus sequence is exactly what is presented here. However, a model for Dps-
dependent genome organization based on toroid formation is possible taking these factors into
consideration. Perhaps Dps first binds this over-represented but less frequent motif during the
transition from log phase to stationary phase, and subsequently binds other loci with low or no
specificity. These parallel phenomena may then converge to form toroids, which are the platform
for biocrystal formation. Further study is needed to determine the validity of the motif presented
in this chapter.
In addition to addressing the phenomenon of Dps DNA binding reproducibility, further
investigation into how Dps organizes the stationary phase chromosome is necessary. Whether Dps
remains at a particular locus throughout stationary phase once bound, for example, or whether Dps
molecules shift loci between the transition to stationary phase and late stationary phase, has yet to
be examined. Higher resolution methods to examine genome-wide accessibility during stationary
phase, such as MNase-Seq, can also address the question probed by this study at a more fine-scale
level (see Chapter 5). Finally, 3D genome deconvolution methods such as HiC, which pairs
Chromosome Conformation Capture (3C) with sequencing to identify long-range genomic
interactions, could provide more insight into how, why, and when Dps affects location in 3D space
of certain genetic loci, especially when compared to this data or data from an MNase-Seq
experiment.
81
Conclusion
By probing the nucleoids of wild-type and dps-null E. coli populations using restriction
enzymes, this study has identified 26 genomic loci that are repeatably protected during stationary
phase by Dps. Reproducible, site-specific binding and protection by Dps has not been shown
before. This reproducibility supports a model of some consensus DNA sequence or structure that
promotes Dps binding. Using MEME to search for motifs conserved among 26 genomic sites
reproducibly protected by Dps in a stationary phase-specific manner, we propose a 21bp consensus
sequence that promotes Dps binding of DNA. While this consensus sequence is overrepresented
in the E. coli genome, suggesting an evolutionary advantage exists in conjunction with these sites
via promotion of Dps binding to increase stationary phase survival, further study is needed to
validate this putative Dps DNA-binding motif.
82
Methods
Bacterial growth conditions
E. coli strains (Table 3.4) stored in frozen LB-glycerol stocks (-80°C) were inoculated into
5 ml of fresh LB medium (Difco) in 23 ml borosilicate test tubes (Thermo Fischer) and grown
overnight. These overnight cultures were used to inoculate fresh LB medium of the same volume
in the same test tube dimensions. All cultures were grown aerobically at 37°C with 70% humidity,
with constant agitation via TC-7 roller drum (New Brunswick Scientific). Once cultures reached
appropriate population density (~3 x 10
7
cfu/ml for log phase, around 2.5 hours post-inoculation;
~2 x 10
9
cfu/ml for late stationary phase, ~48 hours post-inoculation), cultures were harvested for
further analysis.
Nucleoid preparation
Nucleoid preparation was performed similarly to the preparation described by Marbouty
et. al (223), with some changes. Briefly, once at the appropriate growth phase, bacteria were
resuspended in TE. To crosslink chromatin, formaldehyde was added to final concentration 1%;
populations were incubated for 30 minutes at room temperature, then 30 minutes at 4°C.
Crosslinking was quenched with glycine at a final concentration of 0.25M; populations were
incubated at room temperature for 5 minutes, then at 4°C for 15 minutes. Cells were resuspended
in 25 µl TES (TE + 100 mM NaCl) (pH 7.5). 250U (from 250U/µl stock in TES) of Ready-Lyse
lysozyme (Lucigen) was added to each tube, and incubated at room temperature for 15 minutes
with occasional swirling (follow manufacturer’s instructions). To quench the lysozyme reaction,
SDS was added to final concentration 0.5%, and incubated for 10 minutes at room temperature.
Nucleoids were stored in 5 µl aliquots at -80°C until further use.
83
Nucleoid accessibility assays
5 µl crosslinked nucleoids were resuspended in a final concentration 1X Cutsmart Buffer
(NEB). To quench SDS activity, Triton X-100 was added to final concentration 1%, and incubated
for 5 minutes at room temperature. 20U of the appropriate restriction enzyme (AvrII-HF [NEB],
NotI-HF [NEB], or XbaI [NEB]) was added to the tube, and incubated at 37°C (optimal
temperature for enzyme used, follow manufacturer’s instructions) for 3 hours.
Once restriction digest was complete, proteinase K (VWR) was added to a final
concentration of 250 µg/ml (follow manufacturer’s instructions), and incubated at 65°C overnight.
DNA was collected using ethanol precipitation.
To quantify enzyme digestion in a site-specific manner, PCR primers were designed
flanking each restriction site with a target amplicon size of 500bp in the wild-type (ZK126)
genome (Table 3.5). For each site, Taq PCR reactions (Promega) were run for 25 cycles. 1 µl of
each PRC product was run out on a 1% agarose gel; 2 µl of 100 bp ladder (Invitrogen) was run on
each gel. Gel images (d.n.s.) were examined using ImageJ: background was subtracted for PCR
product bands and 500 bp step on 100 bp ladder; PCR product band intensity was normalized to
the 500 bp step. This normalized band intensity was used to compare relative accessibility between
wild-type and dps-null populations. Statistical significance was determined using Student’s t test.
Motif analysis
For each of the stationary phase-specific differentially protected restriction sites, a 2kb
region was used as the input for motif analysis. Each 2kb region was centered on the restriction
site and included 1kb of the DNA upstream and 1kb of the DNA downstream of the restriction
site. These 2kb regions were analyzed for common motifs occurring once per region using MEME
84
(217). To determine whether the stationary phase-specific motif was enriched specifically in these
regions, 2kb regions centered on the non-differentially protected restriction sites were also
searched for motifs using MEME, and statistically significant motifs were compared using
Tomtom (219). Tomtom was also used to compare the stationary phase-specific motif to published
prokaryotic motif databases (219).
85
Figures
Figure 3.1: A PCR-based approach to assess genome accessibility
Top: restriction sites (black ‘X’es) may be occluded by DNA-binding proteins (blue blobs),
including Dps, or left accessible to restriction enzyme action (top two rows). After restriction
enzyme digestion and DNA precipitation (sites noted in red circles), PCR was performed with
primers flanking each restriction site (gray half-arrows). Figure made with Biorender.com.
Bottom: gel image for two replicates at one restriction site (locus 1918694, restriction enzyme
XbaI). From left to right: wild-type during log phase, wild-type during stationary phase, dps-null
during log phase, dps-null during stationary phase.
86
Figure 3.2: Restriction enzyme loci for AvrII, NotI, and XbaI along the E. coli chromosome
Genomic information for restriction sites probed. From outermost to innermost: XbaI sites, NotI
sites, AvrII sites, and some genomic landmarks. On the inner circle, blue ticks denote rRNA genes
(rrnABCDEGH), red ticks denote replication termination regions (terABCD), and yellow indicates
the origin of replication (oriC).
oriC
terC
terD
terA
terB
rrnG
rrnE
rrnB
rrnA
rrnC
rrnD
rrnH
87
Figure 3.3: Nucleoid accessibility changes with growth phase and presence of dps
Log10 transformation of mean intensity ratio of dps-null populations to wild-type populations for
each restriction site during mid-log phase (A) and late stationary phase (B). Y axis ranges from -1
to .2 in each graph. Bars pointing into the circle denote greater accessibility to restriction enzyme
digestion in a dps-null strain; bars pointing out of the circle denote greater accessibility to
restriction enzyme digestion in a wild-type strain. Red bars indicate significance (p < .05, Student’s
t test). n = 6.
A B
88
Figure 3.4: Putative Dps DNA-binding motif
The 26 significantly differentially protected sites between wild-type and dps-null populations in
stationary phase were analyzed using MEME. The resulting 21bp motif, found in each of the 26
differentially protected sites is KGCMGMGAAASCGGCVGCWKM. Inverted dyad symmetry
indicated with black underline.
0
1
2
bits
1
A
C
T
G
2
C
T
G
3
A
C
4
T
A
C
5
C
T
G
6
T
A
C
7
A
C
T
G
8
G
T
A
9
G
C
T
A
10
G
C
A
11
A
C
G
12
A
G
C
13
T
A
G
14
T
C
G
15
A
G
C
16
T A
G
C
17
A
C
G
18
T
A
C
19
C
G
T
A
20
C A
T
G
21
C
A
89
Tables
Table 3.1: Consensus sequence in every stationary phase-specific differentially protected region
Locus Strand Distance p-val Motif
25152 - -2 2.84E-09 TGCTGACACAGCGGCCGCAGA
25681 - -531 2.84E-09 TGCTGACACAGCGGCCGCAGA
131428 + -975 1.30E-07 GGCAGATAAACCGGCGGATAC
497454 + -737 3.54E-07 GCCAGCGAAAACGGGCGCTGA
570508 - -158 3.21E-07 TGCCTGTAAAGCGGCGGCTGA
679192 + -456 1.36E-12 GGCAGCGAAAGCGGCGGCAGA
824657 - -157 4.96E-06 TGCCCTGAAGGCGGCGGCGAA
835259 - 915 8.91E-07 CGCCGTGGTAGCAGCAGCTGC
1088307 + -645 5.17E-07 AGCAGAAACAGCAGCAGCGGA
1100347 + -50 1.44E-07 GGCCGCGAGCGCGGCCCCTTC
1100949 + -652 1.44E-07 GGCCGCGAGCGCGGCCCCTTC
1173000 - 535 1.50E-06 TGAAGCAATACCTGCCGCAAA
1266386 - -47 1.06E-06 ACCAGAGAAACCAGCCGAATA
1374030 - 818 2.64E-07 GGCCGCCATACCTGCGCCGGC
1400533 + -642 1.10E-05 GGCAGAAACAGGGCACGCATA
1516944 - 758 2.64E-07 GGCGGTGAAAGCGGCGGTAGC
1536881 - -384 1.02E-05 GCAAGCTGAAGCGGCAACTGC
1768442 + 208 2.39E-08 TGCCGTTAAAGCTGCTGCTGA
2132544 - 943 6.81E-07 TGCCGCCAAACCGCCGCCCGC
2404317 - -803 2.88E-06 GTCTGCGTAAGCGGCTACATC
2468328 - -365 1.63E-06 TGCCGTGATAGCGGGACCACA
2695231 + -137 2.77E-05 ATCTGACAAACCTTCCCCAAA
2901855 + 588 8.91E-07 TGCAGATAAACCAGATGCATA
3545934 - -126 4.28E-07 CTCAGCGATAGCGCCGGCTTA
4200840 - -209 1.96E-07 CGCCGCGAAACAGGCCACAGA
4205175 + 23 6.67E-06 GGCCGCGTCACCATCAGATGA
90
Table 3.2: Some E. coli NAP DNA binding motifs and their occurrence in the wild-type genome
Nucleoid
associated
protein
DNA binding motif* Motif
stringency
Expected Observed Fold
difference
Dps KGCMGMGAAASCGGCVGCWKM
exact 0 1 ∞
1 mismatch 0 4 ∞
2 mismatches 1 11 11
3 mismatches 11 68 6.2
4 mismatches 87 485 5.57
5 mismatches 397 2,484 6.256
Fis KNNYRNNWNNYRNNM (218)
exact
35,477 67,754 1.9098
1 mismatch
496,682 555,744 1.11891
2 mismatches
2,980,090 2,046,698 0.6867904
3 mismatches
4,257,272 4,530,958 1.064287
4 mismatches
5,676,363 7,011,578 1.235224
5 mismatches
6,811,636 8,499,768 1.247831
IHF WATCAANNNNTTR (220)
exact
69 290 4.2
1 mismatch
2,217 4,756 2.145
2 mismatches
31,319 41,124 1.3130
3 mismatches
167,408 233,885 1.39709
4 mismatches
536,593 901,585 1.68020
5 mismatches
1,401,352 2,472,237 1.764180
*K = G/T
M = A/C
W = A/T
R = A/G
S = C/G
V = A/C/G
Y = C/T
N = A/C/G/T
91
Table 3.3: Putative Dps motif is overrepresented in stationary phase-specific significant sites
sample set Motif stringency expected observed overrepresentation
Stationary
phase-specific
exact 0 1 ∞
1 mismatch 0 2 ∞
2 mismatches 0 3 ∞
3 mismatches 0 15 ∞
4 mismatches 1 31 31
5 mismatches 5 74 14.8
Non-stationary
phase-specific
exact 0 0 —
1 mismatch 0 0 —
2 mismatches 0 0 —
3 mismatches 0 1 ∞
4 mismatches 2 6 3
5 mismatches 8 39 4.875
92
Table 3.4: Strains used in this study
Strain Genotype Reference
ZK126* W3110 DlacU169 tna-2 (214)
ZK1058 ZK126 dps::kan (1)
* ZK126 is the parent strain for ZK1058.
93
Table 3.5: Primers used in this study
Name Restriction
site
Restriction
enzyme
Sequence Direction Length GC
content
Tm
KMO142
25152 NotI
TTTGTCCTGTTGACCTCCGG Forward 20 55% 59.9C
KMO143 GGCCTTCGAGCGATACATCA Reverse 20 55% 60.0C
KMO49
25681 XbaI
GATGTATCGCTCGAAGGCCA Forward 20 55% 60.0C
KMO50 TTTCAGCCCCAACAGGTGTT Reverse 20 50% 60.0C
KMO93
131428 XbaI
GAATGCGGATTTGACGACGG Forward 20 55% 60.0C
KMO94 GCTCCTGTGAGACAGCAGTT Reverse 20 55% 60.0C
KMO136
168926 AvrII
GCTGACCGGTGTCGACTTTA Forward 20 55% 60.0C
KMO137 GCCACCTTCAACCGAGAAGA Reverse 20 55% 60.0C
KMO138*
224041
3320383
3361871
3494335
3588058
AvrII
GACTTAACAAACCGCCTGCG Forward 20 55% 60.1C
KMO139* GCAGGCCTAACACATGCAAG Reverse 20 55% 59.8C
KMO95
237191 XbaI
TGAGTTCAGAGAGCCGCAAG Forward 20 55% 60.0C
KMO96 CGCGGAACTCATCTCCATCA Reverse 20 55% 59.9C
KMO97
270729 XbaI
CCAATTCACGAACGTTCCCG Forward 20 55% 59.8C
KMO98 ACATCAAGAGCGCGGTTGTA Reverse 20 50% 60.0C
KMO51
497454 XbaI
TAATCACCGGTCATCGCCAG Forward 20 55% 59.9C
KMO52 TGGCTGAGTGAGAACTGCTG Reverse 20 55% 60.0C
KMO144
570508 NotI
TGTAGTTTCATCCGCTGCGT Forward 20 50% 60.0C
KMO145 TTGGGCGATGTGATTGTCCA Reverse 20 50% 60.0C
KMO146
679192 NotI
GCAGATGCTAAAGCGAAGGC Forward 20 55% 60.0C
KMO147 TTGATATCGGCCCCTGATGC Reverse 20 55% 60.0C
KMO53
752098 XbaI
ACCAGACCGGCATTTTACGT Forward 20 50% 60.0C
KMO54 GCCAGTGCTGAAAGACATGC Reverse 20 55% 60.1C
KMO55
824657 XbaI
AGAAGAGCTATGCGACTGCC Forward 20 55% 59.9C
KMO56 TTGGCATCAGCGACATCTGT Reverse 20 50% 60.0C
KMO148
835259 NotI
GCCGGATATGTCAGCCTACC Forward 20 60% 60.0C
KMO149 GACGGGAATGGTGTAAGCCA Reverse 20 55% 60.0C
KMO150
1051015 NotI
GTCCATGGCCGTACAACAGA Forward 20 55% 60.0C
KMO151 ATTACCGGCACTGACACAGG Reverse 20 55% 60.0C
KMO57
1088307 XbaI
GCTCGCGAACGTAACCAATC Forward 20 55% 60.0C
KMO58 AACCTTCGCTCAACGCAAAC Reverse 20 50% 60.0C
KMO59**
1100347 AvrII
CGGCCCCTTCACATACATCT Forward 20 55% 59.5C
KMO60**
1100830 XbaI
CCTTGATTTTGGCTGCGGAA Reverse 20 50% 59.4C
KMO99
1100949 XbaI
TTCCGCAGCCAAAATCAAGG Forward 20 50% 59.4C
KMO100 CGGAGAGCTGGAGTCAACAG Reverse 20 60% 60.1C
94
Name Restriction
site
Restriction
enzyme
Sequence Direction Length GC
content
Tm
KMO61
1103016 XbaI
CCCAGCTGCATATGAGCGTA Forward 20 55% 60.0C
KMO62 TTGACTGGGGAGAGGGTTCA Reverse 20 55% 60.1C
KMO152
1173000 NotI
TGGGTTGTGCCTCTTTGGTT Forward 20 50% 60.0C
KMO153 GGCTGGTATGCCGATTACGA Reverse 20 55% 60.0C
KMO154
1266386 NotI
GGTTGGCAGTTTGGCTTCTG Forward 20 55% 59.9C
KMO155 GGTCGTGGTTCCTGGACTTT Reverse 20 55% 60.0C
KMO63
1374030 XbaI
GGAATGGAACTGGCCAGACA Forward 20 55% 60.0C
KMO64 TACCTGCATTAGTGGCGCTC Reverse 20 55% 60.2C
KMO101
1385641 XbaI
CGGGCTCGTGCATTGTATTG Forward 20 55% 60.0C
KMO102 GCCTTGCTCTTCGCGTAATG Reverse 20 55% 60.0C
KMO103
1400533 XbaI
TTGGGTTGTCCGGTTTTGGA Forward 20 50% 60.0C
KMO104 TAAGTCGTGCGCCATTGACT Reverse 20 50% 60.0C
KMO156
1516944 NotI
GCCCTTGCAACATATCGCAG Forward 20 55% 60.0C
KMO157 AGAGTGGCTTTGGTCGTGAG Reverse 20 55% 60.0C
KMO65
1536881 AvrII
ACCCTAAAGTGGTTCCCTGC Forward 20 55% 59.6C
KMO66 GTGGTCTGATCCAGCGTTGA Reverse 20 55% 60.0C
KMO105
1650286 XbaI
CCCCGGTAACTTCCCCTTTC Forward 20 60% 60.0C
KMO106 ACGCGGTAATACAAGGTGGG Reverse 20 55% 60.1C
KMO107
1662509 XbaI
GTTGTGCCGCAGCTTGTTAA Forward 20 50% 60.0C
KMO108 GCGCCCAGTACCAGTTTAGT Reverse 20 55% 60.0C
KMO158
1768442 NotI
GTTGTCGCTGAAGCAACTGG Forward 20 55% 60.0C
KMO159 GAAGTGCAAACTTCGCCGTT Reverse 20 50% 60.0C
KMO69
1877372 XbaI
AAGCAACATACGGCGAATGC Forward 20 50% 59.9C
KMO70 ATGCTGCAAACGCGCATTAA Reverse 20 45% 60.1C
KMO160
1902833 NotI
GTCTCCTGCCTGAAAGCGAT Forward 20 55% 60.1C
KMO161 AAATGCAGTTACTGGTGCGC Reverse 20 50% 59.8C
KMO71
1918694 XbaI
CCAGTCGCTGAAACCTTTGC Forward 20 55% 60.0C
KMO72 TTACGACCACAGACGCAACA Reverse 20 50% 59.9C
KMO162
1943430 NotI
CCATTTGCGCATTCCCAGAC Forward 20 55% 60.2C
KMO163 CGTAGATTTTGGCGGGGAGA Reverse 20 55% 59.8C
KMO109
2012752 XbaI
CTCTGGACTACTCCACCCCA Forward 20 60% 60.0C
KMO110 TTTGAAAGGCTGATGGGGCT Reverse 20 50% 59.9C
KMO111
2061576 XbaI
CTCGACTGGAGCCGGATTAC Forward 20 60% 60.0C
KMO112 GCCTTCATTGCCCATTTGCA Reverse 20 50% 60.0C
KMO113
2084127 XbaI
CACTTTGCTCACCACATCGC
Forward 20 55% 60.1C
KMO114 CGTGCGCTCAGTCATGATTG
Reverse 20 55% 60.0C
KMO73
2132544 XbaI
CAACAATCATCGCGCTACCG Forward 20 55% 60.0C
KMO74 ACACCGGCTACCTGCTTTAC Reverse 20 55% 60.0C
95
Name Restriction
site
Restriction
enzyme
Sequence Direction Length GC
content
Tm
KMO164
2194742 NotI
GGTCAACTTTCACCACCCCA Forward 20 55% 60.1C
KMO165 GCTGCATTTACGTTGGGCAA Reverse 20 50% 60.0C
KMO166
2210185 NotI
GGAAGCCCGCTTTAAACAGC Forward 20 55% 60.1C
KMO167 CAGCAGGAAGAGCAAGGTGA Reverse 20 50% 60.0C
KMO168
2404317 NotI
GTCGTGAATCTATCGCCCGT Forward 20 55% 60.0C
KMO169 CGGGAGAGAACATAGGTGCC Reverse 20 60% 59.9C
KMO170
2419279 NotI
TGGTCAATCGTCCGAAGCAA Forward 20 50% 60.0C
KMO171 TACGCGTTAAAGAGCCGGTT Reverse 20 50% 60.0C
KMO75
2468328 XbaI
CAGGATTACGCCGACGCTAT Forward 20 55% 60.0C
KMO76 ATCTGGCGTGTGTTCTTCGT Reverse 20 50% 60.0C
KMO115
2623625 XbaI
CGAAACAACGCGTAGTTCCC Forward 20 55% 59.8C
KMO116 CGACGTAGCAGAGCTAAGGG Reverse 20 60% 60.0C
KMO140*
2637634
3318964
3360366
3586639
AvrII
ACCGACGCTTATCGCAGATT Forward 20 50% 59.9C
KMO141* TGGGAGTGGGTTGCAAAAGA Reverse 20 50% 59.7C
KMO117
2637706 XbaI
ATTGAGGTCGGCGACTTTCA Forward 20 50% 59.7C
KMO118 GGATCAGAATGCCACGGTGA Reverse 20 55% 60.1C
KMO119
2666693 XbaI
CTCGCATTCTGCCTAGCTCA Forward 20 55% 59.9C
KMO120 TTTGGGAAAGGTGGCATCGT Reverse 20 50% 60.2C
KMO172
2680953 NotI
GCACCGTCGGCATATTGTTC Forward 20 55% 60.0C
KMO173 TGGTGAATATCGTCTCGGCG Reverse 20 55% 60.0C
KMO77
2695231 XbaI
CATTACAGCGCCAGACAGGA Forward 20 55% 60.1C
KMO78 CTTCGGGAGTTTCTGGTCCG Reverse 20 60% 60.4C
KMO79
2901855 XbaI
ATGTGCCTAACCCGCTCAAA Forward 20 50% 60.0C
KMO80 AAAAGGCGCCATCCATGTTG Reverse 20 50% 59.8C
KMO121
3319036 XbaI
GATGTTTCAGTTCCCCCGGT Forward 20 55% 60.0C
KMO122 AAAGTGAAAAGCAAGGCGTC Reverse 20 45% 55C
KMO121
3360438 XbaI
GATGTTTCAGTTCCCCCGGT Forward 20 55% 60.0C
KMO123 GAAAAGCAAGGCGTTTACGC Reverse 20 50% 56C
KMO174
3474718 NotI
AACTGACCTTACACGGCGAG Forward 20 55% 60.0C
KMO175 CCCTTGTTGATAGGCGCTGA Reverse 20 55% 60.1C
KMO176
3510792 NotI
AAGGCATTTCCGGTCTCCTG Forward 20 55% 60.0C
KMO177 GCGACCTGACAGCGAAATTC Reverse 20 55% 59.9C
KMO178
3545934 NotI
GTGGATTCACCGGCACAGTA
Forward 20 55% 60.0C
KMO179 GTGTTAAAAGCGTGCTGGGG
Reverse 20 55% 60.0C
KMO124
3586711 XbaI
CCTTTCCAGACGCTTCCACT Forward 20 55% 60.0C
KMO125 AGGAAGTGAAAAGCAAGGCG Reverse 20 50% 57C
KMO126
3725152 XbaI
TGAAGCGCTGGCAGAAAAAC Forward 20 50% 60.0C
KMO127 AGACCTCGCCGTTCAGTTTT Reverse 20 50% 59.9C
96
Name Restriction
site
Restriction
enzyme
Sequence Direction Length GC
content
Tm
KMO81
3733535 AvrII
GAATGCTGCGTGCCCATATG Forward 20 55% 60.0C
KMO82 GCAACCCTACTCCTGTTCCC Reverse 20 60% 60.0C
KMO180
3754916 NotI
GCTCGGCATATTTGTCTGCG Forward 20 55% 60.0C
KMO181 GGCCGATAAAGCCGACGATA Reverse 20 55% 60.0C
KMO128
3832553 XbaI
GCAACAGGTGGAGAAGTCGA Forward 20 55% 60.0C
KMO129 CGTCATATAGCCGCCTGTGT Reverse 20 55% 60.0C
KMO83
3951009 XbaI
TGAGCGGTGCTTCAAATGGA Forward 20 50% 60.3C
KMO84 AAACAGCGGGCATAGCGATA Reverse 20 50% 59.9C
KMO85
4100219 XbaI
TTCAGGGATCGACAGGCAAC Forward 20 55% 60.0C
KMO86 ACTGTTTTGCCACCCAGTGA Reverse 20 55% 60.0C
KMO182
4200840 NotI
CAACTAATGCATCCACGCCG Forward 20 55% 60.0C
KMO183 CTGTTAGTGGTTGCGTTGGC Reverse 20 55% 60.0C
KMO184
4205175 NotI
TGTCATGACGGCGGATTTCA Forward 20 50% 60.0C
KMO185 GCTTGCCGGTATGGGTTTTG Reverse 20 55% 60.1C
KMO87
4407169 XbaI
TGGATGTTGGCCTTTCCACA Forward 20 50% 59.8C
KMO88 GGAATGGAACTGGCCAGACA Reverse 20 55% 60.0C
KMO89**
4473491 AvrII
GTCAGGGCGCATATCTTCCA Forward 20 55% 59.9C
KMO90
4474052 XbaI
TCCTGGCAACGAATCTGGTC Reverse 20 55% 60.0C
KMO91
4480107 XbaI
ATTTCGTGCTGAGGGGGATG Forward 20 55% 60.1C
KMO92 AGCCGGCAAGCTTTGATTTG Reverse 20 50% 60.0C
* Because a small number of restriction sites fell in rRNA genes, 5 AvrII sites (224041, 3320383,
3361871, 3494335, 3588058) were probed using one set of primers (KMO138+KMO139) and 4
AvrII sites (2637634, 3318964, 3360366, 3586639) were probed using another set of primers
(KMO140+KMO141), without a way to deconvolute variation in sites probed by the same primer
pair.
** A few restriction sites for different enzymes were within a few hundred nucleotides from one
another; these genetically close loci were probed using the same pair of primers.
97
Chapter 4: MNase-Seq reveals E. coli stationary phase genome accessibility
98
Introduction
When incubated under laboratory conditions, Escherichia coli populations exhibit 5 phases
of growth and survival (211). For the first few hours post-inoculation, population density remains
low during lag phase. After lag phase, the population enters log phase, during which time
population density increases exponentially due to rapid growth and cell division. By day 1, the
population enters stationary phase, where cells maintain population density gained during log
phase through day 2. Then, during death phase, approximately 99% of the population dies for
reasons that are not well understood, after which the population strikes a balance between growth
and death during long-term stationary phase, which can persist for years without the addition of
new nutrients.
Population density is not the only measurable change across growth phases. The structure
and protein composition of the nucleoid, the portion of the prokaryotic cell wherein the
chromosome and its associated proteins reside, also changes (19, 24, 25). During log phase, the E.
coli nucleoid is interspersed with translation machinery and the major nucleoid-associated proteins
(NAPs) are Fis, H-NS, HU, and Hfq. During the transition from log phase to stationary phase,
toroidal (ring-shaped) areas of the nucleoid form that are regularly spaced and devoid of
ribosomes. By late stationary phase, the nucleoid is entirely segregated from ribosomes and is
regularly spaced. This late stationary phase nucleoid structure, called “the biocrystal,” is a 3-D
hexacrystalline array composed of DNA and protein. The most notable protein herein, which is
necessary in vivo (24) and sufficient in vitro (25) to form the biocrystal, is the DNA-binding protein
from starved cells, Dps.
Dps is a universally conserved bacterial ferritin and DNA-binding protein that, in E. coli,
is one of the most highly expressed proteins during stationary phase (1, 18). DNA protection is at
99
the heart of Dps function, as the protein protects DNA from nucleases and its ferritin activity has
been shown to confer DNA protection (1, 10). Additionally, E. coli populations that lack dps
exhibit altered growth and survival patterns compared to wild-type populations (see Chapter 2)
(62). Though Dps is highly expressed and important in conferring wild-type growth and survival,
there is no field-wide agreement on DNA sequence or structural specificity for DNA binding
activity. However, some level of targeted binding has been observed (see Chapter 3), suggesting
an underlying sequence or structural motif may be present that determines DNA binding.
This chapter explores which regions of the E. coli chromosome are bound by Dps during
stationary phase. By investigating two stationary phase timepoints, one in early stationary phase
and one in late stationary phase, Dps DNA-binding regions were revealed. These regions were
assessed for commonalities, and a putative Dps DNA-binding motif was revealed.
100
Results
Experimental Design
To determine where Dps binds on the E. coli chromosome, micrococcal nuclease (MNase)
digestion followed by DNA sequencing (MNase-Seq) was performed in duplicate in two strains
(wild-type and dps-null) at two timepoints (early stationary phase and late stationary phase).
MNase is a nonspecific endo-/exo-nuclease. Briefly, crosslinked chromatin was treated with
MNase, de-crosslinked, and the remaining DNA was sent for Illumina sequencing (Figure 4.1).
In this experiment, chromatin would be fixed in its state at harvest, and MNase would digest DNA
not bound by proteins, leaving just DNA that is protein-bound following digestion and de-
crosslinking. Sequencing would then reveal which regions of the chromosome are involved in
DNA-protein interactions. Comparing wild-type populations to dps-null populations would then
reveal which genomic regions may be bound by Dps.
101
Calling peaks
MACS2 was used to determine chromosomal regions overrepresented in the sequencing
dataset (“peaks”) of at least 50 bp in length. This was done for each strain (wild-type or dps-null)
and growth phase (early stationary phase or late stationary phase). Because sequencing was
performed in duplicate, peak sets were combined by strain and timepoint. These peaks are
distributed across the genome in each sample, though the number of peaks varies (Figure 4.2). In
early stationary phase, the wild-type sample has 1927 peaks and the dps-null sample has 2528
peaks. In late stationary phase, the wild-type sample has 308 peaks and the dps-null sample has
1636 peaks. Additionally, different proportions of the chromosome are represented in each
sample’s peak set. 6.6% of the chromosome is represented in peaks in the wild-type strain during
early stationary phase, 12% in the dps-null strain during early stationary phase, 1.6% in the wild-
type strain during late stationary phase, and 6.2% in the dps-null strain during late stationary phase
(Figure 4.3). The difference in protection extends to the length of peaks. Though the majority of
peaks are less than 500bp in length for all samples, the proportion of peaks over 500bp in length
vary by strain and timepoint (Figure 4.4, Table 4.1). During early stationary phase, 1.4% of wild-
type peaks and 6.3% of dps-null peaks are larger than 500bp. During late stationary phase, 10.1%
and 3.6% of peaks are over 500bp in length in wild-type and dps-null strains, respectively. This
disparity expands when evaluating the number of peaks over 1000bp in length. During early
stationary phase, .05% of wild-type peaks are larger than 1000bp and .36% of dps-null peaks are
larger than 1000bp. In late stationary phase, this has grown to 1.3% in the wild-type and remained
relatively constant at .43% in the dps-null strain.
102
Motif analysis
106 peaks are shared between the early stationary phase and late stationary phase wild-type
populations. Shared peak sequences were analyzed using MEME, a motif search tool (217). One
motif was determined with statistical significance. The discerned motif, present in every shared
peak but one, is TTTTTTCDSCWDYW (Figure 4.5). No significant similarity (q < .05) to known
prokaryotic motifs was found when prokaryotic motif databases were queried using Tomtom
(219). This motif is overrepresented in the wild-type (ZK126) E. coli genome. However, the extent
to which it is overrepresented varies with motif stringency (the number of mismatches to the motif
allowed) (Table 4.2). By decreasing motif stringency from exact to 5 mismatches, the number of
expected motif occurrences increases from two (exact) to 104 (one mismatch) to 2,038 (two
mismatches) to 18,837 (three mismatches) to 103,959 (four mismatches) to 400,299 (five
mismatches). The corresponding observed number of motif occurrences are one; 777; 6,221;
34,736; 144,222; and 464,448. These lead to underrepresentation of the motif by a factor of two
for an exact match in the genome and overrepresentation of the motif by 7.5-fold, 3.1-fold, 1.8-
fold, 1.4-fold, and 1.2-fold as motif stringency decreases. This amount of overrepresentation is
similar to the motifs for major NAPs Fis, to which Dps is often compared, and IHF, the other major
NAP during early stationary phase (Table 4.2). However, the motif for Dps occurs less frequently
at every motif stringency tested than those for either Fis or IHF.
This putative Dps DNA-binding motif is also overrepresented in each sample’s peak set
(wild-type or dps-null during early or late stationary phase) (Table 4.3). However, it is
overrepresented to a greater extent in the wild-type population during each timepoint with the most
overrepresentation in the wild-type population during late stationary phase. Wild-type peak sets
display, on average, 3.7X overrepresentation of this motif during early stationary phase and 5.8X
103
overrepresentation of this motif during late stationary phase. This is larger than the
overrepresentation observed for dps-null populations: 2.4X and 3.1X overrepresentation of this
motif during early and late stationary phase, respectively. This disparity is not due to bias in the
raw data: each sample’s sequencing data displays approximately the same overrepresentation of
this motif without statistical significance (Table 4.4). Briefly, in the wild-type early stationary
phase sequencing dataset, this motif is overrepresented by a factor of 3.2 and 2.6 in each replicate;
in dps-null during early stationary phase, this motif is overrepresented by a factor of 3.1 and 3.2;
in wild-type during late stationary phase, this motif is overrepresented by 3.2X apiece; and in dps-
null during late stationary phase, this motif is overrepresented by a factor of 3.4 and 2.8.
104
Discussion
Nucleoids from two strains (wild-type or dps-null) at two timepoints (early stationary phase
or late stationary phase) were crosslinked and treated with MNase, the remaining DNA was
sequenced, and overrepresented regions (‘peaks’) were determined for each of the four groups.
One interpretation of the results is that the peaks present in wild-type samples, but absent from
dps-null samples are overrepresented because of direct Dps binding of those regions of the
chromosome. Another interpretation is that other, non-Dps NAPs act differently when Dps is
absent, causing certain regions of the chromosome to be differentially protected in the presence or
absence of Dps. By focusing on peaks shared between the wild-type timepoints, the second
interpretation may be circumvented. It is important to note that some of these shared peaks are also
present in the dps-null peak sets. However, because deleting a NAP in a strain inherently means
working with a mutant nucleoid, these peaks were determined to be biologically relevant in the
wild-type nucleoid and thus the focus of further investigation.
The late stationary phase wild-type population displayed fewer peaks than the other three
samples. This is expected due to the regular structure and spacing of the late stationary phase
nucleoid. Because the biocrystal is organized into a hexacrystalline array, fewer regions of the
chromosome will be significantly overrepresented in the sequencing dataset. Notably, the late
stationary phase wild-type population had a larger proportion of large (>1000bp) peaks than the
other three samples. This follows from the logic of the previous statement: because the wild-type
nucleoid (the Dps-dependent biocrystal) is so regularly spaced during late stationary phase, larger
stretches of DNA are likely to be protected from degradation, and hence be included in peaks.
While many peaks were unique to strain and time, there is substantial overlap among
populations. Within the early stationary phase timepoint, 1119 peaks (58% of wild-type and 44%
105
of dps-null) are shared. This large overlap is expected, as Dps does not begin to bind substantial
portions of the chromosome until stationary phase. Thus, during early stationary phase, before Dps
has had ample time to restructure the nucleoid, there should still be overlap between wild-type and
dps-null populations. During late stationary phase, 152 peaks (50% of wild-type and 9% of dps-
null) are shared. This smaller overlap is also expected, as Dps has had the entirety of stationary
phase to restructure the nucleoid in the wild-type population, a phenomenon lacking in the dps-
null population.
In addition to overlap between strains within one timepoint, the same strain at different
timepoints also shared many peaks. 106 peaks were shared between the wild-type populations at
early and late stationary phase (6% of early stationary phase and 35% of late stationary phase);
831 peaks were shared between the dps-null populations at early and late stationary phase (33%
of early stationary phase and 51% of late stationary phase). These data are not unexpected when
considering how the nucleoid is restructured throughout stationary phase in wild-type and dps-null
strains. In wild-type nucleoids, the process that initiates during early stationary phase via toroids
continues through late stationary phase in formation of the biocrystal. In dps-null nucleoids, the
process that initiates during early stationary phase continues through late stationary phase in
formation of cholesteric phase chromatin.
The peaks shared between the wild-type populations at early and late stationary phase are
potential Dps binding regions and may even be biocrystal nucleation points. To this end, the DNA
sequences of these peaks were searched for motifs using MEME. The resulting motif, present in
all but one of the shared peaks between wild-type timepoints, is TTTTTTCDSCWDYW. This
motif is different than the motif presented in Chapter 3; these differences will be discussed in
Chapter 5. Notably, no significant similarity to known prokaryotic motifs was found when
106
compared using Tomtom (219). This suggests this motif is unique to Dps. Additionally, this motif
is observed more frequently than it is expected to occur in the wild-type E. coli genome, suggesting
a benefit to having this motif. While this motif is overrepresented in all peak sets defined in this
chapter, it is overrepresented to its greatest extent in the late stationary phase wild-type population,
followed by the early stationary phase wild-type population (Table 4.3). The overrepresentation
observed in the dps-null peak sets is likely due to its overrepresentation across the genome, and
not a biologically-driven phenomenon that functions regardless of dps expression. Moreover, this
motif is significantly more overrepresented in the wild-type peak set during late stationary phase
than in the corresponding dps-null peak set. This all suggests a benefit for wild-type stationary
phase nucleoid structure, particularly during late stationary phase. The benefit in this case is
presumably to direct Dps binding, which results in the wild-type late stationary nucleoid structure
and protects DNA from damaging agents.
The overrepresentation of the motif elucidated through this study is compelling when
compared to DNA binding motif overrepresentation of Fis and IHF (Table 4.2). Though extent of
overrepresentation varies as motif stringency (number of mismatches in motif searched) varies for
each of these motifs, average overrepresentation of this putative Dps DNA-binding motif is similar
to that of the published IHF DNA-binding motif: 2.3X and 2.1X, respectively. This suggests
possible biological relevance of the motif presented in this chapter. However, the number of motif
occurrences in the wild-type E. coli genome is lower for Dps than one might expect, given that
Dps has been shown to bind up to 50% of the chromosome during late stationary phase (22).
Assuming five mismatches between this motif and the chromosome, this motif occurs 10X less
frequently than the motifs for either Fis or IHF with 5 mismatches (Table 4.2), which bind
substantially smaller portions of the E. coli nucleoid than Dps does (22). However, this may be
107
explained by the specificity of these motifs: both the Fis and IHF motifs contain ‘N’s, whereas this
putative Dps motif does not. Perhaps further study can determine which nucleotides in this motif,
if any, are dispensable for Dps binding.
The consensus sequence presented here is derived from peaks shared between wild-type
nucleoids in early stationary phase and late stationary phase. Because these peaks are formed early
in the timeline of Dps DNA binding, one model supposes that these peaks may be anchoring points
for early stationary phase toroid formation. This model also explains the lower-than-expected
frequency of this motif in the wild-type genome. Assuming a Dps-DNA binding scheme predicated
on partial programmed binding (perhaps directed by this motif) and partial nonspecific binding,
the motif presented here might point to biocrystal nucleation points.
This study presents a putative Dps consensus sequence. To confirm that the motif presented
here does direct Dps binding, biochemical and genetic tests can be used. Comparing Dps binding
affinity for the peaks defined in this dataset or other DNA that contain this chapter’s Dps DNA-
binding motif against segments of DNA that do not contain this motif could prove biochemical
preference. Genetically, editing the endogenous shared wild-type peaks to remove the putative Dps
DNA-binding motif and assessing changes in accessibility either through the PCR-based approach
described in Chapter 3 or using MNase-Seq may reveal the impetus for Dps DNA-binding.
Similarly, changing the genetic context of the shared wild-type peaks defined by this MNase-Seq
data by engineering these peaks into plasmids and transforming those plasmids into wild-type and
dps-null E. coli strains might provide a way to examine changes in accessibility using Chapter 3’s
PCR-based method.
108
Conclusion
Crosslinked wild-type or dps-null nucleoids during early or late stationary phase were
treated with MNase, which left only DNA protected by proteins. This DNA was sequenced, and
peaks were called using MACS2. Peaks shared between the wild-type populations during early
stationary phase and late stationary phase, likely points of Dps binding, were searched for a motif.
The resulting motif, TTTTTTCDSCWDYW, is overrepresented in the wild-type E. coli genome
and MNase-Seq peak sets and unique among known prokaryotic motifs, suggesting this may be
the motif for driving DNA binding by Dps.
109
Figures
Figure 4.1: MNase-Seq experimental design
MNase-Seq was performed on four population types in duplicate: wild-type or dps-null during
early stationary phase or late stationary phase. Crosslinked chromatin (top) was digested with
MNase. MNase treatment resulted in digestion of DNA not bound by proteins, leaving only bound
nucleoprotein complexes (middle). After MNase treatment, remaining nucleoprotein complexes
were de-crosslinked, leaving only DNA that was part of nucleoprotein complexes; this DNA was
sent for sequencing (bottom). Yellow, orange, and blue masses represent different NAPs. Figure
made using biorender.com.
110
Figure 4.2: Peaks vary by sample and are distributed across the genome
Peaks were called using MACS2 by strain and timepoint. Y-axis shows -log10(p value); x-axis
shows chromosomal position. From top to bottom: wild-type during early stationary phase, dps-
null during early stationary phase, wild-type during late stationary phase, and dps-null during late
stationary phase.
0
10
20
30
0e+00 1e+06 2e+06 3e+06 4e+06
wild-type early stationary phase
0
5
10
15
20
25
0e+00 1e+06 2e+06 3e+06 4e+06
dps-null early stationary phase
0
10
20
30
40
0e+00 1e+06 2e+06 3e+06 4e+06
-log(pval)
wild-type late stationary phase
0
10
20
30
40
0e+00 1e+06 2e+06 3e+06 4e+06
Position (bp)
dps-null late stationary phase
111
Figure 4.3: Populations reveal differential inclusion in peak sets
Different proportions of the genome are represented in peaks for different samples. From left to
right: wild-type during early stationary phase, dps-null during early stationary phase, wild-type
during late stationary phase, and dps-null during late stationary phase.
0.0
2.5
5.0
7.5
10.0
12.5
Percent of chromosome in peaks
wild-type dps-null wild-type dps-null
early stationary phase late stationary phase
112
Figure 4.4: Peak lengths vary by strain and timepoint
0.0
2.5
5.0
7.5
10.0
Percent
Peak Size
>2000
1501-2000
1001-1500
901-1000
801-900
701-800
601-700
501-600
wild-type dps-null wild-type dps-null
early stationary phase late stationary phase
0
25
50
75
100
Percent
Peak Size
>2000
1501-2000
1001-1500
901-1000
801-900
701-800
601-700
501-600
wild-type dps-null wild-type dps-null
early stationary phase late stationary phase
113
Peak length distribution was examined by strain and time. Top: composition of total peak sets of
peaks over 500bp in length. Bottom: composition of peaks over 500bp, normalized to total peaks
longer than 500bp. From left to right: wild-type during early stationary phase, dps-null during early
stationary phase, wild-type during late stationary phase, and dps-null during late stationary phase.
Peak sizes are color coded from pastel green (501-600 bp) to blue (>2000 bp).
114
Figure 4.5: Putative Dps DNA-binding motif
Peaks shared between the wild-type 8-hour and 48-hour samples were searched for gapped motifs
using MEME (217). The resulting motif (TTTTTTCDSCWDYW) is present in all but one peak
shared between the two wild-type time points.
115
Tables
Table 4.1: Peak length varies by sample
peak length
(bp)
percent of peaks
early stationary phase late stationary phase
wild-type dps-null wild-type dps-null
0-50 0 0 0 0
51-100
35.6 22.9 25.6 38.4
101-200 40.9 38.4 32.1 34.8
201-300
14.6 19.5 20.1 14.9
301-400
6.0 8.1 8.1 5.8
401-500
1.5 4.9 3.9 2.6
501-600 0.78 2.3 4.2 1.2
601-700
0.52 1.3 2.6 0.73
701-800
0.05 1.0 1.3 0.43
801-900
0 0.79 0.32 0.55
901-1000 0 0.47 0.32 0.24
1001-1500
0.05 0.24 0.65 0.37
1501-2000
0 0.08 0.32 0.06
>2000
0 0.04 0.32 0
116
Table 4.2: Putative motif occurrences in the wild-type E. coli genome
motif mismatch expected observed fold
overrepresentation
TTTTTTCDSCWDYW
0 2 1 0.5
1 104 777 7.5
2 2,038 6,221 3.1
3 18,837 34,736 1.8
4 103,959 144,222 1.9
5 400,299 464,448 1.2
Fis
KNNYRNNWNNYRNNM
(218)
0 35,477 67,754 1.9
1 496,682 555,744 1.1
2 2,980,090 2,046,698 0.7
3 4,257,272 4,530,958 1.1
4 5,676,363 7,011,578 1.2
5 6,811,636 8,499,768 1.2
IHF
WATCAANNNNTTR
(220)
0 69 290 4.2
1 2,217 4,756 2.1
2 31,319 41,124 1.3
3 167,408 233,885 1.4
4 536,593 901,585 1.7
5 1,401,352 2,472,237 1.8
117
Table 4.3: Putative motif occurrences in peak sets
timepoint strain # mismatches expected observed fold
overrepresentation
average
overrepresentation
early
stationary
phase
wild-type
0 0 4 ∞
3.7
1 7 78 11
2 134 448 3.3
3 1,243 2,160 1.7
4 6,862 8,565 1.2
5 26,422 27,349 1.0
dps-null
0 0 3 ∞
2.4
1 12 75 6.3
2 244 590 2.4
3 2,252 3,227 1.4
4 12,429 13,482 1.1
5 47,857 44,603 0.9
late
stationary
phase
wild-type*
0 0 1 ∞
5.8
1 2 29 14
2 33 224 6.8
3 305 1,077 3.5
4 1,681 4,043 2.4
5 6,474 11,947 1.8
dps-null
0 0 4 ∞
3.1
1 6 41 6.8
2 126 506 4.0
3 1,169 2,465 2.1
4 6,454 9,728 1.5
5 24,851 30,093 1.2
* The putative Dps-binding motif is significantly more overrepresented in the wild-type peak set
during late stationary phase than in the dps-null peak set during stationary phase (p=.05, Student’s
T test).
118
Table 4.4: Putative motif occurrences in raw data
timepoint strain replicate # mismatches expected* observed* fold
overrepresentation
**
average
overrepresentation
**
early
stationary
phase
wild-
type
A
0 0 1 ∞
3.2
1 4 33 8.3
2 83 253 3.0
3 766 1457 1.90
4 4,230 5,983 1.414
5 16,289 18,999 1.1664
B
0 0 1 ∞
2.6
1 4 23 5.8
2 74 220 3.0
3 685 1,275 1.86
4 3,783 5,155 1.363
5 14,566 16,521 1.1342
dps-
null
A
0 0 3 ∞
3.1
1 4 33 8.3
2 85 259 3.0
3 785 1,436 1.83
4 4,332 5,814 1.342
5 16,682 18,610 1.1156
B
0 0 1 ∞
3.2
1 5 43 8.6
2 93 266 2.9
3 857 1,490 1.74
4 4,728 6,686 1.414
5 18,206 21,445 1.1779
late
stationary
phase
wild-
type
A
0 0 2 ∞
3.2
1 5 40 8
2 104 336 3.23
3 964 1,893 1.963
4 5,320 7,612 1.431
5 20,483 24,185 1.1807
B
0 0 5 ∞
3.2
1 6 51 8.5
2 122 392 3.21
3 1,131 2,089 1.847
4 6,242 8,860 1.419
119
timepoint strain replicate # mismatches expected* observed* fold
overrepresentation
**
average
overrepresentation
**
5 24,034 28,438 1.1832
late
stationary
phase
dps-
null
A
0 0 2 ∞
3.4
1 5 45 9
2 100 314 3.14
3 927 1,807 1.95
4 5,114 7,600 1.486
5 19,692 23,758 1.2065
B
0 0 1 ∞
2.8
1 4 27 7
2 76 218 2.9
3 703 1,293 1.84
4 3,882 5,407 1.393
5 14,947 17,293 1.1570
* Analysis was performed on a set of 1000 randomly sampled DNA sequences for each sample.
** No statistical significance was determined (Student’s T test).
120
Methods
Bacterial culture
Two E. coli K-12 lineage W3110 strains were used in this study: ZK126 (F
-
, D(argF-
lac)169, l
-
, IN(rrnD-rrnE)1, rph-1), referred to here as ‘wild-type’ (214), and ZK1058, a dps::kan
ZK126 (1). Initial cultures of 5 ml LB liquid medium (Difco) in 23 ml borosilicate test tubes
(Thermo Fischer) were inoculated from frozen LB-glycerol stocks (-80°C) containing the strain of
interest, incubated aerobically at 37°C and 70% humidity on a TC-7 roller drum (New Brunswick
Scientific), and incubated overnight. These overnight cultures were then used to inoculate 5ml
fresh LB liquid medium (1:1000) in 23 ml borosilicate test tubes and cultured under the
aforementioned conditions. Cultures were harvested at either 8 hours (for the early stationary phase
time point) or 48 hours (for the late stationary phase time point).
Nucleoid preparation
Nucleoids were prepared for MNase treatment as previously described by Marbouty and
colleagues (223), with a few modifications. Briefly, 10
9
cells from each culture were resuspended
in 100 µl TE (pH 7.5). Formaldehyde was added to a final concentration of 1%; samples were
incubated at room temperature for 30 minutes and then 4°C for 30 minutes. Formaldehyde was
quenched upon addition of glycine at a final concentration of .25M. Samples were incubated at
room temperature for 5 minutes, followed by a 15-minute 4°C incubation. Samples were pelleted
and resuspended in 25 µl TES (pH 7.5). 250 U Ready-Lyse lysozyme (Lucigen) in TES (ph 7.5)
was added to the samples, which were then incubated at room temperature with nutating for 15
minutes (follow manufacturer’s instructions). Lysozyme was quenched with final concentration of
0.5% SDS; samples were incubated at room temperature for 10 minutes. Prepped nucleoids were
stored in 5µl aliquots at -20°C until further use, as applicable.
121
Micrococcal nuclease treatment and DNA collection
To digest prepped nucleoids, 2,000 gel units micrococcal nuclease (NEB), BSA (NEB) to
final concentration .1 mg/ml, MNase buffer (NEB) to final concentration 1X, Triton X-100 to final
concentration 1%, and nuclease-free water were added to 5 µl of prepared nucleoid samples to
final volume 50 µl. Samples were incubated at 37°C for 24 hours. To de-crosslink, 250 µg/ml
proteinase K (VWR) in 6.2 mM EDTA was added to the reaction, and samples were incubated at
65°C overnight (follow manufacturer’s instructions). DNA was collected using ethanol
precipitation.
Sequencing
DNA from MNase treatment was sent to the Microbial Genome Sequencing Center
(MiGS). MiGS performed Illumina DNA library preparation, followed by 150 bp paired-end
sequencing on the NextSeq 2000 platform.
Alignment, peak calling, and differential analysis of sequencing data
Analysis of sequencing data was performed in R. First, three samples (wild-type early
stationary phase B, dps-null early stationary phase A, and dps-null late stationary phase B) had
approximately 1.5X the reads of the remaining five samples (3,000,000 reads compared to
2,000,000 reads), so these three were randomly sampled to yield comparable data to the other five.
Sampling was done as follows: 2,000,000 numbers between 1 and the number of reads for each of
the three samples with higher coverage were acquired using a random number generator, and the
reads corresponding to those numbers were parsed into new fastq files for each of the paired-end
fastq files. fastq files (either provided by MiGS or those that were randomly sampled) were aligned
using Bowtie2 and quality was assessed (measurements included duplicate rate, mitochondria
mapping rate, proper pair rate, unmapped rate, has unmapped mate rate, not passing QC rate, non-
122
redundant fraction, and PCR bottlenecking). Based on these measurements, quality was
determined to be good for five samples, and no further filtering was needed for those. For the
remaining three samples (those that were randomly sampled), duplication rate was higher than
preferred, so duplicate reads were removed.
Peaks were called using MACS2, a count-based peak caller that assumes Poisson read
distribution. Minimum peak length was set to 50 bp to ensure accurate calling of regions bound by
Dps, a relatively small DNA binding protein (compared to eukaryotic DNA binding proteins)
which in vitro binds DNA of length 90bp and larger with strong affinity (18).
The full set of peaks from 1 sample and time (i.e., both wild-type 8-hour samples) was
compared to the full set of peaks from the other sample at the same time (i.e., both wild-type 8-
hour samples compared to both dps-null 8-hour samples). Peaks unique to one strain at one time
were selected for motif analysis. Motif discovery from these unique peak sets was done using
MEME (217).
123
Chapter 5: Dps contributes to wild-type population dynamics and genome organization
124
Introduction
Dps, the DNA-binding protein from starved cells, is universally conserved across bacteria
and present in some archaea (see Chapter 1). Though initially discovered in Escherichia coli as a
DNA-binding protein, structural studies showed striking homology to ferritins (1, 18). This led to
the discovery that Dps is a major protein family of bacterial ferritins.
Ferritins are iron detoxifying and storage proteins that are conserved among all domains of
life. They evolved to help cells cope with a key biological paradox. Iron is necessary for many
biological processes, including such vital pathways as the electron transport chain. However, iron
can also be a source of reactive oxidative species (ROS) when its ferrous form reacts with H2O2
through chemistry known as the Fenton reaction, which can damage macromolecules including
DNA and proteins (46, 86).
The E. coli genome encodes three ferritin proteins: Dps, bacterioferritin (Bfr), and ferritin
(FtnA). The ferritin activity of Dps is unique among ferritins. This is demonstrated in several
ways. First, Dps functions as a dodecamer, which is substantially smaller (both in protein size and
iron storage capacity) than Bfr and FtnA (18, 46). Second, ferroxidase sites are situated at the
interface of two adjacent monomers within a Dps dodecamer, in contrast to the intra-subunit
ferroxidase sites in FtnA and Bfr (92). Finally, H2O2 is the preferred iron oxidizing agent used by
Dps compared to O2 for FtnA and Bfr, which further prevents Fenton chemistry from creating
ROS that can damage DNA (34).
In addition to its unique ferritin activity, the DNA-binding abilities of Dps are unique
among bacterial nucleoid-associated proteins (NAPs). NAPs are DNA-binding proteins that are
usually found associated with the nucleoid, which is the region of the bacterial cell that houses the
chromosome. In the proposed model for Dps-DNA binding, instead of possessing a canonical
125
DNA-binding domain, as many NAPs and eukaryotic histones do, Dps likely binds DNA by charge
attraction, possibly stabilized by divalent cations (18, 24). Additionally, no consensus sequence
for Dps action on DNA has been agreed upon, and as such Dps has been labeled a non-specific
DNA-binding protein (1).
While much has been elucidated about Dps in the three decades since its discovery, much
was still left to be discerned. First, though Dps confers a growth advantage to cells, it was unclear
how much of this advantage is due its two major functions as a DNA-binding protein and a ferritin.
Second, though Dps is necessary in vivo and sufficient in vitro for formation of the stationary
phase-specific nucleoid structure called the biocrystal, little was known about the dynamics of this
structure. This second question was two-fold: 1) is biocrystal formation biologically programmed,
and 2) are there distinct points of Dps-DNA interaction that help restructure the nucleoid over the
course of stationary phase? These points were all investigated over the course of this dissertation.
126
A ferritin hierarchy in E. coil
Dps confers a growth advantage when present in E. coli populations (62). However, how
much of this growth advantage is due to the DNA-binding versus ferritin function of Dps was
elucidated herein. To assess the effect of Dps on population dynamics, a dps-null strain was
cultured aerobically and population density was monitored daily for 14 days. This was compared
to a wild-type population over the same length of time to assess changes in population dynamics.
Additionally, a dps-null population was cultured in the same tube as a wild-type population, and
population density for each strain was monitored daily. To gauge whether the growth defects
observed in dps-null populations were typical for a strain lacking any ferritin, the same experiment
was performed using ftnA-null and bfr-null populations. Ferritin-null mutants were also cultured
with each other, both in pairwise (1v1) and 3-way (1v1v1) environments, to elucidate further
growth differences. Growth and survival studies were developed further by supplementing some
cultures with iron and depleting iron from other cultures. This was done for all strains (wild-type
and the three ferritin-null strains) in monoculture, as well as the dps-null vs. bfr-null vs. ftnA-null
competition schemes.
In monoculture, bfr-null strain growth is similar to the wild-type; ftnA-null populations
experience a 1-day stationary phase and a 2-step death phase; and a dps-null strain experiences
prolonged lag phase, prolonged death phase, and late entry into long-term stationary phase (LTSP).
Iron supplementation or depletion does not rescue the dps-null growth phenotype, but both
treatments partially restore the ftnA-null growth pattern.
In competition with the wild-type, additional growth effects are observed: bfr-null
population density dips below the wild-type by the end of death phase and remains lower through
the end of the experiment; the ftnA-null population outcompetes wild-type within a few days of
127
entering LTSP; and the dps-null population experiences a 1-day stationary phase and is
outcompeted by wild-type after that point. Taken together, these data suggest that wild-type
population dynamics are dependent upon Bfr during death phase and LTSP; rely on FtnA during
stationary phase; and depend on Dps during lag phase, stationary phase, death phase, and LTSP.
Pairwise competitions between ferritin mutants further revealed the temporal importance
of each ferritin. The bfr-null strain outcompetes both other null strains until day 7, at which point
the former’s population density declines precipitously. This further suggests that Bfr is important
for wild-type population dynamics during LTSP. When cocultured with a dps-null population, the
ftnA-null population outcompetes its opponent by day 2, strengthening the case for Dps importance
during stationary phase, death phase, and LTSP.
The final dataset that establishes a hierarchy among ferritins was gathered through 3-way
competitions among the single ferritin-null strains. The bfr-null strain outcompetes the other two
strains through stationary phase; the ftnA-null strain outcompetes the other two strains during death
phase; and the bfr- and ftnA-null strains both outcompete the dps-null strain through several days
of LTSP, after which the bfr-null strain is also outcompeted by the ftnA-null strain. When this
environment was supplemented with iron, the bfr-null strain exhibited less severe death in LTSP,
suggesting at least part of the growth advantage conferred by Bfr is due to its function as an iron
source for the population. Iron sequestration did not rescue any phenotype seen in this competition
scheme.
The growth and survival hierarchy demonstrated among ferritins could be due to
differences in gene expression. This includes levels of transcription and/or translation and mRNA
and/or protein stability. No data has been published regarding levels of translation or mRNA
stability for dps, bfr, or ftnA, and Dps protein has been shown to be rapidly degraded during log
128
phase and incredibly stable during stationary phase (122), but no protein stability data exists for
FtnA or Bfr. Recently, an RNA-Sequencing time course was published in E. coli across growth
phases (212). This data demonstrates that dps is the most highly expressed ferritin gene at every
growth phase examined (log phase, stationary phase, and 3 LTSP timepoints [3 days, 6 days, and
8 days post-inoculation]), with its highest level of expression during stationary phase. Further,
while bfr and ftnA are expressed at similar levels during log phase, ftnA expression drops ~15-fold
by stationary phase, and bfr expression drops ~10-fold by the same timepoint. mRNA levels of all
three ferritin genes fluctuate slightly between stationary phase and the latest LTSP timepoint. This
data strengthens the model proposed: due to its much higher expression than the other two ferritin
genes, dps is the most important of these genes across the growth and survival curve; ftnA and bfr
are less vital to wild-type population dynamics. Additional studies into ferritin gene expression
will test the validity of this hypothesis.
Taken together, the data presented in chapter 2 establishes a fitness hierarchy among
ferritins in E. coli. Dps is most important during lag phase, stationary phase, death phase, and
LTSP; FtnA is second most important during stationary phase; and Bfr is second most important
during LTSP. Moreover, the growth advantage conferred by Dps is not wholly explained by its
ferritin function. This suggests its DNA-binding functions require further investigation to
understand the advantage Dps provides.
129
Investigating the nature of the Dps-DNA biocrystal
After discovering that the ferritin function of Dps cannot explain the growth advantage it
provides, I turned to investigate its DNA-binding function. More specifically, when Dps binds
DNA, does it bind predetermined chromosomal loci or is its DNA binding more plastic? These
two DNA-binding models, it must be noted, are not mutually exclusive and may both be true at
different locations (specific and non-specific binding). To investigate this question, a PCR-based
assay was used. Nucleoids from wild-type or dps-null strains were harvested during mid-log phase
or late stationary phase and treated with one of three endonucleases (AvrII, NotI, or XbaI).
Combined, these three nucleases have 74 restriction sites in the E. coli genome. After nuclease
treatment, nucleoids were de-crosslinked and DNA was precipitated. To quantify the amount of
digestion at each of the 74 restriction sites, PCR was run using primers flanking each restriction
site, PCR products were run on an agarose gel, and band intensity was quantified using ImageJ. In
this experiment, unbound DNA would be more accessible to nuclease digestion, resulting in less
PCR product and a fainter band. Conversely, bound DNA would be less accessible to digestion,
resulting in more PCR product and a more intense band. Sites determined to be significantly
different in their digestion pattern between wild-type and dps-null populations during only late
stationary phase were used in a motif search, as well as 1kb upstream and 1kb downstream of each
site.
26 restriction sites (~35% of sites probed) were determined to be significantly less
protected during stationary phase in dps-null nucleoids compared to wild-type. After motif
analysis, one motif was found to be specific to these stationary phase-unique significant sites:
KGCMGMGAAASCGGCVGCWKM. It is possible that this putative motif might anchor Dps
130
binding. Moreover, this motif is overrepresented in the E. coli genome, supporting its potential
benefit.
The PCR-based assay presented in Chapter 3 demonstrated that at least some Dps DNA-
binding sites are biologically programmed. Not only is this the first time this phenomenon has
been confirmed in a Dps-dependent manner, but this also suggests some driving force contributing
to Dps binding locations, potentially the putative consensus sequence from Chapter 3,
KGCMGMGAAASCGGCVGCWKM. This motif is overrepresented in the genome, the extent to
which varies with motif stringency: from exact matches to five allowed mismatches, zero, zero,
one, 11, 87, and 397 occurrences of this motif are expected; and one, four, 11, 68, 485, and 2,484
occurrences are observed (Table 3.4). While overrepresentation suggests a biological benefit to
having this motif, the number of occurrences is so small that it is difficult to reconcile with the
sheer volume of the genome (up to 50%) bound by Dps during late stationary phase. Additionally,
because this is based on 26 sequences, this exact motif is unlikely to be the only biologically
relevant Dps consensus sequence. A model is possible that assumes there are a relatively small
number of biologically programmed genomic loci that act as Dps anchoring points that potentially
act as precursors to toroid formation during the transition from log phase to stationary phase. In
this model, this putative Dps binding motif that occurs less frequently than one might expect a
major NAP binding motif to occur may act as these primary Dps anchors. To better understand
where Dps binds in the E. coli genome and what a more likely consensus sequence is, a higher
resolution method was needed.
131
Dps-dependent chromatin accessibility revealed
The PCR experiments described in Chapter 3 suggest that, for Dps, at least some of its
DNA-binding sites are biologically programmed. However, those experiments gave a very low-
resolution view of what is happening in the nucleoid since it probed <.01% of the genome. To get
a higher resolution understanding of genome accessibility during stationary phase, micrococcal
nuclease (MNase) treatment followed by sequencing (MNase-Seq) was performed. Briefly, wild-
type or dps-null nucleoids harvested during early or late stationary phase were crosslinked,
digested with MNase, de-crosslinked, and the remaining DNA was sequenced. DNA bound by
protein would be protected from MNase, whereas unbound DNA would be digested, leaving just
protein-bound DNA for sequencing. This would delineate which regions of the genome are
accessible for a given strain and timepoint. Sequencing reads were aligned to the wild-type genome
and peaks were called using MACS2. Peaks shared between early and late stationary phase wild-
type nucleoids were used in a MEME motif search, as these peaks potentially contain points of
Dps binding.
During late stationary phase, the wild-type nucleoid has fewer peaks, but they are longer
in DNA length. This is consistent with a model of the structure of the wild-type nucleoid during
late stationary phase. First, the regular spacing and structure of the biocrystal likely protects longer
stretches of DNA, meaning numerically fewer peaks would pass this threshold. Second, because
so much of the chromosome is bound in the biocrystal, it is expected that larger stretches of DNA
would be overrepresented and thus present in peaks.
106 peaks are shared between the early stationary phase and late stationary phase wild-type
nucleoids. These peaks represent stretches of DNA with potential Dps binding sites, possibly
points of nucleation for the biocrystal, and as such may give insight into a Dps binding motif.
132
Using MEME to compare these shared peaks, a significant motif emerged:
TTTTTTCDSCWDYW. This is overrepresented in the E. coli genome, though to varying degrees
depending on motif stringency: from exact match to allowing 5 mismatches, there are 2, 104,
2,038, 18,837, 103,959, and 400,299 occurrences of this motif expected; and 1, 777, 6,221, 34,736,
144,222, and 464,448 occurrences of this motif observed (Table 4.2). The overrepresentation of
this motif suggests it is beneficial in some way, possibly to direct Dps binding. This is much more
frequent than the motif presented in Chapter 3, though still less frequent when compared to the
consensus sequences for Fis or IHF binding. However, because the model for Dps-DNA binding
is different from that of Fis or IHF in the demonstrated ability for Dps to self-aggregate, it is
possible that this motif acts as a primary Dps binding site, and secondary binding radiates from
these nucleation points.
133
Discussion
Two motifs have been presented in this dissertation. The first, derived from 26 2kb regions
centered at restriction sites, is KGCMGMGAAASCGGCVGCWKM. The second, derived from
106 MNase-Seq peaks, is TTTTTTCDSCWDYW. Both motifs are overrepresented not just in their
own datasets (see Chapters 3 and 4), but also in each other’s datasets. The PCR-derived motif is
modestly overrepresented in the MNase-Seq wild-type shared peak set by a factor of 4 only when
the motif was allowed 5 mismatches (1 occurrence expected, 4 observed); the MNase-Seq-derived
motif is overrepresented in the PCR-based stationary phase-specific Dps-dependent sites by a
factor of 10 (1 mismatch), 4.1 (2 mismatches), 2.4 (3 mismatches), 1.7 (4 mismatches), and 1.3 (5
mismatches) (Table 5.1). While this phenomenon suggests a greater potential for biological
relevance of the MNase-Seq-derived motif, it is possible that both motifs are biologically relevant.
A unifying model might see one motif as primary anchoring points for early stationary phase toroid
formation and the other as secondary toroid or biocrystal binding loci.
Using Tomtom to compare the PCR- and MNase-Seq-derived motifs (219), an alignment
with 11 bp overlap (DGSHGMRAAAV) and 3 bp offset emerged (Figure 5.1). This motif
becomes less overrepresented in the E. coli genome as motif stringency decreases: for an exact
match, 1 mismatch, and 2 mismatches, it is observed 6.8X, 2.7X, and 1.3X more than it is expected,
respectively. It becomes underrepresented as motif stringency continues to decrease, as it is
expected 1.3X more than it is observed when 3 mismatches are allowed, 2.0X at 4 mismatches,
and 2.9X at 5 mismatches. This inversion from overrepresentation at higher motif stringency
(exact, 1 mismatch, or 2 mismatches) to underrepresentation at lower motif stringency (3, 4, or 5
mismatches) is mirrored in both the MNase-Seq shared wild-type peak set and the PCR-derived
late stationary phase-specific Dps-dependent loci. While over-/underrepresentation varies for this
134
motif with motif stringency, its frequency in the E. coli genome is much more similar to those of
Fis and IHF than either of the aforementioned motifs (Table 3.4, Table 4.2). This suggests possible
biological relevance at a scale similar to other major NAPs, a classification of which Dps is a
member.
Biochemical and genetic experiments should be used to confirm the Dps DNA-binding
motif. First, biochemically, binding constants can be tested for motif-containing and non-motif-
containing stretches of DNA. Second, genetically, sites determined to be significant from PCR and
MNase assays can be places in different genetic contexts, like a plasmid, and Dps action can be
assessed. Taken together, these two methods can validate how, where, and why Dps binds, and
may point more definitively to biocrystal nucleation points.
135
Conclusion
Dps is a universally conserved bacterial ferritin and DNA-binding protein. Physiological
studies suggest that, though Dps confers an advantage in maintaining wild-type population
dynamics, this advantage is not due to its ferritin activity. This led to further investigation of the
DNA-binding functions of Dps. First, a PCR-based assay proved that Dps binds at least some
portion of the E. coli genome in a biologically programmed way. Because this assay was only able
to probe <.01% of the chromosome, MNase-Seq was employed to assess the accessibility of the
chromosome when dps was present or absent during early and late stationary phase. These two
experimental designs revealed two putative Dps DNA-binding motifs,
KGCMGMGAAASCGGCVGCWKM and TTTTTTCDSCWDYW. A model for biological
relevance of both of these motifs suggests one might act as primary Dps binding points for toroid
formation during the transition from log phase to stationary phase, and the other might act as
secondary biocrystal anchors. From these motifs, a unifying motif was derived:
DGSHGMRAAAV. Perhaps a more conventional model of NAP-DNA binding applied to Dps
would suggest that this motif directs Dps binding throughout the genome. While much has been
elucidated herein, this dissertation provides a base for future work to fully understand the
biological action of the E. coli DNA-binding protein from starved cells.
136
Figures
Figure 5.1: PCR- and MNase-Seq-derived motif alignment
Motifs derived from the PCR-based assay described in Chapter 3 (bottom) and the MNase-Seq
experiment described in Chapter 4 (top) were aligned using Tomtom. Resulting alignment
(DGSHGMRAAAV) has an 11 bp overlap and a 3 bp offset.
1
1
Tomtom (no SSC) 14.02.2022 14:20
0
1
2
bits
1
C G A
T
2
C T A
G
3
A
C
T
4
C G A
T
5
A
C
T
G
6
A T
C
G
7
A
C
T
8
A
C
T
G
9
C
G
T
A
10
C
G
T
A
11
C
G
T
A
12
C
G
T
A
13
C
G
T
A
14
C
G
T
A
0
1
2
bits
1
A C
G
T
2
A
C
T
G
3
A
G
T
C
4
G T A
C
5
A
C
T
G
6
G T A
C
7
A
C
T
G
8
C
G
T
A
9
C
G
T
A
10
C
G
T
A
11
A T
C
G
12
A
G
T
C
13
A
C
T
G
14
A
C
T
G
15
A
G
T
C
16
A
C
G
17
A
C
T
G
18
A
G
T
C
19
C G A
T
20
A C
G
T
21
G T A
C
137
Tables
Table 5.1: Motif frequency in alternate datasets
motif
dataset
queried
motif
stringency
expected observed overrepresentation
KGCMGMGAAASCGGCVGCWKM
(PCR-derived, Chapter 3)
Mnase-Seq
wild-type
shared peaks
(Chapter 4)
exact 0 0 —
1 mismatch 0 0 —
2 mismatches 0 0 —
3 mismatches 0 0 —
4 mismatches 1 1 —
5 mismatches 3 12 4
TTTTTTCDSCWDYW (Chapter 4)
PCR-derived
stationary
phase-
specific
significant
sites
(Chapter 3)
exact 0 0 —
1 mismatch 1 10 10
2 mismatches 23 95 4.1
3 mismatches 216 509 2.36
4 mismatches 1,190 2,054 1.726
5 mismatches 4,584 6,127 1.337
DGSHGMRAAAV (alignment of
motifs from Chapter 3 and Chapter 4)
Mnase-Seq
wild-type
shared peaks
(Chapter 4)
exact 2 17 8.5
1 mismatch 52 160 3.08
2 mismatches 702 952 1.36
3 mismatches 5,071 4,078 0.8042
4 mismatches 22,968 11,592 0.50470
5 mismatches 72,214 24,789 0.34327
PCR-derived
stationary
phase-
specific
significant
sites
(Chapter 3)
exact 3 31 10
1 mismatch 80 252 3.15
2 mismatches 1,075 1,492 1.388
3 mismatches 7,767 6,069 0.7814
4 mismatches 35,185 17,511 0.49768
5 mismatches 110,627 37,961 0.343144
ZK126
genome
exact 234 1,601 6.842
1 mismatch 7,016 18,860 2.6881
2 mismatches 93,856 121,141 1.29071
3 mismatches 678,437 511,104 0.753355
4 mismatches 3,072,678 1,512,112 0.4921153
5 mismatches 9,660,948 3,315,395 0.3431749
138
References
1. Almirón M, Link AJ, Furlong D, Kolter R. 1992. A novel DNA-binding protein with
regulatory and protective roles in starved Escherichia coli. Genes & Development
6:2646–2654.
2. Gupta S, Chatterji D. 2003. Bimodal protection of DNA by Mycobacterium smegmatis
DNA-binding protein from stationary phase cells. J Biol Chem 278:5235–41.
3. Lomovskaya OL, Kidwell JP, Matin A. 1994. Characterization of the σ38-Dependent
Expression of a Core Escherichia coli Starvation Gene, pexB. J Bacteriol 176:3928–3935.
4. Peña MM, Bullerjahn GS. 1995. The DpsA protein of Synechococcus sp. Strain PCC7942
is a DNA-binding hemoprotein. Linkage of the Dps and bacterioferritin protein families. J
Biol Chem 270:22478–22482.
5. Haikarainen T, Papageorgiou AC. 2010. Dps-like proteins: structural and functional
insights into a versatile protein family. Cellular and Molecular Life Sciences 67:341–351.
6. Chiancone E, Ceci P. 2010. The multifaceted capacity of Dps proteins to combat bacterial
stress conditions: Detoxification of iron and hydrogen peroxide and DNA binding.
Biochimica et Biophysica Acta - General Subjects
https://doi.org/10.1016/j.bbagen.2010.01.013.
7. Shu JC, Soo PC, Chen JC, Hsu SH, Chen LC, Chen CY, Liang SH, Buu LM, Chen CC.
2013. Differential regulation and activity against oxidative stress of Dps proteins in
Bacillus cereus. International Journal of Medical Microbiology 303:662–673.
8. Ramsay B, Wiedenheft B, Allen M, Gauss GH, Martin Lawrence C, Young M, Douglas
T. 2006. Dps-like protein from the hyperthermophilic archaeon Pyrococcus furiosus.
Journal of Inorganic Biochemistry 100:1061–1068.
9. Gauss GH, Benas P, Wiedenheft B, Young M, Douglas T, Martin Lawrence C. 2006.
Structure of the DPS-Like Protein from Sulfolobus solfataricus Reveals a Bacterioferritin-
Like Dimetal Binding Site within a DPS-Like Dodecameric Assembly. Biochemistry
45:10815–10827.
10. Karas VO, Westerlaken I, Meyer AS. 2015. The DNA-Binding Protein from Starved Cells
(Dps) Utilizes Dual Functions To Defend Cells against Multiple Stresses. Journal of
Bacteriology 197:3206–3215.
11. Minsky A, Wolf SG, Frenkiel D, Arad T, Finkel SE, Kolter R. 1999. DNA protection by
stress-induced biocrystallization. Nature 400:83–85.
12. Bhattacharyya G, Grove A. 2007. The N-terminal extensions of Deinococcus radiodurans
Dps-1 mediate DNA major groove interactions as well as assembly of the dodecamer. J
Biol Chem 282:11921–11930.
139
13. Ushijima Y, Yoshida O, Adolf Villanueva MJ, Ohniwa RL, Morikawa K. 2016. Nucleoid
clumping is dispensable for the Dps-dependent hydrogen peroxide resistance in
Staphylococcus aureus. Microbiology (United Kingdom) 162:1822–1828.
14. Grove A, Wilkinson SP. 2005. Differential DNA Binding and Protection by Dimeric and
Dodecameric forms of the Ferritin Homolog Dps from Deinococcus radiodurans. Journal
of Molecular Biology 347:495–508.
15. Ceci P, Mangiarotti L, Rivetti C, Chiancone E. 2007. The neutrophil-activating Dps
protein of Helicobacter pylori, HP-NAP, adopts a mechanism different from Escherichia
coli Dps to bind and condense DNA. Nucleic Acids Research 35:2247–2256.
16. Meyer AS, Grainger DC. 2013. The Escherichia coli Nucleoid in Stationary Phase.
Advances in Applied Microbiology 83:69–86.
17. Liu G, Ma Q, Xu Y. 2018. Physical properties of DNA may direct the binding of
nucleoid-associated proteins along the E. coli genome. Mathematical Biosciences 301:50–
58.
18. Grant RA, Filman DJ, Finkel SE, Kolter R, Hogle JM. 1998. The crystal structure of Dps,
a ferritin homolog that binds and protects DNA. nature structural biology 5:294–303.
19. Ali Azam T, Iwata A, Nishimura A, Ueda S, Ishihama A. 1999. Growth Phase-Dependent
Variation in Protein Composition of the Escherichia coli Nucleoid. JOURNAL OF
BACTERIOLOGY 181:6361–6370.
20. Talukder AA, Ishihama A. 2014. Dps Is a Stationary Phase-Specific Protein of
Escherichia coli Nucleoid. Advances in Microbiology 04:1095–1104.
21. Ishihama A, Kori A, Koshio E, Yamada K, Maeda H, Shimada T, Makinoshima H, Iwata
A, Fujita N. 2014. Intracellular concentrations of 65 species of transcription factors with
known regulatory functions in Escherichia coli. J Bacteriol 196:2718–2727.
22. Talukder A, Ishihama A. 2015. Growth phase dependent changes in the structure and
protein composition of nucleoid in Escherichia coli. Science China Life Sciences 58:902–
911.
23. Dillon SC, Dorman CJ. 2010. Bacterial nucleoid-associated proteins, nucleoid structure
and gene expression. Nature Reviews Microbiology 8:185–195.
24. Frenkiel-Krispin D, Levin-Zaidman S, Shimoni E, Wolf SG, Wachtel EJ, Arad T, Finkel
SE, Kolter R, Minsky A. 2001. Regulated phase transitions of bacterial chromatin: a non-
enzymatic pathway for generic DNA protection. EMBO J 20:1184–1191.
25. Frenkiel-Krispin D, Ben-Avraham I, Englander J, Shimoni E, Wolf SG, Minsky A. 2004.
Nucleoid restructuring in stationary-state bacteria. Molecular Microbiology 51:395–405.
140
26. Kauko A, Haataja S, Pulliainen AT, Finne J, Papageorgiou AC. 2004. Crystal Structure of
Streptococcus suis Dps-like Peroxide Resistance Protein Dpr: Implications for Iron
Incorporation. Journal of Molecular Biology 338:547–558.
27. Roy S, Saraswathi R, Chatterji D, Vijayan M. 2008. Structural Studies on the Second
Mycobacterium smegmatis Dps: Invariant and Variable Features of Structure, Assembly
and Function. Journal of Molecular Biology 375:948–959.
28. Kim J, Yoshimura SH, Hizume K, Ohniwa RL, Ishihama A, Takeyasu K. 2004.
Fundamental structural units of the Escherichia coli nucleoid revealed by atomic force
microscopy. Nucleic Acids Research 32:1982–1992.
29. Takeyasu K, Kim J, Ohniwa RL, Kobori T, Inose Y, Morikawa K, Ohta T, Ishihama A,
Yoshimura SH. 2004. Genome architecture studied by nanoscale imaging: analyses among
bacterial phyla and their implication to eukaryotic genome folding. Cytogenetic and
Genome Research 107:38–48.
30. Morikawa K, Ohniwa RL, Kim J, Takeshita SL, Maruyama A, Inose Y, Takeyasu K, Ohta
T. 2007. Biochemical, Molecular Genetic, and Structural Analyses of the Staphylococcal
Nucleoid. Microscopy and Microanalysis 13:30–35.
31. Ghatak P, Karmakar K, Kasetty S, Chatterji D. 2011. Unveiling the Role of Dps in the
Organization of Mycobacterial Nucleoid. PLoS ONE 6:e16019.
32. Bozzi M, Mignogna G, Stefanini S, Barra D, Longhi C, Valenti P, Chiancone E. 1997. A
novel non-heme iron-binding ferritin related to the DNA-binding proteins of the Dps
family in Listeria innocua. J Biol Chem 272:3259–3265.
33. Tonello F, Dundon WG, Satin B, Molinari M, Tognon G, Grandi G, del Giudice G,
Rappuoli R, Montecucco C. 1999. The Helicobacter pylori neutrophil-activating protein is
an iron-binding protein with dodecameric structure. Molecular Microbiology 34:238–246.
34. Zhao G, Ceci P, Ilari A, Giangiacomo L, Laue TM, Chiancone E, Chasteen ND. 2002.
Iron and Hydrogen Peroxide Detoxification Properties of DNA-binding Protein from
Starved Cells A Ferritin-Like DNA-Binding Protein of Escherichia coli. The Journal of
Biological Chemistry 277:27689–27696.
35. Yamamoto Y, Poole LB, Hantgan RR, Kamio Y. 2002. An Iron-Binding Protein, Dpr,
from Streptococcus mutans Prevents Iron-Dependent Hydroxyl Radical Formation In
Vitro. JOURNAL OF BACTERIOLOGY 184:2931–2939.
36. Ceci P, Ilari A, Falvo E, Chiancone E. 2003. The Dps protein of Agrobacterium
tumefaciens does not bind to DNA but protects it toward oxidative cleavage: x-ray crystal
structure, iron binding, and hydroxyl-radical scavenging properties. J Biol Chem
278:20319–20326.
141
37. Ishikawa T, Mizunoe Y, Kawabata SI, Takade A, Harada M, Wai SN, Yoshida SI. 2003.
The Iron-Binding Protein Dps Confers Hydrogen Peroxide Stress Resistance to
Campylobacter jejuni. JOURNAL OF BACTERIOLOGY 185:1010–1017.
38. Su M, Cavallo S, Stefanini S, Chiancone E, Chasteen ND. 2005. The So-Called Listeria
innocua Ferritin Is a Dps Protein. Iron Incorporation, Detoxification, and DNA Protection
Properties. Biochemistry 44:5572–5578.
39. Liu X, Kim K, Leighton T, Theil EC. 2006. Paired Bacillus anthracis Dps (mini-ferritin)
have different reactivities with peroxide. J Biol Chem 281:27827–27835.
40. Li X, Pal U, Ramamoorthi N, Liu X, Desrosiers DC, Eggers ChristiaNH, Anderson JF,
Radolf JD, Fikrig E. 2007. The Lyme disease agent Borrelia burgdorferi requires
BB0690, a Dps homologue, to persist within ticks. Molecular Microbiology 63:694–710.
41. Pulliainen AT, Haataja S, Kähkönen S, Finne J. 2003. Molecular basis of H2O2 resistance
mediated by Streptococcal Dpr. J Biol Chem 278:7996–8005.
42. Wiedenheft B, Mosolf J, Willits D, Yeager M, Dryden KA, Young M, Douglas T. 2005.
An archaeal antioxidant: characterization of a Dps-like protein from Sulfolobus
solfataricus. Proc Natl Acad Sci U S A 102:10551–10556.
43. Franceschini S, Ceci P, Alaleona F, Chiancone E, Ilari A. 2006. Antioxidant Dps protein
from the thermophilic cyanobacterium Thermosynechococcus elongatus. FEBS Journal
273:4913–4928.
44. Tu WY, Pohl S, Gizynski K, Harwood CR. 2012. The iron-binding protein Dps2 confers
peroxide stress resistance on Bacillus anthracis. J Bacteriol 194:925–931.
45. Haikarainen T, Tsou C-C, Wu J-J, Papageorgiou AC. 2010. Crystal structures of
Streptococcus pyogenes Dpr reveal a dodecameric iron-binding protein with a ferroxidase
site. JBIC Journal of Biological Inorganic Chemistry 15:183–194.
46. Harrison PM, Arosio P. 1996. The ferritins: molecular properties, iron storage function
and cellular regulation. Biochimica et Biophysica Acta (BBA) - Bioenergetics 1275:161–
203.
47. Roy S, Gupta S, Das S, Sekar K, Chatterji D, Vijayan M, IUCr. 2003. Crystallization and
preliminary X-ray diffraction analysis of Mycobacterium smegmatis Dps. Acta
Crystallographica 59:2254–2256.
48. Williams SM, Chandran AV, Prakash S, Vijayan M, Chatterji D. 2017. A Mutation
Directs the Structural Switch of DNA Binding Proteins under Starvation to a Ferritin-like
Protein Cage. Structure 25:1449–1454.
49. Ceci P, Cellai S, Falvo E, Rivetti C, Rossi GL, Chiancone E. 2004. DNA condensation
and self-aggregation of Escherichia coli Dps are coupled phenomena related to the
properties of the N-terminus. Nucleic Acids Research 32:5935–5944.
142
50. Roy S, Saraswathi R, Gupta S, Sekar K, Chatterji D, Vijayan M. 2007. Role of N and C-
terminal Tails in DNA Binding and Assembly in Dps: Structural Studies of
Mycobacterium smegmatis Dps Deletion Mutants. Journal of Molecular Biology 370:752–
767.
51. Roy S, Gupta S, Das S, Sekar K, Chatterji D, Vijayan M. 2004. X-ray Analysis of
Mycobacterium smegmatis Dps and a Comparative Study Involving Other Dps and Dps-
like Molecules. Journal of Molecular Biology 339:1103–1113.
52. Zhang Y, Fu J, Chee SY, Ang EXW, Orner BP. 2011. Rational disruption of the
oligomerization of the mini-ferritin E. coli DPS through protein-protein interface
mutation. Protein Science 20:1907–1917.
53. Stillman TJ, Upadhyay M, Norte VA, Sedelnikova SE, Carradus M, Tzokov S, Bullough
PA, Shearman CA, Gasson MJ, Williams CH, Artymiuk PJ, Green J. 2005. The crystal
structures of Lactococcus lactis MG1363 Dps proteins reveal the presence of an N-
terminal helix that is required for DNA binding. Molecular Microbiology 57:1101–1112.
54. Ceci P, Ilari A, Falvo E, Giangiacomo L, Chiancone E. 2005. Reassessment of protein
stability, DNA binding, and protection of Mycobacterium smegmatis Dps. Journal of
Biological Chemistry 280:34776–34785.
55. Bessonova TA, Shumeiko SA, Purtov YuA, Antipov SS, Preobrazhenskaya E v., Tutukina
MN, Ozoline ON. 2016. Hexuronates influence the oligomeric form of the Dps structural
protein of bacterial nucleoid and its ability to bind to linear DNA fragments. Biophysics
(Oxf) 61:825–832.
56. Rothe M, Alpert C, Loh G, Blaut M. 2013. Novel Insights into E. coli’s Hexuronate
Metabolism: KduI Facilitates the Conversion of Galacturonate and Glucuronate under
Osmotic Stress Conditions. PLOS ONE 8:e56906.
57. Wang SW, Chen CY, Tseng JT, Liang SH, Chen SC, Hsieh C, Chen YH, Chen CC. 2009.
Orf4 of the Bacillus cereus sigB gene cluster encodes a general stress-inducible dps-like
bacterioferritin. Journal of Bacteriology 191:4522–4533.
58. Ohniwa RL, Morikawa K, Kim J, Ohta T, Ishihama A, Wada C, Takeyasu K. 2006.
Dynamic state of DNA topology is essential for genome condensation in bacteria. The
EMBO Journal 25:5591–5602.
59. Ohniwa RL, Muchaku H, Saito S, Wada C, Morikawa K. 2013. Atomic Force Microscopy
Analysis of the Role of Major DNA-Binding Proteins in Organization of the Nucleoid in
Escherichia coli. PLoS ONE 8:e72954.
60. Zimmerman SB. 2006. Cooperative transitions of isolated Escherichia coli nucleoids:
Implications for the nucleoid as a cellular phase. Journal of Structural Biology 153:160–
175.
143
61. Minsky A, Shimoni E, Frenkiel-Krispin D. 2002. Stress, order and survival. Nature
Reviews Molecular Cell Biology 3:50–60.
62. Nair S, Finkel SE. 2004. Dps protects cells against multiple stresses during stationary
phase. J Bacteriol 186:4192–4198.
63. Sato YT, Watanabe S, Kenmotsu T, Ichikawa M, Yoshikawa Y, Teramoto J, Imanaka T,
Ishihama A, Yoshikawa K. 2013. Structural change of DNA induced by nucleoid proteins:
Growth phase-specific fis and stationary phase-specific dps. Biophysical Journal
105:1037–1044.
64. Lee SY, Lim CJ, Dröge P, Yan J. 2015. Regulation of Bacterial DNA Packaging in Early
Stationary Phase by Competitive DNA Binding of Dps and IHF. Nature Publishing Group
https://doi.org/10.1038/srep18146.
65. Hong Y, Wang G, Maier RJ. 2006. Helicobacter hepaticus Dps protein plays an important
role in protecting DNA from oxidative damage. Free Radical Research 40:597–605.
66. Morikawa K, Ohniwa RL, Kim J, Maruyama A, Ohta T, Takeyasu K. 2006. Bacterial
nucleoid dynamics: oxidative stress response in Staphylococcus aureus. Genes to Cells
11:409–423.
67. Foley PL, Wilson DB, Shuler ML. 2010. Macromolecular crowding can account for
RNase-sensitive constraint of bacterial nucleoid structure. Biochemical and Biophysical
Research Communications 395:42–47.
68. Pinson V, Takahashi M, Rouviere-Yaniv J. 1999. Differential binding of the Escherichia
coli HU, homodimeric forms and heterodimeric form to linear, gapped and cruciform
DNA. Journal of Molecular Biology 287:485–497.
69. Gulvady R, Gao Y, Kenney LJ, Yan J. 2018. A single molecule analysis of H-NS
uncouples DNA binding affinity from DNA specificity. Nucleic Acids Research
46:10216–10224.
70. Stella S, Cascio D, Johnson RC. 2010. The shape of the DNA minor groove directs
binding by the DNA-bending protein Fis. Genes and Development 24:814–826.
71. Shao Y, Feldman-Cohen LS, Osuna R. 2008. Functional Characterization of the
Escherichia coli Fis-DNA Binding Sequence. Journal of Molecular Biology 376:771–785.
72. Wang S, Cosstick R, Gardner JF, Gumport RI. 1995. The specific binding of Escherichia
coli integration host factor involves both major and minor grooves of DNA. Biochemistry
34:13082–13090.
73. Windbichler N, von Pelchrzim F, Mayer O, Csaszar E, Schroeder R. 2008. RNA Biology
Isolation of small RNA-binding proteins from E. coli: Evidence for frequent interaction of
RNAs with RNA polymerase. RNA Biology 5.
144
74. Joyeux M. 2015. Compaction of bacterial genomic DNA: clarifying the concepts. Journal
of Physics: Condensed Matter 27:383001.
75. Antipov SS, Tutukina MN, Preobrazhenskaya E V, Kondrashov FA, Patrushev M V,
Toshchakov S V, Dominova I, Shvyreva US, Vrublevskaya V V, Morenkov OS,
Sukharicheva NA, Panyukov V V, Ozoline ON. 2017. The nucleoid protein Dps binds
genomic DNA of Escherichia coli in a non-random manner. PLoS ONE 12.
76. Ishihama A, Shimada T, Yamazaki Y. 2016. Transcription profile of Escherichia coli:
genomic SELEX search for regulatory targets of transcription factors. Nucleic Acids
Research 44:2058–2074.
77. Ueshima J, Shoji M, Ratnayake DB, Abe K, Yoshida S, Yamamoto K, Nakayama K.
2003. Purification, Gene Cloning, Gene Expression, and Mutants of Dps from the
Obligate Anaerobe Porphyromonas gingivalis. INFECTION AND IMMUNITY 71:1170–
1178.
78. Ping L, Büchler R, Mithöfer A, Svatoš A, Spiteller D, Dettner K, Gmeiner S, Piel J,
Schlott B, Boland W. 2007. A novel Dps-type protein from insect gut bacteria catalyses
hydrolysis and synthesis of N-acyl amino acids. Environmental Microbiology 9:1572–
1583.
79. Kim SG, Bhattacharyya G, Grove A, Lee YH. 2006. Crystal Structure of Dps-1, a
Functionally Distinct Dps Protein from Deinococcus radiodurans. Journal of Molecular
Biology 361:105–114.
80. Santos SP, Cuypers MG, Round A, Finet S, Narayanan T, Mitchell EP, Romão C v. 2017.
SAXS Structural Studies of Dps from Deinococcus radiodurans Highlights the
Conformation of the Mobile N-Terminal Extensions. Journal of Molecular Biology
429:667–687.
81. Grainger DC, Goldberg MD, Lee DJ, Busby SJW. 2008. Selective repression by Fis and
H-NS at the Escherichia coli dps promoter. Molecular Microbiology 68:1366–1377.
82. Browning DF, Grainger DC, Busby SJ. 2010. Effects of nucleoid-associated proteins on
bacterial chromosome structure and gene expression. Current Opinion in Microbiology
13:773–780.
83. Graham JS, Johnson RC, Marko JF. 2011. Concentration-dependent exchange accelerates
turnover of proteins bound to double-stranded DNA. Nucleic Acids Res 39:2249–2259.
84. Hadizadeh N, Johnson RC, Marko JF. 2016. Facilitated dissociation of a nucleoid protein
from the bacterial chromosome. Journal of Bacteriology 198:1735–1742.
85. Imlay JA. 2013. The molecular mechanisms and physiological consequences of oxidative
stress: lessons from a model bacterium. Nature Reviews Microbiology 11:443–454.
145
86. Andrews SC. 2010. The Ferritin-like superfamily: Evolution of the biological iron
storeman from a rubrerythrin-like ancestor. Biochimica et Biophysica Acta (BBA) -
General Subjects 1800:691–705.
87. Alaleona F, Franceschini S, Ceci P, Ilari A, Chiancone E. 2010. Thermosynechoccus
elongatus DpsA binds Zn(II) at a unique three histidine-containing ferroxidase center and
utilizes O2 as iron oxidant with very high efficiency, unlike the typical Dps proteins.
FEBS Journal 277:903–917.
88. Schwartz JK, Liu XS, Tosha T, Diebold A, Theil EC, Solomon EI. 2010. CD and MCD
Spectroscopic Studies of the Two Dps Miniferritin Proteins from Bacillus anthracis: Role
of O2 and H2O2 Substrates in Reactivity of the Diiron Catalytic Centers. Biochemistry
49:10516–10525.
89. Howe C, Moparthi VK, Ho FM, Persson K, Stensjö Id K. 2019. The Dps4 from Nostoc
punctiforme ATCC 29133 is a member of His-type FOC containing Dps protein class that
can be broadly found among cyanobacteria. PLOS ONE 14.
90. Kauko A, Pulliainen AT, Haataja S, Meyer-Klaucke W, Finne J, Papageorgiou AC. 2006.
Iron Incorporation in Streptococcus suis Dps-like Peroxide Resistance Protein Dpr
Requires Mobility in the Ferroxidase Center and Leads to the Formation of a Ferrihydrite-
like Core. Journal of Molecular Biology 364:97–109.
91. Huergo LF, Rahman H, Ibrahimovic A, Day CJ, Korolik V. 2013. Campylobacter jejuni
Dps protein binds DNA in the presence of iron or hydrogen peroxide. Journal of
Bacteriology 195:1970–1978.
92. Ilari A, Ceci P, Ferrari D, Rossi GL, Chiancone E. 2002. Iron incorporation into
Escherichia coli Dps gives rise to a ferritin-like microcrystalline core. J Biol Chem
277:37619–37623.
93. Haggis G. 1965. The iron oxide core of the ferritin molecule. J Mol Biol 14:598–602.
94. Bevers LE, Theil EC. 2011. Maxi- and mini-ferritins: minerals and protein nanocages.
Prog Mol Subcell Biol 52:29–47.
95. Castruita M, Saito M, Schottel PC, Elmegreen LA, Myneni S, Stiefel EI, Morel FMM.
2006. Overexpression and characterization of an iron storage and DNA-binding Dps
protein from Trichodesmium erythraeum. Appl Environ Microbiol 72:2918–2924.
96. Reindel S, Schmidt CL, Anemüller S, Matzanke BF. 2002. Characterization of a non-
haem ferritin of the Archaeon Halobacterium salinarum, homologous to Dps (starvation-
induced DNA-binding protein). Biochem Soc Trans 30:713–715.
97. Havukainen H, Haataja S, Kauko A, Pulliainen AT, Salminen A, Haikarainen T, Finne J,
Papageorgiou AC. 2008. Structural basis of the zinc- and terbium-mediated inhibition of
ferroxidase activity in Dps ferritin-like proteins. Protein Sci 17:1513–1521.
146
98. Haikarainen T, Tsou C-C, Wu J-J, Papageorgiou AC. 2010. Structural characterization
and biological implications of di-zinc binding in the ferroxidase center of Streptococcus
pyogenes Dpr. Biochemical and Biophysical Research Communications 398:361–365.
99. Romão C v., Mitchell EP, McSweeney S. 2006. The crystal structure of Deinococcus
radiodurans Dps protein (DR2263) reveals the presence of a novel metal centre in the N
terminus. JBIC Journal of Biological Inorganic Chemistry 11:891–902.
100. Haikarainen T, Thanassoulas A, Stavros P, Nounesis G, Haataja S, Papageorgiou AC.
2011. Structural and thermodynamic characterization of metal ion binding in
Streptococcus suis Dpr. Journal of Molecular Biology 405:448–460.
101. Ardini M, Fiorillo A, Fittipaldi M, Stefanini S, Gatteschi D, Ilari A, Chiancone E. 2013.
Kineococcus radiotolerans Dps forms a heteronuclear Mn-Fe ferroxidase center that may
explain the Mn-dependent protection against oxidative stress. Biochimica et Biophysica
Acta - General Subjects 1830:3745–3755.
102. Johnson JL, Cannon M, Watt RK, Frankel RB, Watt GD. 1999. Forming the Phosphate
Layer in Reconstituted Horse Spleen Ferritin and the Role of Phosphate in Promoting
Core Surface Redox Reactions. Biochemistry 38:6706–6713.
103. Horsburgh MJ, Clements MO, Crossley H, Ingham E, Foster SJ. 2001. PerR Controls
Oxidative Stress Resistance and Iron Storage Proteins and Is Required for Virulence in
Staphylococcus aureus. INFECTION AND IMMUNITY 69:3744–3754.
104. Dowds BC, Murphy P, McConnell DJ, Devine KM. 1987. Relationship among oxidative
stress, growth cycle, and sporulation in Bacillus subtilis. J Bacteriol 169:5771–5775.
105. Wei Q, le Minh PN, Dötsch A, Hildebrand F, Panmanee W, Elfarash A, Schulz S,
Plaisance S, Charlier D, Hassett D, Häussler S, Cornelis P. 2012. Global regulation of
gene expression by OxyR in an important human opportunistic pathogen. Nucleic Acids
Research 40:4320–4333.
106. Ohniwa RL, Ushijima Y, Saito S, Morikawa K. 2011. Proteomic Analyses of Nucleoid-
Associated Proteins in Escherichia coli, Pseudomonas aeruginosa, Bacillus subtilis, and
Staphylococcus aureus. PLoS ONE 6:e19172.
107. Bsat N, Chen L, Helmann JD. 1996. Mutation of the Bacillus subtilis Alkyl
Hydroperoxide Reductase (ahpCF) Operon Reveals Compensatory Interactions among
Hydrogen Peroxide Stress GenesJournal of Bacteriology.
108. Reindel S, Schmidt CL, Anemüller S, Matzanke BF. 2005. Expression and Regulation
Pattern of Ferritin-like DpsA in the Archaeon Halobacterium salinarum. BioMetals
18:387–397.
147
109. Velayudhan J, Castor M, Richardson A, Main-Hester KL, Fang FC. 2007. The role of
ferritins in the physiology of Salmonella enterica sv. Typhimurium: a unique role for
ferritin B in iron-sulphur cluster repair and virulence. Molecular Microbiology 63:1495–
1507.
110. Shrivastava AK, Pandey S, Dietz KJ, Singh K, Singh S, Rai R, Rai LC. 2016.
Overexpression of AhpC enhances stress tolerance and N 2-fixation in Anabaena by
upregulating stress responsive genes https://doi.org/10.1016/j.bbagen.2016.07.031.
111. Altuvia S, Almirón M, Huisman G, Kolter R, Storz G. 1994. The dps promoter is
activated by OxyR during growth and by IHF and σS in stationary phase. Mol Microbiol
13:265–72.
112. Rothe M, Alpert C, Engst W, Musiol S, Loh G, Blaut M. 2012. Impact of nutritional
factors on the proteome of intestinal Escherichia coli: Induction of OxyR-dependent
proteins AhpF and Dps by a lactose-rich diet. Applied and Environmental Microbiology
78:3580–3591.
113. Toledano MB, Kullik I, Trinh F, Baird PT, Schneider TD, Storz G. 1994. Redox-
dependent shift of OxyR-DNA contacts along an extended DNA-binding site: a
mechanism for differential promoter selection. Cell 78:897–909.
114. Bsat N, Herbig A, Casillas‐Martinez L, Setlow P, Helmann JD. 1998. Bacillus subtilis
contains multiple Fur homologues: identification of the iron uptake (Fur) and peroxide
regulon (PerR) repressors. Molecular Microbiology 29:189–198.
115. de Castro Ferreira IG, Rodrigues MM, da Silva Neto JF, Mazzon RR, do Valle Marques
M. 2016. Role and regulation of ferritin-like proteins in iron homeostasis and oxidative
stress survival of Caulobacter crescentus. BioMetals 29:851–862.
116. Yu M, Ren J, Zeng Y, Zhou S, Lu Y. 2009. The Legionella pneumophila Dps homolog is
regulated by iron and involved in multiple stress tolerance. Journal of Basic Microbiology
49:S79–S86.
117. Pesek J, Büchler R, Albrecht R, Boland W, Zeth K. 2011. Structure and mechanism of
iron translocation by a Dps protein from Microbacterium arborescens. J Biol Chem
286:34872–34882.
118. Polidoro M, de Biase D, Montagnini B, Guarrera L, Cavallo S, Valenti P, Stefanini S,
Chiancone E. 2002. The expression of the dodecameric ferritin in Listeria spp. is induced
by iron limitation and stationary growth phase. Gene 296:121–128.
119. Wolfe AJ, Chang D-E, Walker JD, Seitz-Partridge JE, Vidaurri MD, Lange CF, Prüß BM,
Henk MC, Larkin JC, Conway T. 2003. Evidence that acetyl phosphate functions as a
global signal during biofilm development. Molecular Microbiology 48:977–988.
148
120. Williams E, Lowe TM, Savas J, DiRuggiero J. 2007. Microarray analysis of the
hyperthermophilic archaeon Pyrococcus furiosus exposed to gamma irradiation.
Extremophiles 11:19–29.
121. Nandal A, Huggins CCO, Woodhall MR, McHugh J, Rodríguez-Quiñones F, Quail MA,
Guest JR, Andrews SC. 2010. Induction of the ferritin gene ftnA of Escherichia coli by
Fe2+ -Fur is mediated by reversal of H-NS silencing and is RyhB independent. Molecular
Microbiology 75:637–657.
122. Stephani K, Weichart D, Hengge R. 2003. Dynamic control of Dps protein levels by
ClpXP and ClpAP proteases in Escherichia coli. Molecular Microbiology 49:1605–1614.
123. Flynn JM, Neher SB, Kim Y-I, Sauer RT, Baker TA. 2003. Proteomic Discovery of
Cellular Substrates of the ClpXP Protease Reveals Five Classes of ClpX-Recognition
Signals. Molecular Cell 11:671–683.
124. Ninnis RL, Spall SK, Talbo GH, Truscott KN, Dougan DA. 2009. Modification of
PATase by L/F-transferase generates a ClpS-dependent N-end rule substrate in
Escherichia coli. The EMBO Journal 28:1732–1744.
125. Schmidt R, Zahn R, Bukau B, Mogk A. 2009. ClpS is the recognition component for
Escherichia coli substrates of the N-end rule degradation pathway. Molecular
Microbiology 72:506–517.
126. Xia X, Larios-Valencia J, Liu Z, Xiang F, Kan B, Wang H, Zhu J. 2017. OxyR-Activated
expression of Dps is important for Vibrio cholerae oxidative stress resistance and
pathogenesis. PLoS ONE 12:e0171201.
127. Li X, Sandh G, Nenninger A, Muro-Pastor AM, Stensjö K. 2015. Differential
transcriptional regulation of orthologous dps genes from two closely related heterocyst-
forming cyanobacteria. FEMS Microbiology Letters 362:fnv017.
128. Moparthi VK, Li X, Vavitsas K, Dzhygyr I, Sandh G, Magnuson A, Stensjö K. 2016. The
two Dps proteins, NpDps2 and NpDps5, are involved in light-induced oxidative stress
tolerance in the N2-fixing cyanobacterium Nostoc punctiforme. Biochimica et Biophysica
Acta - Bioenergetics 1857:1766–1776.
129. Facey PD, Hitchings MD, Saavedra-Garcia P, Fernandez-Martinez L, Dyson PJ, del Sol
R. 2009. Streptomyces coelicolor Dps-like proteins: differential dual roles in response to
stress during vegetative growth and in nucleoid condensation during reproductive cell
division. Molecular Microbiology 73:1186–1202.
130. Chowdhury RP, Gupta S, Chatterji D. 2007. Identification and characterization of the dps
promoter of Mycobacterium smegmatis: Promoter recognition by stress-specific
extracytoplasmic function sigma factors σH and σF. Journal of Bacteriology 189:8973–
8981.
149
131. Ekman M, Sandh G, Nenninger A, Oliveira P, Stensjö K. 2014. Cellular and functional
specificity among ferritin-like proteins in the multicellular cyanobacterium Nostoc
punctiforme. Environmental Microbiology 16:829–844.
132. Diaz PI, Slakeski N, Reynolds EC, Morona R, Rogers AH, Kolenbrander PE. 2006. Role
of oxyR in the oral anaerobe Porphyromonas gingivalis. Journal of Bacteriology
188:2454–2462.
133. McLeod SM, Johnson RC. 2001. Control of transcription by nucleoid proteins. Current
Opinion in Microbiology 4:152–159.
134. Azam TA, Hiraga S, Ishihama A. 2000. Two types of localization of the DNA-binding
proteins within the Escherichia coli nucleoid. Genes Cells 5:613–626.
135. Calhoun LN, Kim JN, Ren Y, Song JJ, Kwon M. 2011. The DNA-binding protein Dps
functions as a global regulator in starved Salmonella enterica serovar enteritidis during
starvation. International Journal of Microbiology Research 3:136–147.
136. Narayan OP, Kumari N, Bhargava P, Rajaram H, Rai LC. 2016. A single gene all3940
(Dps) overexpression in Anabaena sp. PCC 7120 confers multiple abiotic stress tolerance
via proteomic alterations. Functional and Integrative Genomics 16:67–78.
137. Janissen R, Arens MMA, Vtyurina NN, Rivai Z, Sunday ND, Eslami-Mossallam B,
Gritsenko AA, Laan L, de Ridder D, Artsimovitch I, Dekker NH, Abbondanzieri EA,
Meyer AS. 2018. Global DNA Compaction in Stationary-Phase Bacteria Does Not Affect
Transcription. Cell 174:1188–1199.
138. MN M, T M, R S, R F, IM E. 2018. The complex underpinnings of genetic background
effects. Nat Commun 9.
139. Harrison P, Fischbach F, Hoy T, Haggis G. 1967. Ferric oxyhydroxide core of ferritin.
Nature 216:1188–1190.
140. Dittmann C, Han HM, Grabenbauer M, Laue M. 2015. Dormant Bacillus spores protect
their DNA in crystalline nucleoids against environmental stress. Journal of Structural
Biology 191:156–164.
141. Frenkiel-Krispin D, Sack R, Englander J, Shimoni E, Eisenstein M, Bullitt E, Horowitz-
Scherer R, Hayes CS, Setlow P, Minsky A, Wolf SG. 2004. Structure of the DNA-SspC
complex: Implications for DNA packaging, protection, and repair in bacterial spores.
Journal of Bacteriology 186:3525–3530.
142. Abbondanzieri EA, Meyer AS. 2019. More than just a phase: the search for membraneless
organelles in the bacterial cytoplasm. Current Genetics. Springer Verlag
https://doi.org/10.1007/s00294-018-00927-x.
150
143. Boeynaems S, Alberti S, Fawzi NL, Mittag T, Polymenidou M, Rousseau F, Schymkowitz
J, Shorter J, Wolozin B, van den Bosch L, Tompa P, Fuxreiter M. 2018. Protein Phase
Separation: A New Phase in Cell Biology. Trends in Cell Biology. Elsevier Ltd
https://doi.org/10.1016/j.tcb.2018.02.004.
144. Vtyurina NN, Dulin D, Docter MW, Meyer AS, Dekker NH, Abbondanzieri EA. 2016.
Hysteresis in DNA compaction by Dps is described by an Ising model. Proc Natl Acad Sci
U S A 113:4982–7.
145. Riback JA, Katanski CD, Kear-Scott JL, Pilipenko E v., Rojek AE, Sosnick TR,
Drummond DA. 2017. Stress-Triggered Phase Separation Is an Adaptive, Evolutionarily
Tuned Response. Cell 168:1028-1040.e19.
146. Chen L, Helmann JD. 1995. Bacillus subtilis MrgA is a Dps (PexB) homologue: evidence
for metalloregulation of an oxidative-stress gene. Molecular Microbiology 18:295–300.
147. Martinez A, Kolter R. 1997. Protection of DNA during Oxidative Stress by the
Nonspecific DNA-Binding Protein Dps. JOURNAL OF BACTERIOLOGY 179:5188–
5194.
148. Ljungman M, Hanawalt PC. 1992. Efficient protection against oxidative DNA damage in
chromatin. Mol Carcinog 5:264–269.
149. Chodavarapu S, Gomez R, Vicente M, Kaguni JM. 2008. Escherichia coli Dps interacts
with DnaA protein to impede initiation: a model of adaptive mutation. Molecular
Microbiology 67:1331–1346.
150. Silva LSO, Baptista JM, Batley C, Andrews SC, Saraiva LM. 2018. The di-iron RIC
protein (YtfE) of Escherichia coli interacts with the DNA-binding protein from starved
cells (Dps) to diminish RIC protein-mediated redox stress. Journal of Bacteriology
200:e00527-18.
151. Lacqua A, Wanner O, Colangelo T, Martinotti MG, Landini P. 2006. Emergence of
Biofilm-Forming Subpopulations upon Exposure of Escherichia coli to Environmental
Bacteriophages. APPLIED AND ENVIRONMENTAL MICROBIOLOGY 72:956–959.
152. Theoret JR, Cooper KK, Zekarias B, Roland KL, Law BF, Curtiss R, Joens LA. 2012. The
Campylobacter jejuni dps homologue is important for in vitro biofilm formation and cecal
colonization of poultry and may serve as a protective antigen for vaccination. Clinical and
Vaccine Immunology 19:1426–1431.
153. Wu L, Lin X, Peng X. 2009. From Proteome to Genome for Functional Characterization
of pH-Dependent Outer Membrane Proteins in Escherichia coli. Journal of Proteome
Research 8:1059–1070.
154. Li H, Wang BC, Xu WJ, Lin XM, Peng XX. 2008. Identification and Network of Outer
Membrane Proteins Regulating Streptomysin Resistance in Escherichia coli. Journal of
Proteome Research 7:4040–4049.
151
155. Evans DJ, Evans DG, Takemura T, Nakano H, Lampert HC, Graham DY, Granger DN,
Kvietys PR. 1995. Characterization of a Helicobacter pylori neutrophil-activating protein.
Infect Immun 63:2213–2220.
156. Montemurro P, Nishioka H, Dundon WG, Bernard M de, Giudice G del, Rappuoli R,
Montecucco C. 2002. The neutrophil‐activating protein (HP‐NAP) of Helicobacter pylori
is a potent stimulant of mast cells. European Journal of Immunology 32:671–676.
157. Zanotti G, Papinutto E, Dundon WG, Battistutta R, Seveso M, Giudice G del, Rappuoli R,
Montecucco C. 2002. Structure of the Neutrophil-activating Protein from Helicobacter
pylori. Journal of Molecular Biology 323:125–130.
158. Halsey TA, Vazquez-Torres A, Gravdahl DJ, Fang FC, Libby SJ. 2004. The Ferritin-Like
Dps Protein Is Required for Salmonella enterica Serovar Typhimurium Oxidative Stress
Resistance and Virulence. INFECTION AND IMMUNITY 72:1155–1158.
159. Thumiger A, Polenghi A, Papinutto E, Battistutta R, Montecucco C, Zanotti G. 2005.
Crystal structure of antigen TpF1 from Treponema pallidum. Proteins: Structure,
Function, and Bioinformatics 62:827–830.
160. Olsen KN, Larsen MH, Gahan CGM, Kallipolitis B, Wolf XA, Rea R, Hill C, Ingmer H.
2005. The Dps-like protein Fri of Listeria monocytogenes promotes stress tolerance and
intracellular multiplication in macrophage-like cells. Microbiology (N Y) 151:925–933.
161. Amedei A, Cappon A, Codolo G, Cabrelle A, Polenghi A, Benagiano M, Tasca E, Azzurri
A, D’Elios MM, del Prete G, de Bernard M. 2006. The neutrophil-activating protein of
Helicobacter pylori promotes Th1 immune responses. J Clin Invest 116:1092–1101.
162. Polenghi A, Bossi F, Fischetti F, Durigutto P, Cabrelle A, Tamassia N, Cassatella M,
Montecucco C, Tedesco F, de Bernard M. 2007. The neutrophil-activating protein of
Helicobacteria pylori crosses endothelia to promote neutrophil adhesion in vivo. The
Journal of Immunology 178:1312–1320.
163. Kottakis F, Befani C, Asiminas A, Kontou M, Koliakos G, Choli-Papadopoulou T. 2009.
The C-terminal Region of HPNAP Activates Neutrophils and Promotes Their Adhesion to
Endothelial Cells. Helicobacter 14:177–179.
164. Zughaier SM, Cornelis P. 2018. Editorial: Role of Iron in Bacterial Pathogenesis.
Frontiers in Cellular and Infection Microbiology 8.
165. Ciacci C, Sabbatini F, Cavallaro R, Castiglione F, di Bella S, Iovino P, Palumbo A,
Tortora R, Amoruso D, Mazzacca G. 2004. Helicobacter pylori impairs iron absorption in
infected individuals. Digestive and Liver Disease 36:455–460.
166. Durham KA, Bullerjahn GS. 2002. Immunocytochemical localization of the stress-
induced DpsA protein in the cyanobacterium Synechococcus sp. strain PCC 7942. Journal
of Basic Microbiology 42:367–372.
152
167. Yamashita I, Iwahori K, Kumagai S. 2010. Ferritin in the field of nanodevices.
Biochimica et Biophysica Acta (BBA) - General Subjects 1800:846–857.
168. He D, Marles-Wright J. 2015. Ferritin family proteins and their use in bionanotechnology.
New Biotechnology 32:651–657.
169. Gálvez N, Sánchez P, Domínguez-Vera JM. 2005. Preparation of Cu and CuFe Prussian
Blue derivative nanoparticles using the apoferritin cavity as nanoreactor. Dalton
Transactions 2492–2494.
170. Kramer RM, Sowards LA, Pender MJ, Stone MO, Naik RR. 2005. Constrained Iron
Catalysts for Single-Walled Carbon Nanotube Growth. Langmuir 21:8466–8470.
171. Kang S, Lucon J, Varpness ZB, Liepold L, Uchida M, Willits D, Young M, Douglas T.
2008. Monitoring Biomimetic Platinum Nanocluster Formation Using Mass Spectrometry
and Cluster-Dependent H2 Production. Angewandte Chemie International Edition
47:7845–7848.
172. Cornell TA, Ardejani MS, Fu J, Newland SH, Zhang Y, Orner BP. 2018. A Structure-
Based Assembly Screen of Protein Cage Libraries in Living Cells: Experimentally
Repacking a Protein-Protein Interface to Recover Cage Formation in an Assembly-
Frustrated Mutant. Biochemistry 57:604–613.
173. Okuda M, Suzumoto Y, Iwahori K, Kang S, Uchida M, Douglas T, Yamashita I. 2010.
Bio-templated CdSe nanoparticle synthesis in a cage shaped protein, Listeria-Dps, and
their two dimensional ordered array self-assembly. Chemical Communications 46:8797–
8799.
174. Budiman MF, Hu W, Igarashi M, Tsukamoto R, Isoda T, Itoh KM, Yamashita I,
Murayama A, Okada Y, Samukawa S. 2012. Control of optical bandgap energy and
optical absorption coefficient by geometric parameters in sub-10 nm silicon-nanodisc
array structure. Nanotechnology 23:065302.
175. Li CQ, Soistman E, Carter DC. 2006. Ferritin nanoparticle technology. A new platform
for antigen presentation and vaccine development. Industrial Biotechnology 2:143–147.
176. Wu H, Wang J, Wang Z, Fisher DR, Lin Y. 2008. Apoferritin-Templated Yttrium
Phosphate Nanoparticle Conjugates for Radioimmunotherapy of Cancers. Journal of
Nanoscience and Nanotechnology 8:2316–2322.
177. Uchida M, Terashima M, Cunningham CH, Suzuki Y, Willits DA, Willis AF, Yang PC,
Tsao PS, McConnell M v., Young MJ, Douglas T. 2008. A human ferritin iron oxide
nano-composite magnetic resonance contrast agent. Magnetic Resonance in Medicine
60:1073–1081.
178. Gupta AK, Gupta M. 2005. Synthesis and surface engineering of iron oxide nanoparticles
for biomedical applications. Biomaterials 26:3995–4021.
153
179. Dilweg IW, Dame RT. 2018. Post-translational modification of nucleoid-associated
proteins: an extra layer of functional modulation in bacteria? Biochemical Society
Transactions 46:1381–1392.
180. Hanna ES, Roque-Barreira M-C, Bernardes ES, Panunto-Castelo A, Sousa M v, Almeida
IC, Brocchi M. 2007. Evidence for glycosylation on a DNA-binding protein of Salmonella
enterica. Microbial Cell Factories 6:11.
181. Stillman TJ, Hempstead PD, Artymiuk PJ, Andrews SC, Hudson AJ, Treffry A, Guest JR,
Harrison PM. 2001. The high-resolution X-ray crystallographic structure of the ferritin
(EcFtnA) of Escherichia coli; comparison with human H ferritin (HuHF) and the
structures of the Fe3+ and Zn2+ derivatives. Journal of Molecular Biology 307:587–603.
182. Crow A, Lawson TL, Lewin A, Moore GR, le Brun NE. 2009. Structural basis for iron
mineralization by bacterioferritin. J Am Chem Soc 131:6808–6813.
183. Papinutto E, Dundon WG, Pitulis N, Battistutta R, Montecucco C, Zanotti G. 2002.
Structure of two iron-binding proteins from Bacillus anthracis. J Biol Chem 277:15093–
15098.
184. Santos SP, Mitchell EP, Franquelim HG, Castanho MARB, Abreu IA, Romão C v. 2015.
Dps from Deinococcus radiodurans: oligomeric forms of Dps1 with distinct cellular
functions and Dps2 involved in metal storage. FEBS Journal 282:4307–4327.
185. Cuypers MG, Mitchell EP, Romão C v., McSweeney SM. 2007. The Crystal Structure of
the Dps2 from Deinococcus radiodurans Reveals an Unusual Pore Profile with a Non-
specific Metal Binding Site. Journal of Molecular Biology 371:787–799.
186. Chiancone E, Ilari A, Stefanini S, Tsernoglou D. 2000. The dodecameric ferritin from
Listeria innocua contains a novel intersubunit iron-binding site. Nature Structural Biology
7:38–43.
187. Gupta S, Pandit SB, Srinivasan N, Chatterji D. 2002. Proteomics analysis of carbon-
starved Mycobacterium smegmatis: induction of Dps-like protein. Protein Engineering,
Design and Selection 15:503–511.
188. Williams SM, Chatterji D. 2017. Flexible aspartates propel iron to the ferroxidation sites
along pathways stabilized by a conserved arginine in Dps proteins from: Mycobacterium
smegmatis. Metallomics 9.
189. Howe C, Ho F, Nenninger A, Raleiras P, Stensjö K. 2018. Differential biochemical
properties of three canonical Dps proteins from the cyanobacterium Nostoc punctiforme
suggest distinct cellular functions. Journal of Biological Chemistry 293.
190. Moparthi VK, Moparthi SB, Howe C, Raleiras P, Wenger J, Stensjö K. 2019. Structural
diffusion properties of two atypical Dps from the cyanobacterium Nostoc punctiforme
disclose interactions with ferredoxins and DNA. Biochimica et Biophysica Acta -
Bioenergetics 1860.
154
191. Ushijima Y, Ohniwa RL, Maruyama A, Saito S, Tanaka Y, Morikawa K. 2014. Nucleoid
compaction by MrgAAsp56Ala/Glu60Ala does not contribute to staphylococcal cell
survival against oxidative stress and phagocytic killing by macrophages. FEMS
Microbiology Letters 360:144–151.
192. Niven DF, Ekins A. 2001. Iron content of Streptococcus suis and evidence for a dpr
homologue. Canadian Journal of Microbiology 47:412–416.
193. Hitchings MD, Townsend P, Pohl E, Facey PD, Jones DH, Dyson PJ, del Sol R. 2014. A
tale of tails: Deciphering the contribution of terminal tails to the biochemical properties of
two Dps proteins from Streptomyces coelicolor. Cellular and Molecular Life Sciences
71:4911–4926.
194. Choi SH, Baumler DJ, Kaspar CW. 2000. Contribution of dps to acid stress tolerance and
oxidative stress tolerance in Escherichia coli O157:H7. Appl Environ Microbiol 66:3911–
3916.
195. Barth E, Gora K v, Gebendorfer KM, Settele F, Jakob U, Winter J. 2009. Interplay of
cellular cAMP levels, σS activity and oxidative stress resistance in Escherichia coli.
Microbiology (N Y) 155:1680–1689.
196. Tsou CC, Chiang-Ni C, Lin YS, Chuang WJ, Lin MT, Liu CC, Wu JJ. 2008. An iron-
binding protein, Dpr, decreases hydrogen peroxide stress and protects Streptococcus
pyogenes against multiple stresses. Infect Immun 76:4038–4045.
197. Narayan OP, Kumari N, Rai LC. 2010. Heterologous expression of Anabaena PCC 7120
all3940 (a Dps family gene) protects Escherichia coli from nutrient limitation and abiotic
stresses. Biochemical and Biophysical Research Communications 394:163–169.
198. Hébraud M, Guzzo J. 2000. The main cold shock protein of Listeria monocytogenes
belongs to the family of ferritin-like proteins. FEMS Microbiology Letters 190:29–34.
199. Nicodeme M, Perrin C, Hols P, Bracquart P, Gaillard J-L. 2004. Identification of an Iron-
Binding Protein of the Dps Family Expressed by Streptococcus thermophilus. Current
Microbiology 48:51–56.
200. Perrin C, Guimont C, Bracquart P, Gaillard JL. 1999. Expression of a New Cold Shock
Protein of 21.5 kDa and of the Major Cold Shock Protein by Streptococcus thermophilus
After Cold Shock. Current Microbiology 39:342–347.
201. Thieme D, Grass G. 2010. The Dps protein of Escherichia coli is involved in copper
homeostasis. Microbiological Research 165:108–115.
202. Malone AS, Chung Y-K, Yousef AE. 2006. Genes of Escherichia coli O157:H7 That Are
Involved in High-Pressure Resistance. APPLIED AND ENVIRONMENTAL
MICROBIOLOGY 72:2661–2671.
155
203. Li X, Mustila H, Magnuson A, Stensjö K. 2018. Homologous overexpression of NpDps2
and NpDps5 increases the tolerance for oxidative stress in the multicellular
cyanobacterium Nostoc punctiforme. FEMS Microbiology Letters 365:fny198.
204. Pacello F, Ceci P, Ammendola S, Pasquali P, Chiancone E, Battistoni A. 2008.
Periplasmic Cu,Zn superoxide dismutase and cytoplasmic Dps concur in protecting
Salmonella enterica serovar Typhimurium from extracellular reactive oxygen species.
Biochimica et Biophysica Acta (BBA) - General Subjects 1780:226–232.
205. Yamamoto Y, Higuchi M, Poole LB, Kamio Y. 2000. Role of the dpr product in oxygen
tolerance in Streptococcus mutans. J Bacteriol 182:3740–3747.
206. Brenot A, King KY, Caparon MG. 2004. The PerR regulon in peroxide resistance and
virulence of Streptococcus pyogenes. Molecular Microbiology 55:221–234.
207. Pulliainen AT, Kauko A, Haataja S, Papageorgiou AC, Finne J. 2005. Dps/Dpr ferritin-
like protein: insights into the mechanism of iron incorporation and evidence for a central
role in cellular iron homeostasis in Streptococcus suis. Molecular Microbiology
57:1086–1100.
208. Bradley JM, Le Brun NE, Moore GR. 2016. Ferritins: furnishing proteins with iron. J Biol
Inorg Chem 21:13–28.
209. Abdul-Tehrani H, Hudson AJ, Chang Y-S, Timms AR, Hawkins C, Williams JM,
Harrison PM, Guest JR, Andrews SC. 1999. Ferritin Mutants of Escherichia coli Are Iron
Deficient and Growth Impaired, and fur Mutants are Iron Deficient. Journal of
Bacteriology 181:1415–1428.
210. Andrews SC, Robinson AK, Rodríguez-Quiñones F. 2003. Bacterial Iron Homeostasis.
FEMS Microbiology Reviews 27:215–237.
211. Finkel SE. 2006. Long-term survival during stationary phase: evolution and the GASP
phenotype. Nature Reviews Microbiology 4:113–120.
212. Kram KE, Henderson AL, Finkel SE. 2020. Escherichia coli Has a Unique Transcriptional
Program in Long-Term Stationary Phase Allowing Identification of Genes Important for
Survival https://doi.org/10.1128/mSystems.00364-20.
213. Carrondo MA. 2003. Ferritins, iron uptake and storage from the bacterioferritin viewpoint.
The EMBO Journal 22:1959.
214. Zambrano MM, Siegele DA, Almirón M, Tormo A, Kolter R. 1993. Microbial
competition: Escherichia coli mutants that take over stationary phase cultures. Science
(1979) 259:1757–1760.
215. Nair S. 2005. Role of Dps in protection against various stresses and maintaining iron
homeostasis during stationary phase. Los Angeles.
156
216. Schneider CA, Rasband WS, Eliceiri KW. 2012. NIH Image to ImageJ: 25 years of image
analysis. Nature Methods 2012 9:7 9:671–675.
217. Bailey TL, Elkan C. 1994. Fitting a mixture model by expectation maximization to
discover motifs in biopolymers.
218. Pan CQ, Finkel SE, Cramton SE, Feng JA, Sigman DS, Johnson RC. 1996. Variable
structures of Fis-DNA complexes determined by flanking DNA-protein contacts. Journal
of Molecular Biology 264:675–695.
219. Gupta S, Stamatoyannopoulos JA, Bailey TL, Noble WS. 2007. Quantifying similarity
between motifs. Genome Biology 8:1–9.
220. Goodrich JA, Schwartz ML, Mcclure WR. 1990. Searching for and predicting the activity
of sites for DNA binding proteins: Compilation and analysis of the binding sites for
Escherichia coli integration host factor (IHF). Nucleic Acids Research 18:4993–5000.
221. Arakawa K, Uno R, Nakayama Y, Tomita M. 2007. Validating the significance of
genomic properties of Chi sites from the distribution of all octamers in Escherichia coli.
Gene 392:239–246.
222. Finkel SE, Johnson RC. 1992. The Fis protein: it’s not just for DNA inversion anymore.
Molecular Microbiology 6:3257–3265.
223. Marbouty M, Cournac A, Flot JF, Marie-Nelly H, Mozziconacci J, Koszul R. 2014.
Metagenomic chromosome conformation capture (meta3C) unveils the diversity of
chromosome organization in microorganisms. Elife 3:e03318.
Abstract (if available)
Abstract
The DNA-binding protein from starved cells, Dps, is a highly abundant protein during stationary phase in Escherichia coli. It combines two distinct biological functions: as a ferritin it detoxifies and stores iron, and as a DNA-binding protein it binds and structures the chromosome. These activities, though biochemically separable, function jointly to protect the cell from a variety of stresses, the most well-studied of which is oxidative stress. This dissertation focuses on both major aspects of Dps function. First, in comparing the growth and survival phenotypes of ferritin- null E. coli populations, growth phase-dependent importance of each ferritin gene is elucidated. While Dps has the most dramatic impact on growth and survival of the E. coli ferritins, iron availability experiments suggest this impact is only minimally due to its ferritin function, leading to the hypothesis that its major role in growth and survival is as a DNA-binding protein. However, Dps has been identified as a non-specific DNA-binding protein. As such, the next question addressed in this dissertation is whether Dps binds the same genomic loci in a repeatable manner. Using a novel assay to probe chromatin accessibility, Dps is shown here to bind at least a portion of the chromosome repeatably. Finally, MNase treatment followed by DNA sequencing (MNase- Seq) was used to determine chromosome accessibility on a more high-resolution scale. The two different chromatin accessibility experiments described in this dissertation reveal two putative Dps-DNA binding motifs. The first, determined by comparing Dps-dependent protected chromosomal regions between mid-log phase and late stationary phase, may denote a primary Dps binding motif. The second, determined by comparing Dps-dependent protected chromosomal regions shared in early and late stationary phase, may denote a secondary Dps binding motif.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Genomic and physiological characterization of Escherichia coli evolving in long-term batch culture
PDF
Bacterial survival under low nutrient conditions in the laboratory and the deep biosphere
PDF
Investigating the E. Coli DNA polymerase V mutasome
PDF
Genetic diversity and bacterial death in the context of adaptive evolution
PDF
Mechanisms of long-term survival of Vibrio harveyi and Escherichia coli: linking damage and senescence
PDF
Mechanisms of long-term survival in Escherichia coli
PDF
The function of Rpd3 in balancing the replicaton initiation of different genomic regions
PDF
Decrypting Escherichia coli DNA polymerase V mutasome ATP regulation
PDF
Molecular genetics of longevity in Escherichia coli
PDF
Physiological roles and evolutionary implications of alternative DNA polymerases in Escherichia coli
PDF
Characterizing and developing E. coli Type I-E CRISPR adaptation as a DNA recording and genome engineering tool
PDF
Nutritional competence in Escherichia coli: the use of double-stranded DNA as a nutrient
PDF
Genome-wide GATC methylation in long-term cultures of Escherichia coli K-12
PDF
Investigating highly mutagenic polymerases in pathogenic bacteria
PDF
The physics of emergent membrane phenomena
PDF
Maintenance of genome stability at fragile sites in Schizosaccharomyces pombe
PDF
From single molecules to bacterial nanowires: functional and dynamic imaging of the extracellular electron transfer network in Shewanella oneidensis MR-1
PDF
Genetic and molecular insights into the genotype-phenotype relationship
PDF
Structural and biochemical studies of large T antigen: the SV40 replicative helicase
PDF
Nanoscale dynamics and nuclear envelope organization of the muscular dystrophy related protein emerin
Asset Metadata
Creator
Orban, Katie
(author)
Core Title
Dps contributes to typical growth, survival, and genome organization in E. coli
School
College of Letters, Arts and Sciences
Degree
Doctor of Philosophy
Degree Program
Molecular Biology
Degree Conferral Date
2022-08
Publication Date
06/10/2022
Defense Date
04/26/2022
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
bacteria,DNA,Dps,E. coli,ferritin,Iron,nucleoid,OAI-PMH Harvest
Format
application/pdf
(imt)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Finkel, Steve (
committee chair
), Ehrenreich, Ian (
committee member
), El-Naggar, Moh (
committee member
), Phillips, Carolyn (
committee member
), Pinaud, Fabien (
committee member
)
Creator Email
korban@usc.edu,orban.katie@gmail.com
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-oUC111339681
Unique identifier
UC111339681
Document Type
Dissertation
Format
application/pdf (imt)
Rights
Orban, Katie
Internet Media Type
application/pdf
Type
texts
Source
20220613-usctheses-batch-946
(batch),
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 author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright. The original signature page accompanying the original submission of the work to the USC Libraries is retained by the USC Libraries and a copy of it may be obtained by authorized requesters contacting the repository e-mail address given.
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Repository Email
cisadmin@lib.usc.edu
Tags
DNA
Dps
E. coli
ferritin
nucleoid