CENTRAL CARBON METABOLISM REGULATES CELLULOSE PRODUCTION IN THE
PLANT PATHOGEN ERWINIA AMYLOVORA

By
Jared Zaporski

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Plant Pathology – Master of Science
2023



ABSTRACT
The bacterium Erwinia amylovora is a gram-negative pathogen of rose family plants
worldwide and is the causal agent of the devastating fire blight disease, causing substantial
damage to apple and pear trees annually. Cellulose is an exopolysaccharide produced by Erwinia
amylovora from host carbohydrates that improves the three-dimensional structure of biofilms in
the xylem of apple host plants, contributing to the devastating fire blight disease. Cellulose
biosynthesis is regulated by the second messenger c-di-GMP, which is produced and degraded
by diguanylate cyclases (DGCs) and specific phosphodiesterases (PDEs), respectively. While
several DGCs and PDEs have the potential to contribute to cellulose production, the specificity
of these enzymes is poorly known. In this study, host carbohydrates were added to growth media
in vitro, and knockout mutant strains were used to examine the specificity of PDEs and a DGC
with respect to cellulose production under carbohydrate-rich conditions. Experimental results
indicate that glucose, sucrose, and sorbitol each resulted in an increase in cellulose production,
which caused bacteria to engage in autoaggregation behavior in liquid medium. The pdeC gene
had the strongest negative impact on cellulose production out of the E. amylovora PDEs, while
mutation of dgcB abolished cellulose production. A mutant strain unable to metabolize sorbitol
was unable to produce cellulose on sorbitol-containing medium, and sugar fermentation occurred
in all sugar-enriched media, indicating that fermentative metabolism may favor cellulose
production in specific host environments such as the xylem.



This work is dedicated to Naomi
and also Martin, Julie, Ronald, and Pauline

iii



ACKNOWLEDGEMENTS
I would first like to thank my advisor, Dr. George Sundin, for his advice, encouragement,
and especially his patience. My graduate studies took many twists, turns, and dead ends and he
was willing to work with me every step of the way. His humor and enthusiasm for bacteria and
pudding stones is always uplifting.
I would also like to thank my graduate committee members Dr. Quan Zeng and Dr. Sarah
Lebeis for their teaching, support, and consultation during my time as a student at MSU. Their
voluntary donation of time and expertise bears testament to their dedication and generosity.
Bacterial genetics and molecular biology would still be a complete mystery for me
without the help and mentorship of Dr. Roshni Kharadi and Dr. Xiaochen Yuan, both of whom
taught me everything I know about cloning and various other laboratory procedures essential for
this work. In addition, lab members Dr. Kristi Gdanetz MacCready, Erin Lauwers, Katherine
Olive, Lindsay Brown, Cory Outwater, and Dr. Tyre Proffer contributed in many ways to my
graduate school experience, and I am grateful to each of them for it.
Special thanks also to Dr. Ziming Yang (Oakland University), who saw potential in me
before I even knew what a pipet was, and introduced me to the research world.
Finally, I would like to thank my wife Naomi and our families for their love and support.

iv



TABLE OF CONTENTS
LIST OF SYMBOLS ........................................................................................................ vi
LIST OF ABBREVIATIONS .......................................................................................... vii
Chapter 1. Introduction and Literature Review .................................................................. 1
1.1. Fire blight disease and the causal agent Erwinia amylovora ...........................2
1.2. Virulence factors of E. amylovora ...................................................................2
1.3. Bacterial cellulose production ..........................................................................5
1.4. Cyclic di GMP signaling in E. amylovora .......................................................7
1.5. Other mechanisms of virulence regulation ....................................................12
1.6. Objectives ......................................................................................................14
Chapter 2. Fermentative sugar metabolism stimulates cellulose production
in Erwinia amylovora .......................................................................................................15
2.1. Abstract ..........................................................................................................16
2.2. Introduction ....................................................................................................17
2.3. Materials and Methods ................................................................................... 20
2.4. Results ............................................................................................................ 28
2.5. Discussion ......................................................................................................45
REFERENCES .................................................................................................................54
APPENDIX A: BIOINFORMATIC OBSERVATIONS ...................................................62
APPENDIX B: DOMAIN-DIRECTED MUTAGENESIS ..............................................65
APPENDIX C: AMINO ACID SCREENING ..................................................................68

v



LIST OF SYMBOLS
α - alpha
β - beta
µ - micro
Δ - delta

vi



LIST OF ABBREVIATIONS
c-di-GMP – cyclic dimeric guanosine monophosphate
CFU – colony forming unit
DGC – diguanylate cyclase
EPS – exopolysaccharide
LB – Luria-Bertani medium or Lysogeny Broth
PDE - phosphodiesterase
ROS – reactive oxygen species
sRNA – small RNA
T3SS – type 3 secretion system
TCA – tricarboxylic acid
TCS – two component system

vii



Chapter 1. Introduction and Literature Review

1



1.1. Fire blight disease and the causal agent Erwinia amylovora
The Gram-negative bacterium Erwinia amylovora has been named one of the top 10 most
important plant pathogenic bacteria for causing fire blight disease on apple, pear, and other
Rosaceous plants (Mansfield et al., 2012). One of the first bacterial pathogens known, E.
amylovora is a member of the family Enterobacteriaceae, similar to other plant and animal
pathogens (Piqué et al., 2015). The disease caused by this bacterium is complex and devastating,
with severe outbreaks leading to the death of entire apple and pear trees and costing local and
global economies millions of dollars annually (Norelli et al., 2003). The disease cycle begins in
spring, when E. amylovora is transferred to the flower stigma by wind, rain, or insects.
Populations on the stigma can grow up to 1 x 106 CFU/flower before migrating to the floral cup
and entering the plant through the nectaries (Slack et al., 2022). Systemic infection then occurs,
with bacterial population within the plant causing ‘scorched’ symptoms on flowers, leaves, and
shoots. E. amylovora forms biofilms in the xylem tissue of leaves and shoots, blocking water
flow and contributing to the distinctive symptoms (Koczan et al., 2009). Secondary infections
can occur via wounds and natural openings. Eventually, the bacteria form cankers on the trunk
and crown of the tree, which can cause tree death and also provides an overwintering site for the
pathogen (Norelli et al., 2003). Periodically throughout the disease cycle, masses of bacterial
cells, exopolysaccharides (EPS), and sugars will burst through the plant tissue to form ooze
droplets on the surface, which serve as a source of inoculum for further infection (Bennett &
Billing, 1980; Slack et al., 2017).
1.2. Virulence factors of E. amylovora
During the various infection stages on different host tissues, E. amylovora deploys a
number of virulence factors to cause disease. During flower colonization, flagellar motility is

2



important for the migration of the bacteria from the stigma into the hypanthium of the flower
(Schachterle, Zeng, et al., 2019). The bacteria also induce a ROS burst from the host immune
system, but are able to overcome the oxidative effects due to the activity of catalase enzymes
(Schachterle, Onsay, et al., 2019). In addition to catalases, the siderophore desferrioxamine has
shown to be a key factor contributing to survival of ROS production in E. amylovora, and also
plays a key role in scavenging bioavailable iron, a limiting nutrient in apple flowers (Dellagi et
al., 1998).
An important virulence factor in E. amylovora is the ability to metabolize host
carbohydrates. In the apple host, several carbohydrates are consumed over the course of the
disease cycle. Analysis of the stigma tip revealed that glucose and fructose contributed to sigma
colonization, while in the nectar, sucrose is the primary sugar (Pusey et al., 2008). Interestingly,
while sucrose is the primary transport carbohydrate in most plants, apple, pear, and other hosts of
E. amylovora contain sorbitol as the main carbohydrate (Aldridge et al., 1997). While glucose
and fructose can be immediately processed via glycolysis, sucrose and sorbitol require specific
enzymes to transport and convert these sugars into glycolytic intermediates. The five-gene scr
operon is responsible for sucrose uptake and metabolism in E. amylovora, and deletion of these
genes reduced disease symptoms in apple seedlings (Bogs & Geider, 2000).Similarly, the srl
operon encodes six genes that phosphorylate and translocate sorbitol molecules into the cell,
where sorbitol-6-phosphate 2-dehydrogenase (SrlD) converts sorbitol-6-phosphate to fructose-6phosphate, which is then able to be further processed by glycolysis (Salomone-Stagni et al.,
2018). Mutant strains lacking the srl operon were found to have greatly decreased virulence in
apple plants, highlighting the importance of sorbitol metabolism in fire blight disease
development (Aldridge et al., 1997).

3



A facultative anaerobe, E. amylovora can metabolize carbon using oxidative metabolism
and the Tricarboxilic Acid (TCA) cycle, or fermentative metabolism. Fermentation pathways
may be very important for E. amylovora growth and survival, especially in anaerobic or semianaerobic conditions. A study conducted on E. amylovora glucose metabolism in aerobic and
semi-anaerobic conditions found that, in both conditions, ethanol, lactate, and carbon dioxide
were the primary metabolic products (Sutton & Starr, 1959). In addition, smaller amounts of
acetic, formic, and succinic acid were also detected in this study. The high yields of ethanol
formation reported by this study and others, even in aerobic conditions, can be explained by the
presence of pyruvate decarboxylase and ethanol dehydrogenase enzymes in E. amylovora. In
fact, the thiamine pyrophosphate-dependent pyruvate decarboxylase is quite rare among bacteria,
and E. amylovora was the first member of the Enterobacteriaceae reported to possess this
enzyme (Haq & Dawes, 1971). The fermentative pathway for carbon substrates seems to be an
important factor for growth in vitro for this pathogen, and researchers have speculated that the
bacteria may encounter different oxygen levels as they move from flower and shoot surfaces to
more interior tissues such as xylem (Schachterle, Onsay, et al., 2019). How carbon metabolism
affects the regulation of other virulence factors in E. amylovora remains poorly understood.
One of the most important virulence factors in E. amylovora is the type III secretion
system (T3SS), used to translocate effector proteins into the host. These effectors suppress host
immunity and cause cell death, and electrolyte leakage, contributing to the cells’ survival of
reactive oxygen species (ROS) produced by the host during early infection (Piqué et al., 2015). A
single effector, DspA/E, is largely responsible for these properties (Boureau et al., 2006).
Equally important for pathogenicity is the EPS amylovoran, a heteropolymer of glucose,
galactose, and pyruvate (Koczan et al., 2009). Mutagenesis of the ams gene cluster, responsible

4



for the biosynthesis of amylovoran, renders the bacteria unable to infect the host (Bellemann &
Geider, 1992). Amylovoran is an important component of biofilms, contributing to host
colonization, and is the major constituent of bacterial ooze, an important stage in the disease
cycle (Koczan et al., 2009; Slack et al., 2017). E. amylovora also produces the EPS levan, a
homopolymer of fructose, which also contributes to biofilm formation and is found in ooze and
together with amylovoran contributes to sliding motility (Koczan et al., 2009; Slack et al., 2017;
Yuan, Eldred, & Sundin, 2022). A third EPS, cellulose, has been more recently discovered to
play a role in virulence. Cellulose has been shown to be an important component of xylem
biofilms, contributing to their three-dimensional structure, and is important for full virulence in
planta (Castiblanco & Sundin, 2018). Interestingly, cellulose was not found to contribute to
sliding motility in E. amylovora, in contrast to amylovoran and levan (Yuan, Eldred, & Sundin,
2022). It is currently unknown if cellulose is only produced in xylem biofilms or is involved with
other aspects of the fire blight disease cycle.
1.3. Bacterial cellulose production
Cellulose is one of the most abundant biological polymers on Earth and is the primary
component of the cell wall in plants (Abidi et al., 2022). By no means limited to plants, however,
cellulose is produced by a wide diversity of bacteria as well (Römling & Galperin, 2015).
Although cellulose in its purest form is a homopolymer of glucose molecules connected by β-1,4
linkages, there are differences in the properties of bacterial and plant cellulose. For example,
cellulose from bacterial sources has been noted to display much higher purity, crystallinity,
malleability, and porosity than plant cellulose, which can contain similar polymers with different
bond structures such as lignin and hemicellulose (Naomi et al., 2020). These characteristics, in
addition to high tensile strength, biodegradability, and low toxicity, have led to great interest in

5



various applications of bacterial cellulose, including wound healing, drug delivery, tissue
grafting, toxin removal, soil additives, and even as a thickener and stabilizing agent in foods
(Mohite & Patil, 2014).
Bacterial cellulose polymers, as well as other bacterial exopolysaccharides, are key
constituents of biofilms, and can play a fundamental role in the biology of many pathogenic
bacteria. Cellulose is a component of Salmonella and Escherichia coli biofilms, contributing to
cell-cell interaction, macrocolony structure, biofilm architecture, and resistance to environmental
stresses (Römling & Galperin, 2015; Serra et al., 2013). Due to its insolubility and high
hydrophilicity, cellulose has significant capacity to trap and retain water. In Es. coli, cellulose
production increased resistance to osmotic stress and chlorine treatment in pathogenic strain
O157, and another study found exopolysaccharide overproduction in response to osmotic stress
in strain K-12 (Ionescu & Belkin, 2009; Yoo & Chen, 2010). Plant-pathogenic bacteria Dickya
dadantii and Agrobacterium tumefaciens produce cellulose to aid colonization of plant surfaces,
and cellulose production enhances phyllosphere persistence in epiphytic Pseudomonas
fluorescens (Römling & Galperin, 2015). In E. amylovora, cellulose was discovered to play a
major role in the architecture of xylem biofilms, modulating the three-dimensional structure in
vitro and in planta (Castiblanco & Sundin, 2018). This study also reported that strains unable to
produce cellulose were less virulent on apple shoots, demonstrating the importance of cellulose
as a virulence factor in a shoot blight infection model.
Bacterial cellulose synthase (Bcs) proteins are responsible for the synthesis and
translocation of bacterial cellulose. These Bcs proteins form a multiprotein complex spanning the
inner and outer membranes and are encoded by the bcs operon, typically consisting of three to
seven genes (Römling & Galperin, 2015). The protein products of bcsA and bcsB are essential

6



for cellulose production, while the other associated bcs genes are involved in regulation,
transport and translocation of the cellulose polymer, or have uncharacterized functions (Römling
& Galperin, 2015). According to proposed classification, E. amylovora contains a Type 1b bcs
operon, containing the periplasmic protein bcsD as well as the putative endogluconase bcsZ
(Castiblanco & Sundin, 2018; Römling & Galperin, 2015).
The Bcs system was famously discovered to be allosterically activated by the ubiquitous
second messenger molecule Cyclic dimeric Guanosine Monophosphate (c-di-GMP). The BcsA
protein is one of the most well-studied subunits of the Bcs system because it contains a Cterminal cytoplasmic PilZ domain, a known receptor for c-di-GMP (Ryjenkov et al., 2006).
Upon allosteric binding of c-di-GMP to the PilZ domain of BcsA, an intermolecular salt bridge is
broken which releases an inhibitory gating loop, resulting in access of the UDP-glucose substrate
to the enzymatic active site of BcsA (Morgan et al., 2014). Because of the nature of the BcsA
gating loop, the Bcs complex is constitutively inactive in the absence of c-di-GMP. In E.
amylovora, c-di-GMP strongly regulates cellulose production by binding to the PilZ domain on
BcsA. Mutation of a key amino acid in this binding site resulted in a lack of cellulose production,
even in the presence of high c-di-GMP levels (Castiblanco & Sundin, 2018).
1.4. Cyclic di GMP signaling in E. amylovora
Cyclic dimeric Guanosine Monophosphate (c-di-GMP) is a ubiquitous second messenger
molecule found in a diverse array of bacteria. First discovered in the 1970’s as an activator of
cellulose biosynthesis in Acetobacter xylinum, c-di-GMP has since become known as one of the
most complex and important prokaryotic signal transduction mechanisms, and genes related to cdi-GMP metabolism have been found in all major bacterial phyla (Römling et al., 2013; Ross et
al., 1973). Many important cellular functions, including motility, cell division, differentiation,

7



virulence, and biofilm formation are regulated by this key molecule (Römling et al., 2013).
Composed of two GMP molecules, c-di-GMP forms a relatively stable 12-member ribosephosphate ring structure, and is synthesized by diguanylate cyclase enzymes (DGCs). While
often having a diverse domain architecture, DGCs all share a GGDEF domain, named for the
conserved enzymatically active site responsible for the synthesis of c-di-GMP (Ausmees et al.,
2001). The GGDEF domain is enzymatically active as a homodimer, and the other components
of DGC proteins often have a significant effect on the ability of the proteins to dimerize,
influencing the amount of c-di-GMP output in the cell. Conversely, c-di-GMP is degraded by
specific phosphodiesterase enzymes (PDEs), which contain either an EAL or HD-GYP domain
that contains the active site. The more common EAL domain linearizes the c-di-GMP molecule
to the degradation product 5’-pGpG, whereas the HD-GYP domain is able to complete
hydrolysis and produce two molecules of GMP (Bobrov et al., 2005; Galperin et al., 1999). The
EAL domain does not require dimerization for activity, unlike the GGDEF domain, but its
activity is often increased as a dimer (Bobrov et al., 2005). Interestingly, tandem enzymes
containing both a GGDEF and an EAL domain appear to be relatively common. Some of these
are truly bifunctional with the ability to both synthesize and degrade c-di-GMP, but more often
one of the two domains is inactive due to a degenerate active site motif (Römling et al., 2013;
Tarutina et al., 2006).
While the GGDEF, EAL, and HD-GYP domains are conserved across different bacterial
lineages, the overall protein architecture of DGCs and PDEs is highly variable. Many bacterial
species have up to several dozen various c-di-GMP enzymes, indicating that the variety of
different protein structures plays a key role in c-di-GMP signal transduction. For example, the
model organism Escherichia coli contains 29 DGC and PDE genes with a variety of sensory

8



domains and proteins structures (Hengge et al., 2016). DGCs and PDEs can be part of onecomponent, two-component, or chemosensory systems, and may be cytoplasmic or membranebound, often with N-terminal sensory domains (Randall et al., 2022). The sensory domains
percieve a variety of external stimuli such as temperature, light, nutrients, oxidative state,
molecular oxygen levels, quorum sensing molecules, and others (Almblad et al., 2021; Enomoto
et al., 2023; Herbst et al., 2018; Huang et al., 2013; Rossi et al., 2022; Waldron et al., 2019).
Transmitting these signals into a c-di-GMP output, c-di-GMP enzymes may affect the entire cell
globally, or may provide more localized signal transduction that only affects a specific
downstream process (Reinders et al., 2016; Richter et al., 2020). However, even with these
recent advances in knowledge, most of the specific signals affecting the expression or function of
individual DGCs or PDEs remain a mystery, and many of the sensory domains are poorly
characterized.
Five DGC enzymes (edcA, edcB, edcC, edcD, edcE) have been characterized in the
genome of E. amylovora, possessing active GGDEF domains with all residues necessary for cdi-GMP synthesis (Edmunds et al., 2013). Except for edcA, each of the enzymes is predicted to
be membrane-associated with multiple transmembrane domains, and edcA, edcD, and edcE each
have predicted sensory domains at the N-terminus. In contrast, edcC and edcB do not contain a
predicted sensory domain, and only have multiple transmembrane domains and the GGDEF
domain. Overexpression of edcB. edcC, edcD, and edcE each increased intracellular c-di-GMP
levels, amylovoran production, and biofilm formation while decreasing swimming motility
(Edmunds et al., 2013). In addition, edcB, edcC, and edcD were shown to upregulate cellulose
production in E. amylovora (Castiblanco & Sundin, 2018). This study also suggests that, while cdi-GMP produced from DgcC and DgcE can activate cellulose biosynthesis in E. amylovora,

9



DgcB may be more specifically related to cellulose, as overexpression of this enzyme increased
the production of cellulose but not amylovoran. In addition, a transposon mutagenesis
experiment implicated DgcB as a key positive regulator of cellulose production in E. amylovora,
but the factors regulating the expression of the dgcB gene as well as the regulation of the PDE
enzymes remain unknown (Yuan, Eldred, Kharadi, et al., 2022).
E. amylovora also contains three active PDE enzymes (pdeA, pdeB, and pdeC). All three
PDEs have active EAL domains and are predicted to be membrane-associated, with pdeA and
pdeB containing an N-terminal CSS sensory domain commonly found in PDE enzymes (Kharadi
et al., 2019a). Interestingly, pdeC has a more complex protein architecture with a predicted Nterminal GAPES3 sensory domain, a HAMP domain usually involved in transmitting
conformational changes, and a degenerate GGDEF domain lacking the required active site
residues for c-di-GMP synthesis. The importance or function of the GAPES3 domain is not
known, even though it is found in other important pathogens such as Yersinia pestis.
Mutagenesis studies by Kharadi et al., (2019) revealed that PdeC is the most active PDE in E.
amylovora, with a knockout mutant showing significantly higher intercellular c-di-GMP levels.
However, all three PDE knockouts displayed similar increases in biofilm formation and
amylovoran production, while not appearing to significantly affect swimming motility. A triple
mutant lacking all three PDE genes showed the highest c-di-GMP levels, significantly increasing
amylovoran production while decreasing motility, ooze production, hrp gene expression, and
virulence on immature pears. This triple mutant also autoaggregated in liquid culture, a
phenotype associated with sessile growth and biofilm formation. The lack of distinct phenotype
in the pdeC single mutant, even with significantly increased c-di-GMP levels, is interesting and
could indicate that pdeC plays a more specific role in other aspects of virulence. However, the

10



effects of pdeC on other virulence factors, such as cellulose production, have not yet been
investigated.
Further delving into the complexity of c-di-GMP signaling in E. amylovora, (Kharadi et
al., 2022a) constructed a strain missing all five active DGCs and all three active PDEs (Δ8) and
another strain missing an additional four c-di-GMP related genes (Δ12). The authors discovered
that, despite high expression of the T3SS, the Δ12 strain is unable to effectively colonize the
apple shoot and is deficient in surface attachment and biofilm formation. In addition, the
transcriptome was substantially altered in the absence of c-di-GMP signaling genes, with over
500 genes differentially expressed, including genes for metabolic pathways and other regulatory
systems. Each DGC contributed to biofilm formation when complemented into the Δ8 strain,
suggesting that effective host colonization requires the concerted orchestration of all DGCs in E.
amylovora.
Biofilm formation has been one of the most well-studied c-di-GMP regulated processes,
and biofilms have also been implicated as an important part of the fire blight disease cycle
(Kharadi & Sundin, 2021). A related behavior in E. amylvora is autoaggregation, in which
visible bacterial clumps settle to the bottom of liquid culture media. Research on autoaggregation
of this pathogen discovered that the autoaggregation process is highly c-di-GMP dependent, but
does not necessarily correlate with biofilm formation. In fact, mutant strains that had high levels
of c-di-GMP and displayed autoaggregation showed greatly decreased biofilm formation
(Kharadi & Sundin, 2019). The authors attribute this interesting result to the increased propensity
for cell-cell interaction compared to cell-surface interaction. In addition, autoaggregation also
involved cell division genes sulA and ftsZ, with decreased expression of the cytoskeletal ftsZ
gene. One of the key factors contributing to autoaggregation in E. amylovora is EPS production,

11



specifically the EPSs amylovoran and cellulose. Strains unable to produce one or both EPSs
showed marked decreases in autoaggregation, despite high c-di-GMP levels (Kharadi & Sundin,
2019). These cell-cell interactions are likely important for forming biofilm structures and
establishing and maintaining sessile growth.
1.5. Other mechanisms of virulence regulation
While c-di-GMP is one of the most important mechanisms of virulence gene regulation,
bacteria use a variety of other systems to alter virulence strategies and adapt to changing
environments. Two-component systems (TCS) are commonly found signal transduction
mechanisms usually consisting of a receptor kinase, which perceives a signal and phosphorylates
the response regulator, which then performs a specific function. Response regulators are often
enzymatically active or are transcriptional regulators, but may have other functions. Analysis of
the E. amylovora genome revealed forty-six predicted TCS genes, and mutagenesis of many of
these had significant effects on virulence factors such as amylovoran production and swarming
motility, and impacted virulence on immature pear fruit (Zhao, Wang, et al., 2009). In particular,
the EnvZ/OmpR TCS was discovered to positively regulate amylovoran production and
swarming motility. EnvZ is well known in other organisms as a sensor of acidic pH and changes
in osmotic pressure, and has a pleiotropic effect on the cell. Conversely, the GrrS/GrrA TCS
negatively regulated amylovoran and swarming motility in E. amylovora (Zhao, Wang, et al.,
2009).The Rcs system, a phosphorelay containing a TCS, has been demonstrated to be critical in
regulating amylovoran production. Mutations of the rcs genes resulted in a lack of virulence in
immature pear fruit (D. Wang et al., 2009). Other important TCS in E. amylovora include the
ArcA/B system, known in other organisms to regulate anaerobic metabolism, the PhoP/Q
system, sensitive to environmental magnesium and pH levels, and the HrpX/Y system, involved

12



in the T3SS (Zhao, Wang, et al., 2009).
TCS and c-di-GMP are important mechanisms by which bacteria perceive and respond to
the environment, but in recent years small noncoding RNAs (sRNA) have emerged as critical
regulators of virulence in many pathogens. These 50-100 nucleotide molecules of RNA play
diverse roles in regulating gene expression and protein activity. Most commonly, sRNAs act as
antisense RNA that regulates mRNA transcripts by complementary binding. Many bacterial
sRNAs require the chaperone protein Hfq for full stability and activity. In other instances,
sRNAs may act as riboswitches or bind to proteins to directly alter their function (Gottesman &
Storz, 2011). E. amylovora contains up to 40 Hfq-dependent sRNAs, several of which have
important functions in virulence (Zeng & Sundin, 2014). For example, the sRNA RmaA
negatively regulates amylovoran production and positively regulates swimming motility, and the
sRNA ArcZ regulates a variety of processes including biofilm formation, cell-surface
attachment, T3SS, amylovoran production, and motility (Schachterle, Zeng, et al., 2019; Zeng et
al., 2013). Additionally, ArcZ is important for surviving ROS produced by the host immune
system by regulating catalase genes in E. amylovora, highlighting the importance of this sRNA
in causing disease (Schachterle, Onsay, et al., 2019).
Carbon metabolism could be a key interface in the regulation of various virulence factors.
A recent study of E. amylovora discovered that thiamine availability and biosynthesis increased
production of the pathogenicity factor amylovoran (Yuan et al., 2021). Another interesting result
from that study is that enzymes involved with oxidative metabolism greatly decreased
amylovoran production. The EPS cellulose is one of the more recently discovered virulence
factors in E. amylovora, and the regulation of cellulose in response to environmental conditions
and potential roles in virulence could provide insight into bacterial signal transduction and plant

13



disease.
1.6. Objectives
In this study, my objectives are: 1) evaluate the specificity of DGC and PDE enzymes
that regulate cellulose production, and 2) the mode of regulation for these enzymes. I also
investigate 3) potential metabolic pathways that could facilitate cellulose production, and explore
the effects of cellulose on bacterial behavior.

14



Chapter 2. Fermentative sugar metabolism stimulates cellulose production in
Erwinia amylovora

15



2.1. Abstract
Cellulose is an exopolysaccharide produced by Erwinia amylovora that improves the
three-dimensional structure of biofilms in the xylem of apple host plants, contributing to the
devastating fire blight disease. Cellulose is composed of glucose molecules, and is energetically
expensive for bacteria to produce. However, the relationship between carbon metabolism and
cellulose production in Erwinia amylovora remains understudied. Cellulose biosynthesis is
regulated by the second messenger c-di-GMP, which is produced and degraded by diguanylate
cyclases (DGCs) and specific phosphodiesterases (PDEs), respectively. While several DGCs and
PDEs have the potential to contribute to cellulose production, the specificity of these enzymes is
poorly known. In this study, host carbohydrates were added to growth media in vitro, and
knockout mutant strains were used to examine the specificity of PDEs and a DGC with respect to
cellulose production under carbohydrate-rich conditions. Experimental results indicate that
glucose, sucrose, and sorbitol each resulted in an increase in cellulose production, which caused
bacteria to engage in autoaggregation behavior in liquid medium. The pdeC gene had the
strongest negative impact on cellulose production out of the E. amylovora PDEs, while mutation
of dgcB abolished cellulose production. A mutant strain unable to metabolize sorbitol was unable
to produce cellulose on sorbitol-containing medium, and sugar fermentation occurred in all
sugar-enriched media. Mutations of enzymes in the Tricarboxylic Acid (TCA) cycle decreased,
but did not abolish, cellulose production, indicating that fermentative metabolism may favor
cellulose production compared to other EPS such as amylovoran.

16



2.2. Introduction
The Gram-negative plant pathogen Erwinia amylovora causes the destructive fire blight
disease on rose family plants, most notably apple and pear. The bacteria are able to navigate
through a wide variety of plant parts and tissues during the disease cycle, and they employ a
variety of virulence mechanisms to do so. In the spring, bacteria emerge from trunk and stem
cankers as ooze which is then transferred to flower stigmas by insects, wind, and rain (Slack et
al., 2017). After building up large populations epiphytically, E. amylovora uses flagellar motility
to migrate to the floral cup and enter the nectarthodes. Once inside the plant, E. amylovora can
systemically infect shoot and stem tissue, periodically bursting forth with masses of
exopolysaccharides (EPS) as ooze droplets which initiate secondary infections on growing shoot
tips (Slack et al., 2022). Once inside the xylem of the host plant, E. amylovora attaches to the
xylem wall and forms biofilms, which can slow water flow and contribute to wilting symptoms
and disease development (Koczan et al., 2009).
E. amylovora produces three EPS: amylovoran, which is a pathogenicity factor important
for all infection stages; levan, which is a component of ooze droplets and biofilms while
contributing to swarming motility; and cellulose, which is important for the three-dimensional
structure of biofilms (Yuan, Eldred, & Sundin, 2022; Yuan, Eldred, Kharadi, et al., 2022).
Cellulose, synthesized by the Bcs complex, is regulated by the second messenger cyclic-dimeric
guanosine monophosphate (c-di-GMP), which allosterically activates BcsA in E. amylovora
(Koczan et al., 2009). The c-di-GMP molecule is synthesized by diguanylate cyclases (DGCs)
and degraded by phosphodiesterases (PDEs) (Kharadi et al., 2019b). In other bacteria, such as
Escherichia coli and Yersinia pestis, a specific DGC/PDE enzyme pair regulates the c-di-GMP
levels near the Bcs system, thus controlling cellulose production (Bobrov et al., 2005; Richter et

17



al., 2020). DGCs and PDEs are well known for their diversity of roles in bacterial signal
transduction, and often contain signaling domains that perceive external cues and modulate c-diGMP levels accordingly (Randall et al., 2022). Of particular importance in the change from a
motile, planktonic lifestyle to a sessile, biofilm lifestyle, c-di-GMP is a critical mechanism that
E. amylovora uses to adapt virulence strategies to the specific host environment (Kharadi et al.,
2022b).
Mutations that alter c-di-GMP levels can have drastic effects on bacterial behavior in
vitro. For example, mutagenesis of the three PDEs in E. amylovora resulted in high c-di-GMP
levels, which increased biofilm formation and caused autoaggregation behavior, when cells and
EPS clump together in culture media (Kharadi & Sundin, 2019). Conversely, strains that lacked
one of the DGCs showed greatly decreased biofilm formation (Edmunds et al., 2013). Previous
studies of DGCs in E. amylovora discovered that overexpression of several individual DGC
enzymes each increased cellulose production, and a separate transposon mutagenesis study
revealed there may be a specific effect of DgcB on the regulation of cellulose in E. amylovora
(Koczan et al., 2009; Yuan, Eldred, Kharadi, et al., 2022). The specificity of PDE enzymes on
cellulose production in E. amylovora has not yet been explored. However, a mutation of pdeC
increased intracellular c-di-GMP levels without having a distinct effect on amylovoran
production and biofilm formation compared to mutation of other active PDE enzymes in E.
amylovora (Kharadi et al., 2019b). This could potentially indicate that PdeC regulates some other
process, such as cellulose production.
EPSs such as cellulose are a key factor in autoaggregation and biofilm formation, and
offer stability and protection for E. amylovora during sessile stages of infection. Composed of
glucose polymers, production of cellulose necessitates that host carbohydrates be directed from

18



bacterial metabolism to EPS synthesis, representing a significant energy cost to the pathogen.
However, how availability of host carbohydrates influences cellulose production has not been
studied.
An important virulence factor in E. amylovora is the ability to metabolize host
carbohydrates. Analysis of the stigma tip revealed that glucose and fructose contributed to sigma
colonization while in the nectar sucrose is the primary sugar. Sorbitol, however is the most
abundant carbohydrate in apple plants, especially in shoot and leaf tissues (Pusey et al.,
2008).Mutant strains lacking the ability to metabolize sorbitol were found to have greatly
decreased virulence in apple plants, highlighting the importance of sorbitol metabolism in fire
blight disease development (Aldridge et al., 1997). While these various carbohydrates are
abundant in the apple host tissues, the common LB growth medium used for in vitro studies of E.
amylovora does not have a significant amount of carbohydrates (Kharadi et al., 2019b). In this
study, we investigate the effect of carbohydrate metabolism on cellulose production in vitro, and
observe the impact of cellulose production on bacterial behavior. In addition, we hypothesize
that DgcB is an essential positive regulator of cellulose production, and that PdeC negatively
regulates cellulose production. We also investigate mutagenesis of the signaling domain of PdeC,
and propose potential regulatory pathways for these enzymes.

19



2.3. Materials and Methods
The bacterial strains and plasmids used in this study are compiled in Table 1. Unless
otherwise noted, E. amylovora and Escherichia coli were grown in LB broth at 28° or 37° C,
respectively. Appropriate antibiotics were added to growth media at the following
concentrations: ampicillin (Ap) at 100 μg/mL, chloramphenicol (Cm) at 10 μg/mL, gentamicin
(Gm) at 15 μg/mL, and kanamycin (Km) at 30 μg/mL. The genome sequence of E. amylovora
ATCC 49946 was obtained from GenBank (accession no. FN666575) and used to determine
sequences of oligonucleotide primers used for the construction of mutants and complementation
clones (Sebaihia et al., 2010).
Table 1. Table of all bacterial strains and plasmids used in this study. Strain names in the left
column are followed by relevant characteristics and reference publications.
Strain
name
Ea1189

Characteristics

Source

Wild type

ΔbcsA

bcsA deletion mutant Ea1189

(Yuan, Eldred, Kharadi, et al.,
2022)
(Castiblanco & Sundin, 2018)

ΔpdeA

pdeA deletion mutant Ea1189

(Kharadi et al., 2019b)

ΔpdeB

pdeB deletion mutant Ea1189

(Kharadi et al., 2019b)

ΔpdeC

pdeC deletion mutant Ea1189

(Kharadi et al., 2019b)

Δ8

pdeA, pdeB, pdeC, dgcA, dgcB, dgcC, dgcD,
dgcE combination deletion mutant Ea1189
dgcB deletion mutant Ea1189

(Kharadi et al., 2022b)

ΔdgcB

20

(Yuan, Eldred, Kharadi, et al.,
2022)



Table 1. (cont’d)
Δsrl

srlAEBDMR operon mutant Ea1189

This study

ΔaceE

aceE deletion mutant Ea1189

(Yuan et al., 2021)

ΔsucA

sucA deletion mutant Ea1189

(Yuan et al., 2021)

Escherichia
coli DH5α
Plasmid
name
pBBR1MCS5:: pdeC
pLFC34::dgcB

Cloning strain

Lab Stock

Characteristics

Source

pdeC cloned into pBBR1-MCS5; Gmr

(Kharadi et al., 2019b)

dgcB overexpression vector with IPTGinducible promoter; Kmr
Arabinose-inducible lambda red
recombinase, Apr
CmR cassette flanked by FRT sites
pBAD LIC 8A plasmid, Apr

(Castiblanco & Sundin, 2018)

pKD46
pKD3
pBAD

(Datsenko & Wanner, 2000)
(Datsenko & Wanner, 2000)
AddGene commercial vector

Bioinformatics
Accessions EAM_RS16620 and EAM_RS02755 were used to find sequences for pdeC and
dgcB in the National Center for Biotechnology Information Gene database (NCBI Gene),
respectively, and homology to genes from other organisms was assessed by aligning genes using
the NCBI BLAST tool (Edmunds et al., 2013; Kharadi et al., 2019). Protein domain structure was
analyzed using SMART database, and amino acid sequence alignments were generated using the
MEGA version 7.0 program (Finn et al., 2016; Tamura et al., 2013).
Genetic manipulations
Chromosomal mutants were constructed in E. amylovora using the lambda red
recombinase method previously described (Datsenko & Wanner, 2000; Yuan, Eldred, Kharadi, et
al., 2022; Zhao, Sundin, et al., 2009). In summary, primers were designed with a target genespecific overhang attached to flanking sequences of the short flippase recognition target (FRT)
21



sites of the Cm resistance cassette from the plasmid pKD3. The Cm resistance cassette was
amplified with these primers from the pKD3 plasmid, which after gel electrophoresis were purified
using a QIAEX II Gel Extraction Kit (Qiagen, Hilden, Germany). Purified PCR products were
electroporated into E. amylovora cells containing the helper plasmid pKD46 and plated onto LB
medium supplemented with Cm to select for chromosomal insertions of the Cm cassette.
Validation that the correct mutation occurred was confirmed with PCR using primers outside the
mutagenized region and confirmed with Sanger sequencing. For complementation of target genes,
the open reading frame (ORF) of the target gene was amplified via PCR and cloned into the
complementation or overexpression vector. The correct insertion was confirmed with Sanger
sequencing and the plasmid was electroporated into the appropriate E. amylovora strain.
A truncated pdeC gene was designed by designing primers to start after the N-terminal
GAPES3 domain region of the pdeC gene and insert a start codon, creating a shortened pdeC gene
with a start codon that was cloned into an arabinose-inducible plasmid vector.
Domain-directed mutagenesis was performed on the GAPES3 domain of pdeC using the
EZ Clone GeneMorph II Domain Mutagenesis Kit (Agilent) according to manufacturer
instructions. In brief, error-prone polymerases were used to amplify the GAPES3 region of the
pdeC gene, and the product of this PCR was used as a megaprimer to resynthesize the pBBR1MCS5::pdeC plasmid, with the result of an intact plasmid and full-length pdeC gene with a
randomly mutagenized GAPES3 domain. Results were analyzed via Sanger sequencing.
Cellulose production assays
Cellulose production of E. amylovora strains was observed using saltless LB medium
amended with 40 µg/mL Congo Red dye and 20 µg/mL Coomassie Brilliant Blue dye, as
previously described (Castiblanco & Sundin, 2018; Yuan, Eldred, Kharadi, et al., 2022). E.

22



amylovora cells were grown overnight in LB medium to an optical density at 600 nm (O.D600) of
0.5 and spotted onto Congo Red medium in 5 µL aliquots. Plates were incubated at 28° for 24
hours before images were taken using a Leica M165 C microscope equipped with a DFC295
camera. Cellulose production was quantified from these images as per previous studies (Yuan,
Eldred, Kharadi, et al., 2022). To measure colony redness, representing cellulose production,
images were split into 3 RGB (red, green, and blue) channels using ImageJ and the resulting green
images were chosen to quantify the redness. Grayscale densities of the entire bacterial colonies
were quantified using the default settings of ImageJ. In order to qualitatively assess the presence
of cellulose in liquid cell culture, cell cultures were washed with sterile water and stained with a
mixture of Calcofluor white and Evans blue at concentrations of 1 g/L and 0.5 g/L, respectively.
After overnight staining, cells were washed three times with sterile water, placed in a clear 24-well
plate, and photographed under UV light to visually observe the fluorescence of Calcofluor bound
to cellulose (Castiblanco & Sundin, 2018). Washed cell cultures and aggregates were also placed
in a 0.1 M potassium phosphate buffer (pH 5) and exposed to 10 u/mL purified fungal cellulase
enzyme (Trichoderma viride cellulase, Sigma-Aldrich, Germany) and shaken gently at 37° C to
assess sensitivity of cell aggregates to cellulase, similar to previous studies (Jahn et al., 2011; Yap
et al., 2005). Aggregates with and without exposure to cellulase were photographed at 0 and 120
minutes to observe any degradation. Experiments were repeated three times, with three technical
replicates per experiment.
Determination of aggregation factor
Methods from previous studies of autoaggregation in E. amylovora were used to calculate
the aggregation factor (Kharadi & Sundin, 2019). In brief, bacterial strains were grown for 20 h at
28°. A 200 μL sample of growing medium was removed from the culture. Following this, the

23



cultures were vortexed for 30 s, and 200 μL of growing medium was removed from the culture.
The aggregation factor was calculated by determining the OD600 ratio of the medium removed
post-homogenization to that removed pre-homogenization. Experiments were repeated three
times, with three technical replicates per experiment.
Gene expression analysis
E. amylovora RNA was extracted using a Zymo RNA Clean and Concentrator kit
(Research Products International, USA) according to the manufacturer’s instructions. Genomic
DNA was removed using Turbo DNA-free DNase (Ambion, Austin, TX, USA), and cDNA was
synthesized using SuperScript III First-Strand synthesis system (Invitrogen, CA, USA). To
evaluate the mRNA levels of bcsA, dgcB, and pdeC, quantitative real-time PCR (qRT-PCR) was
conducted using cDNA as the template. Levels of cDNAs from different samples were quantified
using SYBR green PCR master mix (Applied Biosystems, Foster City, CA, USA), and the relative
expression levels of each target gene were calculated using the 2–ΔΔCT method, with the recA gene
as the internal control (Livak & Schmittgen, 2001; Takle et al., 2007). Primers used for qRT-PCR
are listed in Table 2. Experiments were repeated three times.
Table 2. Primer names, sequences, and sources of all primers used in this study.
Primer

Sequence

Source

srl
mutagenesis
F
srl
mutagenesis
R
pdeC qPCR
F
pdeC qPCR
R
dgcB qPCR
F

5’ –ATGATTGAAGCTATCA CACATGGGGCCGAATG
GTTTATCGGTCTTTTCCAGTGTAGGCTGGAGCTGCTTC –
3’
5’ –TCAGTCCTCTCCTGCAGTGATGACGCTGATAT
TCATCCCTGACAGCTGTTCATATGAATATCCTCCTTA - 3’

This study

5’- TTACGGTAATGGTGCTGCGT- 3’

This study

5’- GCTGCGAATTTTGCTCACCA - 3’

This study

5’- GCGCTTATCTTCTGGTTGCG - 3’

This study

24

This study



Table 2. (cont’d)
dgcB qPCR
5’- GATCGAAGTCGGTGCCGTAT - 3’
R
bcsA qPCR F 5’- AACCACGCCATGCAAATCAC - 3’

This study
(Yuan,
Eldred,
Kharadi,
et al.,
2022)
(Yuan,
Eldred,
Kharadi,
et al.,
2022)
This study

bcsA qPCR
R

5’- AGTAGTGCGGCGTCTGTAAC - 3’

pdeC cloning
primer with
overhang F
pdeC cloning
primer R

5’ttaactttaagaaggagatatagatTTGCGCGTCAGCCGTTCATTAAAGA
- 3’
5’This study
caagggatccttatggagttgggatTCAGTAACTGGCCAGATAGCGCT
- 3’
5’- atgccatagcatttttatcc- 3’
This study

pBAD
cloning site
primer F
pBAD
cloning site
primer R
Truncated
pdeC cloning
primer F
Truncated
pdeC cloning
primer F
Truncated
pdeC cloning
primer R
pBBr1-MCS
M13 F
pBBr1-MCS
M13 R
GAPES3
domain
region F

5’GATTTAATCTGTATCAGG
- 3’
5’ttaactttaagaaggagatatagatTTGGCTGTATCGACGATGGTCACC
- 3’
5’ttaactttaagaaggagatatagatTTGGCTGTATCGACGATGGTCACC
- 3’
5’caagggatccttatggagttgggatTCAGTAACTGGCCAGATAGCGCT
- 3’
5’CCCAGTCACGACGTTGTAAAACG
- 3’
5’AGCGGATAACAATTTCACACAGG
- 3’
5’TTTATCGTTATCCAGCTTTTTCA
- 3’

25

This study

This study

This study

This study

This study

This study

This study



Table 2. (cont’d)
GAPES3
domain
region R
pBAD
cloning site
primer F
pBAD
cloning site
primer R
GAPES3
domain
region F
GAPES3
domain
region R

5’GATAGGCGGTGACCATCG
- 3’
5’- atgccatagcatttttatcc- 3’

This study

5’GATTTAATCTGTATCAGG
- 3’
5’TTTATCGTTATCCAGCTTTTTCA
- 3’
5’GATAGGCGGTGACCATCG
- 3’

This study

This study

This study

This study

Analysis of sorbitol fermentation ability using MacConkey medium
MacConkey indicator plates (10 g/L peptone, 2.5 g/L NaCl, 30 mg/L Phenol Red) with 1%
sorbitol were used to qualitatively analyze the ability of the mutant strains to ferment sorbitol.
Phenol Red in the medium serves as a pH indicator; sorbitol fermentation is signaled by color
change from red (> ~7.5) to yellow (< ~6.8) with intermediate shades of orange.
C-di-GMP quantification
Intracellular levels of c-di-GMP were quantified using ultraperformance liquid
chromatography-tandem mass spectrometry (UPLC-MS/MS), as described previously (Massie et
al., 2012). Overnight cultures were diluted 1:30 and grown until they reached an OD600 of 0.8 to
1.0 and then normalized to an OD600 of 0.7, 5 ml of this new culture was pelleted, and the
supernatant was discarded. The pelleted cells were treated with 100 µL of an extraction buffer
solution (40% acetonitrile and 40% methanol) and incubated overnight at 20° C, after which 10
µL of supernatant was analyzed by UPLC-MS/MS on a Waters TQ-S instrument. All samples were
quantified by comparison against a standard curve generated by using synthetic c-di-GMP (Axxora
26



Life Sciences Inc., San Diego, CA). This assay was repeated three times.

27



2.4. Results
Addition of host sugars stimulates cellulose production and autoaggregation in E. amylovora.
Glucose, sucrose, and sorbitol are the primary free carbohydrates found in the apple and
pear host plants of E. amylovora, and are often found at concentrations up to 80 mg/mL (Wang
& Stutte, 1992). When 20 mg/mL of these three host carbohydrates were added to LB agar
medium, we observed that all three host sugars resulted in significantly brighter red
microcolonies when assayed for cellulose production, indicating that sugars stimulated cellulose
production in WT E. amylovora. . In contrast, the ΔbcsA mutant strain appeared unaffected by
the addition of sugars to the growth medium and displayed a lack of cellulose production
(Fig.1A). When WT cells were grown in liquid LB medium amended with 20 mg/mL sugars, the
cells remained clumped in aggregates on the bottom of the culture tube or clinging to the side of
the culture tube, whereas in unamended LB medium E. amylovora grew planktonically, creating
a turbid culture. To quantify this, the autoaggregation factor was measured, and amendment of
LB medium with each of the three sugars resulted in a significant increase in the autoaggregation
factor of the WT strain. Similarly to the cellulose assay, the ΔbcsA mutant strain appeared
unaffected by the addition of sugars to the growth medium, showing planktonic growth in the
presence or absence of sugars (Fig.1B). This suggests that cellulose is a key component of the
sugar-stimulated autoaggregation behavior in E. amylovora. In addition, the washed cell
aggregates fluoresced bluish white when stained with Calcofluor white, and the aggregates were
sensitive to enzymatic digestion with cellulase, further supporting this point (Fig.1C-D). The
aggregates from all three sugar treatments completely degraded within 120 minutes after
exposure to cellulase, suggesting that cellulose was the primary structural component.
All three sugars tested produced very similar phenotypic results in our experiments. Since

28



sorbitol has been previously demonstrated to be the most abundant transport carbohydrate in
apple tissue, and the ability of E. amylovora to metabolize sorbitol is an important virulence
factor contributing to fire blight disease (Aldridge et al., 1997; Z. Wang & Stutte, 1992), we
chose to focus on the effects of sorbitol on signal transduction in E. amylovora for the remainder
of this study.

29



C

Cellulose production on sugar-amended LB medium
T

Relative cellulose production
(fold -change LB/LB sugars)

A

T
glucose

Glucose

Sucrose

T

B

T

Sorbitol

T

T No sugar

T

Autoaggregation in liquid culture at 20 hours
Autoaggregation factor
(arbitrary units)

T
sorbitol

T
sucrose

D

0 min
Buffer

cellulase

120 min
Buffer

cellulase

T
glucose

N.S

T
sucrose
No sugars

Glucose

Sucrose

Sorbitol

T
sorbitol

Figure 1. Addition of host sugars stimulates cellulose production and autoaggregation in E.
amylovora. A) Cellulose production by Ea1189 WT and bcsA mutant strains on Congo Red LB
agar medium amended with 20 mg/mL sugars. Quantification of macrocolony redness was
measured with WT on unamended medium as a reference and denoted by the dotted line. B)
Quantification of autoaggregation factor in liquid LB medium amended with sugars after 20
hours of growth. Error bars signify standard deviation; presence of an asterisk above the bars
indicates statistically significant difference observed (P > 0.05 by Student’s t test). C) Calcoflour
white binding on DI-washed cellulose aggregates produced by Ea1189 WT, visualized under UV
light. Background fluorescence is due to Evans Blue dye added to the reagent for contrast. D)
Enzymatic digestion of DI-washed cellulose aggregates by cellulase over time.

30



PdeC and DgcB are the main cyclic-di-GMP enzymes involved in cellulose production in E.
amylovora.
E. amylovora contains 8 active cyclic-di-GMP metabolizing enzymes, and we compared
these proteins to known regulators of cellulose production in other organisms to predict potential
roles of c-di-GMP enzymes in regulating cellulose production. In the widespread human
pathogen Escherichia coli, the bcs operon is associated with a specific diguanylate cyclase and a
specific phosphodiesterase that regulate the local pool of cyclic-di-GMP available to activate
cellulose synthase (Richter et al., 2020). To predict if E. amylovora contains any homologs to
these proteins, a protein BLAST search was performed using the amino acid sequences of Es.
coli proteins DgcC and PdeK (accessions NP_414919.1 and NP_417986.4, respectively) against
the genome of E. amylovora. These cellulose-associated proteins showed significant homology
to the E. amylovora proteins DgcB and PdeC. Es. coli DgcC (371 amino acids) showed 36.57%
sequence similarity to E. amylovora DgcB (365 amino acids). Es. coli PdeK (649 amino acids)
and E. amylovora PdeC (664 amino acids) showed 50.45% amino acid sequence similarity with
99% query cover (Appendix 1). In addition, E. amylovora DgcB and PdeC also have similar
domain structures to corresponding Es. coli proteins DgcC and PdeK, with 5 predicted
transmembrane helices followed by the GGDEF domain for the DGCs, and a periplasmic domain
flanked by two transmebrane helices followed by GGDEF and EAL domains for the PDEs
(Appendix 1). These E. amylovora proteins are also similar to the HmsT and HmsP proteins in
the animal pathogen Yersinia pestis, both of which are associated with exopolysaccharide
production (Bobrov et al., 2005). The location of the pdeC gene in the E. amylovora
chromosome was also located immediately next to the bcs operon, suggesting a role in cellulose
production (Appendix 1). Based on this bioinformatic prediction, we hypothesized that DgcB

31



and PdeC play an important role in cellulose production in E. amylovora.
Knockout mutants of each of the three PDE enzymes in E. amylovora were assayed for
cellulose production. The ΔpdeA and ΔpdeB deletion mutant strains did not show a visibly
different phenotype from either the

T strain or ΔbcsA, with whitish macrocolonies observed on

Congo Red plates. In contrast, ΔpdeC formed a red macrocolony, suggesting that PdeC
negatively regulates cellulose production (Fig2.A). Complementation of the pdeC gene in the
ΔpdeC mutant background abolished the cellulose production phenotype, suggesting that PdeC is
the most active in regulating cellulose production out of the three E. amylovora PDEs. Previous
studies of cellulose production in E. amylovora have suggested that, out of the five active
diguanylate cyclases, DgcB could be most important for the upregulation of cellulose
(Castiblanco & Sundin, 2018; Yuan, Eldred, Kharadi, et al., 2022) further confirm this, the dgcB
gene in a plasmid vector was introduced into a strain containing knockout mutations of all five
diguanylate cyclases and all three phosphodiesterases. This Δ8 strain containing only the dgcB
gene was then assayed for cellulose production. Interestingly, the Δ8 pdgcB strain displayed a
red macrocolony, providing evidence that out of all E. amylovora c-di-GMP enzymes, only
DgcB is necessary to stimulate cellulose production (Fig.2A).
Similar results were observed when autoaggregation was measured, with ΔbcsA, ΔpdeA,
and ΔpdeB growing planktonically and not forming an aggregate at the bottom of the culture
tube. Although a small aggregate was visible in the ΔpdeC mutant, this strain had sufficient
planktonic growth so as to render the difference in autoaggregation factor statistically
insignificant (Fig.2B). The Δ8 pdgcB strain had an autoaggregation factor significantly higher
than the wild type, suggesting that this DGC alone can result in the transition from planktonic to
sessile growth. This strain, as well as the ΔpdeC mutant to a much lesser extent, formed a large

32



visible aggregate (Fig.2C). Washed aggregates fluoresced bluish white after staining with
Calcofluor white and were sensitive to cellulase digestion, confirming that the aggregates formed
by the mutant strains contained mostly cellulose, similar to the sugar amended medium
(Fig.2D,E). No aggregate was formed or able to be collected for the Calcofluor and cellulase
assays in the ΔbcsA, ΔpdeA, ΔpdeB, and ΔpdeC ppdeC strains.

33



Relative cellulose production
(fold -change T/mutant)

A

B

C

Cellulose production in knockout strains

8p

0 min

D

Buffer

cellulase

120 min
Buffer

cellulase

8

Autoaggregation in Pde knockout strains at 20 hours
Autoaggregation factor
(arbitrary units)

T

p

N.S

E
8p

Figure 2. PdeC and DgcB are the main cyclic-di-GMP enzymes involved in cellulose production
in E. amylovora. A) Cellulose production by Ea1189 mutant strains on Congo Red LB agar
medium amended with 20 mg/mL sugars. Quantification of macrocolony redness was measured
with WT on unamended medium as a reference and denoted by the dotted line. B) Quantification
of autoaggregation factor in liquid LB medium amended with sugars after 20 hours of growth.
Error bars signify standard deviation; presence of an asterisk above the bars indicates statistically
significant difference observed (P > 0.05 by Student’s t test). C) Image showing visible
aggregates settling to the bottom of culture tubes. Red boxes highlight these aggregates. D)
Enzymatic digestion of DI-washed cellulose aggregates by cellulase over time. E) Calcoflour
white binding on DI-washed cellulose aggregates produced by Ea1189 WT, visualized under UV
light. Background fluorescence is due to Evans Blue dye added to the reagent for contrast.

34



The dgcB gene is necessary for E. amylovora to produce cellulose in response to sugars.
Since the addition of host sugars resulted in an increase of cellulose production and
cellulose-mediated autoaggregation (Fig.1), and the pdeC and dgcB genes appeared to
significantly influence cellulose production and autoaggregation (Fig.2), we hypothesized that
these genes were responsible for the phenotype observed upon sugar addition to the growth
medium. Mutant strains were assayed for cellulose production and autoaggregation and
compared between growth conditions with and without sorbitol. As observed earlier, the WT
strain showed significantly higher levels of cellulose production on sorbitol-amended medium,
whereas the ΔbcsA strain did not appear to produce cellulose in either condition. The ΔpdeC
strain displayed higher levels of cellulose production in both conditions, both comparable to the
WT on sorbitol-amended medium, but there was no significant difference in cellulose production
in this strain between conditions (Fig.3A). When this gene was complemented on a plasmid
vector, however, there appeared to be a lack of cellulose production even in the presence of
sorbitol, and there was no significant difference in cellulose production between conditions for
the ΔpdeC ppdeC strain. These results indicate that the pdeC gene is a powerful negative
regulator of cellulose production, but may not be specifically involved in the activation of this
process in response to host carbohydrates. A single knockout mutation of the dgcB gene resulted
in a lack of cellulose production, even in the presence of sorbitol, indicating that this enzyme is a
critical positive regulator of cellulose production. As before, the Δ8 pdgcB strain, expression of
this gene alone resulted in high cellulose production, similar to the WT on sorbitol-amended
medium and the ΔpdeC strain, but there was no significant difference in cellulose production
between growth conditions (Fig.3A).
Autoaggregation assays of the mutant strains with and without sorbitol showed a similar

35



overall trend, with the WT displaying significantly higher autoaggregation with sorbitol and the
ΔbcsA strain growing planktonically under both conditions. The ΔpdeC strain mirrored the
phenotype of the WT strain, showing a significantly higher autoaggregation factor when sorbitol
was added to the growth medium (Fig.3B). Complementation of this gene abolished the
phenotype, with cultures growing planktonically in both conditions. Interestingly, the ΔdgcB
mutation also abolished autoaggregation (Fig.3B). Together with the cellulose assays, these
results suggest that, while PdeC is a key regulator of cellulose biosynthesis in E. amylovora,
DgcB may be more important in stimulating cellulose production in response to host
carbohydrates such as sorbitol.

36



Cellulose production on sugar-amended LB medium
Relative cellulose production
(fold -change LB/ LB sugars)

A

LB sorbitol

N.S
N.S

B

N.S

N.S

Autoaggregation in liquid culture at 20 hours
Autoaggregation factor
(arbitrary units)

LB

LB
LB sorbitol

N.S

N.S

N.S

Figure 3. The dgcB gene is necessary for E. amylovora to produce cellulose in response to
sugars. A) Cellulose production by Ea1189 mutant strains on Congo Red LB agar medium
amended with 20 mg/mL sugars. Quantification of macrocolony redness was measured with WT
on unamended medium as a reference and denoted by the dotted line. B) Quantification of
autoaggregation factor in liquid LB medium amended with sugars after 20 hours of growth. Error
bars signify standard deviation; presence of an asterisk above the bars indicates statistically
significant difference observed (P > 0.05 by Student’s t test). Non-significant differences are
denoted by N.S.

37



Addition of sorbitol to LB medium increases expression of dgcB and intracellular c-di-GMP
levels.
Previous bioinformatic analysis has predicted that DgcB has a relatively simple protein
domain architecture, with a C-terminal GGDEF domain and a series of N-terminal
transmembrane domains and lacking a predicted sensory domain (Edmunds et al., 2013). This
information led us to hypothesize that a key factor in cellulose regulation and associated
phenotypes in response to host sugars was the expression of the dgcB gene, rather than a proteinlevel sensory mechanism of the DgcB protein product. We tested the effect of dgcB expression
induction on autoaggregation using an IPTG-inducible promoter in the Δ8 pdgcB strain, so that
no other active DGCs or PDEs would be present.

hen cultures of the Δ8 pdgcB strain were

grown overnight with no IPTG, the unamended medium showed a higher autoaggregation factor
than the sorbitol amended medium. These cultures were then vortexed and IPTG was added to
induce dgcB expression. Interestingly, within two hours of gene induction the sorbitol-amended
culture had formed a dense aggregate, showing a very high autoaggregation factor, whereas the
culture without sorbitol had only partially restored autoaggregation behavior upon dgcB
expression (Fig.4A). These results indicate that expression of dgcB is a primary regulator in the
transition from planktonic to sessile growth and biosynthesis of cellulose.
Expression of three target genes, bcsA, dgcB, and pdeC was measured in the WT strain
using qPCR to quantify the effects of sorbitol addition on gene expression. The bcsA and dgcB
genes showed approximately 2.5 and 3-fold increases in expression, respectively (Fig.4B). These
results appear to be consistent with the other experiments in this study, as sorbitol increased
cellulose production, which requires c-di-GMP as an activator. Unexpectedly, however, the pdeC
gene was also upregulated about 2-fold (Fig.4B). We found this result surprising because PdeC is

38



a negative regulator of cellulose production. However, our bioinformatic analysis indicates that
the pdeC gene is immediately adjacent to the bcs operon in E. amylovora, and this could
potentially explain the co-expression of bcsA and pdeC (Appendix 1).
Our results suggest that sorbitol increases expression of c-di-GMP metabolizing genes,
and we measured intracellular levels of c-di-GMP in cultures grown with or without sorbitol to
assess the effect of sorbitol on overall c-di-GMP levels. The WT strain grown in sorbitolamended medium showed significantly higher c-di-GMP levels, providing further evidence that
the presence of sugars induces c-di-GMP mediated cellulose production and autoaggregation
(Fig.4C). The dgcB mutant strain, however, showed no change in c-di-GMP levels, revealing that
this enzyme is responsible for the sorbitol-induced increase in c-di-GMP levels.

39



A

Autoaggregation induced by

expression

Autoaggregation factor
(arbitrary units)

LB
LB sorbitol

C

Relative gene expression in sorbitol amended
medium
Relative expression levlel
(Log2 fold change treated/untreated

B

LB

120 min
after IPTG

Concentration (nM)

Before
IPTG

Intracellular c-di-GM P

LB sorbitol
N.S

Figure 4. Addition of sorbitol to LB medium increases expression of dgcB and intracellular c-diGMP levels. A) Autoaggregation of Ea1189 Δ8 pdgcB mutant strain in liquid LB medium
amended with 20 mg/mL sorbitol. The dgcB gene was cloned on a plasmid with an IPTGinducible promoter; autoaggregation measurements were taken before 0.1 mM IPTG was added
and two hours after. B) mRNA levels of Ea1189 WT measured by qRT-PCR in liquid LB
medium with or without sorbitol after 20 hours of growth. Values are shown relative to
endogenous control recA. C) Intracellular c-di-GMP concentrations measured with HPLCMS/MS. Error bars signify standard deviation; presence of an asterisk above the bars indicates
statistically significant difference observed (P > 0.05 by Student’s t test). Non-significant
differences are denoted by N.S.

40



Fermentation is associated with cellulose production in E. amylovora.
We observed increases in dgcB and bcsA expression, c-di-GMP levels, and cellulosemediated autoaggregation in response to the addition of sorbitol to the growth medium. In light
of these results, we questioned whether these phenotypes were the result of sorbitol itself acting
as a molecular signal, or via more indirect means through general metabolic changes. To test
this, we used a mutant strain lacking the srl operon, a six-gene operon necessary for the
conversion of sorbitol into glycolytic intermediates, and assayed for sorbitol metabolism and
cellulose production. MacConkey indicator plates revealed a pH decrease in response to WT
bacterial growth, with the medium changing from magenta to yellow immediately around the
bacterial colony after overnight growth. As expected, the srl mutant strain was unable to ferment
the sorbitol in the medium, and no color change was observed, while the fermentation ability was
restored by complementing the srl mutant with the whole srl operon on a plasmid vector
(Fig.5A). The same strains showed the same pattern with respect to cellulose production on
sorbitol-amended medium, with the srl mutant lacking cellulose production even in the presence
of sorbitol, and complementation of this operon restoring cellulose production (Fig.5B). These
results suggest that cellulose production requires fermentative sorbitol metabolism under our
experimental conditions. Liquid growth medium pH was measured at 18 and 40 hours, and the
pH of sorbitol-amended medium decreased by about 1 pH unit, while the pH of unamended
medium increased by a similar amount, providing further evidence that sorbitol is being
fermented by E. amylovora under our experimental conditions (Fig.5C). In addition, while c-diGMP levels significantly increased with sorbitol addition in the WT strain as observed before,
there was not an increase in c-di-GMP in the srl mutant strain. These results suggest that the
uptake and metabolism of sorbitol are essential for the increase in c-di-GMP that activates

41



cellulose biosynthesis. Together, these experiments indicate that cellulose production and
autoaggregation in response to sorbitol are likely because of sorbitol metabolism, and not some
other factor, such as changes in osmotic pressure.
The link between sorbitol fermentation and autoaggregation prompted us to assess the
effect of two mutations affecting respiration via the TCA cycle, used in a recent study of E.
amylovora metabolism (Yuan et al., 2021). These mutations were used to inhibit respiration and
completion of the TCA cycle. Interestingly, the ΔaceE strain, lacking a pyruvate dehydrogenase
gene, did not appear to greatly affect cellulose production or autoaggregation (Fig.6A). The
ΔsucA strain, lacking an α-ketoglutarate dehydrogenase, showed a slight negative effect on
cellulose production, with no significant difference between sorbitol and plain LB media
(Fig.6A). In contrast, ΔsucA did not change the phenotype with respect to autoaggregation
(Fig.6B). Since these mutant strains have an impaired TCA cycle and respiratory pathway, yet
some level of cellulose production and autoaggregation is still observed, this could indicate that
the TCA cycle is not essential for cellulose production.

42



p

C

9

p units

A

of bacterial culture medium over time

10
8
7
6
5

T

rl E

4
-1

rl E
p rl E

B

14

LB

19

24

29

34

39

44

rl E
p rl E

LB sorbitol

Intracellular c-di-GM P
Concentration (nM)

rl E

9

Time (h)

D
T

4

LB
LB sorbitol
N.S

rl

Figure 5. Fermentation is associated with cellulose production in E. amylovora. A) Fermentation
of sorbitol on sorbitol-MacConkey agar. The yellowish color of colonies indicates an acidic pH
from fermentation. B) Cellulose production of Ea1189 WT and srl operon mutants on Congo
Red amended LB agar with sorbitol. Reddish or pinkish color indicates cellulose production. C)
Change in pH of liquid LB growth medium with and without sorbitol over time. D) Intracellular
c-di-GMP concentrations measured with HPLC-MS/MS. Error bars signify standard deviation;
presence of an asterisk above the bars indicates statistically significant difference observed (P >
0.05 by Student’s t test). Non-significant differences are denoted by N.S.

43



Figure 6. TCA cycle enzymes may not be essential for cellulose production. A) Cellulose
production by Ea1189 mutant strains on Congo Red LB agar medium amended with 20 mg/mL
sugars. Quantification of macrocolony redness was measured with WT on unamended medium
as a reference and denoted by the dotted line. B) Quantification of autoaggregation factor in
liquid LB medium amended with sugars after 20 hours of growth. Error bars signify standard
deviation; presence of an asterisk above the bars indicates statistically significant difference
observed (P > 0.05 by Student’s t test). Non-significant differences are denoted by N.S.

44



2.5. Discussion
The apple and pear host plants of E. amylovora are rich in carbohydrates, mainly sorbitol,
sucrose, and small amounts of fructose and glucose. In vitro studies of this economically
important bacterium often use LB growth medium, which is rich in amino acids and short
peptides but does not contain a significant source of carbohydrates. When high levels of host
sugars glucose, sucrose, and sorbitol were added to LB medium, we observed visible differences
in bacterial growth: in 20 mg/mL sugar treatments, dilution plating revealed CFU/mL were
roughly 6-fold lower than unamended LB medium at the sampling time (1.7 x 107 vs. 1.1 x 108
CFU/mL). In addition, the cells were mostly confined to the bottom of the culture tube in visible
aggregates, displaying autoaggregation behavior as previously observed in E. amylovora
(Kharadi & Sundin, 2019). Cellulose production assays revealed that WT Ea1189 produced
significantly more cellulose when grown on medium amended with sugars than plain LB, and
this phenotype was reflected in liquid media by significantly increased autoaggregation (Fig.1AB). Interestingly, each sugar produced a very similar phenotype, suggesting that under our
experimental conditions there was no specific effect of a certain sugar. The addition of host
sugars increased the nutrient density of the growth medium, yet there were far fewer cells of WT
Ea1189 in the sugar amended medium than unamended LB at the sampling timepoint. These
results seem to suggest some kind of metabolic shift in E. amylovora upon sugar treatment, in
which energy usage favors biosynthesis of EPS over cell division and proliferation, creating a
mass of EPS and relatively few cells. Kharadi & Sundin (2019) discovered similar
autoaggregation in a triple PDE mutant which had constitutively high levels of c-di-GMP. The
authors noted that cell division, and especially cell separation post division, was impaired in this
autoaggregating strain. These findings, in concert with those of this study, could suggest that c-

45



di-GMP may control carbon metabolism in E. amylovora, and vice versa, based on the
environmental conditions.
The aggregates produced by WT Ea1189 stained positive for cellulose using Calcoflour
white, and were totally degraded by cellulase enzymes on a timescale of hours (Fig.1C-D). These
cellulose aggregates displayed interesting physical properties. During the washing stage of these
assays, the whitish aggregates appeared very fluffy and fragile, almost cloudlike, and would
occasionally break apart into smaller pieces. These smaller pieces were far sturdier and more
resistant to shearing even with gentle pipetting and vortexing (data not shown). At no time,
however, was any reduction in the aggregate apparent, as cellulose is highly insoluble. These
physiochemical properties of cellulose differ greatly from the two other EPSs of E. amylovora,
amylovoran and levan, which are water soluble, in contrast to cellulose which is water insoluble
by definition (Bennett & Billing, 1980; Slack et al., 2017). This basic chemical difference begs
the question of whether cellulose plays a different function in the fire blight disease cycle
compared with amylovoran and levan. Amylovoran and levan make up a large portion of
bacterial ooze, and amylovoran is thought to be constitutively produced throughout the disease
cycle, as mutations in the ams gene cluster render the bacteria nonpathogenic (Bellemann &
Geider, 1992; Slack et al., 2017). These two EPSs also play a role in sliding motility, an
important mechanism for bacterial movement in confined spaces, and contribute to biofilm
formation (Koczan et al., 2009; Yuan, Eldred, & Sundin, 2022). Cellulose appears to be of more
limited use in virulence, to date only observed in xylem tissue with electron microscopy,
although tests for the presence of bacterial cellulose in other plant tissues have not been done and
would be difficult to perform. Nonetheless, cellulose may be a key EPS for E. amylovora to
effectively colonize the xylem. Because of its insolubility and ability to resist shear stresses,

46



cellulose could provide a better matrix to protect bacterial cells from the flow of xylem fluid,
where water-soluble amylovoran and levan might be washed away. Furthermore, the ability of
cellulose to facilitate cell-cell interaction and autoaggregation may be an important mechanism
to keep groups of cells together, where a group would have a higher likelihood of successfully
colonizing new areas of the host than a single cell.
Cellulose is an important EPS produced by E. amylovora and is under strict regulation by
c-di-GMP, a regulator of many virulence factors in this bacterium (Castiblanco & Sundin, 2018;
Edmunds et al., 2013). The experiments in this study present evidence that specific enzymes
DgcB and PdeC control c-di-GMP levels that influence activity of the bcs system, modulating
the phenotype of E. amylovora. Previous research on this pathogen reveals that although DgcB
impacts other virulence factors such as amylovoran production and swimming motility, it exerts
a particularly powerful effect on cellulose production. Mutation of this gene alone abolished
cellulose production in a transposon mutagenesis study (Yuan, Eldred, Kharadi, et al., 2022), and
the results shown here reveal a striking lack of cellulose production even in culture conditions
where cellulose production is favored (Fig.3). Furthermore, the expression of the dgcB gene has
a strong positive effect on cellulose production and caused a distinct increase in autoaggregation
in 120 minutes. In sorbitol treated growth medium, intracellular c-di-GMP levels increased in the
WT strain, but this increase was not observed in the dgcB mutant. This suggests that the effect of
sugar amendment on high c-di-GMP phenotypes such as cellulose production and
autoaggregation is mediated at least primarily, and potentially exclusively, by DgcB-synthesized
c-di-GMP. Bioinformatic analysis of the predicted DgcB protein structure shows, in contrast to
many DGCs, that DgcB does not contain a predicted signaling domain; instead, the entire Nterminus is a series of five transmembrane domains followed only by the C-terminal GGDEF

47



domain (Edmunds et al., 2013). This comparatively simple structure could suggest that DgcB
could be regulated more strictly at the transcriptional level, as opposed to the many posttranslational mechanisms discovered in other DGCs (Randall et al., 2022). Indeed, mRNA
quantification of WT Ea1189 in growth media with and without sorbitol showed an
approximately 3-fold increase in bcsA and dgcB expression in the sorbitol treatment, where
cellulose was produced (Fig.4B). In addition, expression of dgcB in both growth conditions
produced similar levels of cellulose on Congo Red plates (Fig.3A), suggesting that the sorbitol
treatment affects transcription of the dgcB gene but may not influence protein-level regulation of
enzyme activity. However, in liquid medium the Δ8 pdgcB strain had a markedly higher
autoaggregation factor when sorbitol was added compared to the control condition (Fig.4A). This
could be from some unknown effect on protein-level enzyme activity, or it could simply be an
effect of abundant carbon substrates facilitating more growth and the biosynthesis of larger
cellulose aggregates. While these experiments demonstrate that expression of dgcB is likely the
most important positive regulator of cellulose production and autoaggregation, it is still unknown
what genetic mechanisms facilitate the perception of sugars in the environment and increase
transcription of this important gene.
Regardless of enzymatic regulation, part of the pronounced impact of E. amylovora DgcB
on cellulose could be from membrane localization. In a detailed study of pEtN-cellulose
production in Es. coli, researchers discovered that DgcC directly interacts with the BcsB subunit
of the Bcs complex via the transmembrane domain; a truncated DgcC protein lacking the
transmembrane domain failed to produce cellulose (Richter et al., 2020). Since our bioinformatic
analysis of Es. coli DgcC predicts an identical domain structure and some degree of amino acid
sequence similarity to E. amylovora DgcB, perhaps a similar mechanism could be at play, where

48



the membrane localization of DgcB as a part of the Bcs complex gives this DGC such a potent
effect on cellulose production. However, it is important to note that the Es. coli DgcC activated
cellulose production specifically and locally, with no impact on the intracellular levels of c-diGMP (Richter et al., 2020). This is in contrast to DgcB of E. amylovora, mutation of which can
change the overall levels of intracellular c-di-GMP, and can have an effect on multiple virulence
factors as observed in previous studies (Edmunds et al., 2013; Kharadi et al., 2022b). How single
enzymes influence both local and global c-di-GMP signal transduction remains a mysterious
research area of molecular biology.
Analysis of the three PDE enzymes of E. amylovora revealed PdeC to be the only one
affecting cellulose production and autoaggregation under our experimental conditions (Fig.2A).
This phenotype is interesting given that comparison of the pdeC mutant to pdeA and pdeB
mutants found no change in virulence factors such as biofilm formation, swimming motility, and
amylovoran production, even though this gene had the strongest individual impact on c-di-GMP
levels (Kharadi et al., 2019b). This suggests that the strong PDE activity of PdeC is directed
mostly towards downregulating cellulose. The effect of PdeC was noticeable but less distinct in
LB liquid media; the ΔpdeC strain produced a visible cellulose-based aggregate even with no
sorbitol (Fig2C-E), and the autoaggregation factor was slightly higher, but not statistically
significant compared to the wild type (Fig.2B). Mutation of pdeC also had no effect on the
sorbitol treatments compared to the WT strain (Fig.3A,B). These results in addition to the effects
of DgcB may indicate that the stimulation of cellulose production when sorbitol is added is a
result of dgcB expression and activity, rather than some inactivation or downregulation of PdeC.
Bioinformatic comparison of E. amylovora PdeC to Es. coli PdeK showed very similar
domain structures, and high amino acid sequence homology (Appendix 1). Es. coli PdeK is a

49



known specific regulator of cellulose production and was demonstrated to associate with both Es.
coli DgcC and the BcsB subunit of the Bcs complex (Richter et al., 2020). PdeC and PdeK also
share similar placement in the genome, immediately next to the bcs operon (Appendix 1). This
comparison, when combined with the experimental data shown here, leads us to propose that
PdeC is a specific regulator of cellulose production in E. amylovora.
Previous bioinformatic analysis predicts a periplasmic signaling domain between two
transmembrane domains at the N-terminus of PdeC and similar proteins such as Es. coli PdeK
(Kharadi et al., 2019b). Annotated as Gamma-proteobacterial periplasmic signaling domain 3
(GAPES3), the function of this domain and any possible roles in PDE enzyme activity are
completely unknown. Truncation of the pdeC gene to eliminate the N-terminal GAPES3 domain
appeared to abolish PDE activity when assayed for cellulose production (Appendix 2). However,
truncated genes can be prone to misfolding, and it should be noted that a transmembrane domain
was also eliminated in the truncated version. To more clearly ascertain if the GAPES3 domain is
truly necessary for enzymatic activity of PdeC, domain-directed mutagenesis was used to
randomly mutate base pairs in the GAPES3 region. A strain was isolated that had a completely
intact full-length pdeC gene except for 6 point mutations in the target region. This strain also
showed a loss of PDE activity when assayed for cellulose production (Appendix 2). This
presents strong evidence that the N-terminal GAPES3 domain controls the activity of the active
C-terminal EAL domain. We speculate that a post-translational protein-level interaction
regulates the activity of PdeC via the GAPES3 domain. While the specific mechanism of signal
perception by this domain is still unknown, this information does help explain the surprising
finding that the pdeC gene was upregulated in response to sorbitol addition (Fig.4B). Since
cellulose production is favored under these conditions, and PdeC is a negative regulator of

50



cellulose production, we expected this gene to be downregulated. However, perhaps inactivation
of the enzyme could occur via the GAPES3 domain while allowing the pdeC gene to be
expressed, allowing E. amylovora to have the negative regulator ready for speedy shutoff of
cellulose production via protein-level activation of PdeC.
While the precise mechanisms of how addition of host sugars influences c-di-GMP and
cellulose via DgcB and PdeC was beyond the scope of this study, the experiments presented here
suggest that changes in central carbon metabolism may broadly regulate these processes. The srl
mutant strain, unable to take up or metabolize sorbitol via glycolysis, failed to ferment sorbitol
when this sugar was added to the growth medium (Fig. 5A). This strain showed a concomitant
lack of cellulose production and a failure to increase c-di-GMP levels when grown with sorbitol,
as in the WT (Fig.5B,D). We found this association between cellulose production and
fermentative metabolism interesting given the 6-fold decrease in growth we observed upon
addition of sorbitol to LB medium. Glycolysis yields 2 ATP molecules per glucose molecule in
addition to a fermentation product such as lactate or acetate, whereas full oxidative metabolism
with the Tricarboxylic Acid (TCA) cycle yields up to 31 ATP molecules per glucose molecule
(Mookerjee et al., 2017). This supports our data showing slower cell growth associated with less
energy-productive glycolysis and fermentation.
Recent research on E. amylovora revealed that mutations in aceE, a pyruvate
dehydrogenase, and sucA, an α-ketoglutarate dehydrogenase, impaired the TCA cycle and
amylovoran production, with resulting amylovoran levels near the ams mutant control (Yuan et
al., 2021). We tested these strains for cellulose production and autoaggregation, hypothesizing
that if glycolysis and fermentation favor cellulose production, the TCA cycle is less important
for this process. The results of these experiments show that, overall, mutation of aceE and sucA

51



did not abolish these phenotypes, although the Δ u

strain did show reduced cellulose

production on solid medium (Fig.6A,B). While this does potentially support our hypothesis that
glycolysis favors cellulose production compared to complete oxidation via the TCA cycle, we do
not place too much emphasis on these experiments, as these key metabolic genes may have wideranging effects on many cellular processes, and may even be involved in certain fermentation
pathways as well (Clark, 1989). A more detailed chemical analysis would likely be necessary to
draw conclusions more accurately from this experiment.
Previous chemical analysis of E. amylovora have shown that fermentation and oxidation
pathways happen simultaneously under ambient oxygen levels, with bacterial cultures producing
significant amounts of lactic acid, ethanol, and other organic acids from glycolysis of glucose in
addition to CO2 gas from oxidative metabolism via the TCA cycle (Sutton & Starr, 1959). This
study noted that fermentation end products were favored over CO2 in microaerobic and
anaerobic conditions. The effect of producing less energy-efficient fermentative byproducts
when nutrients are abundant, even when oxygen is available to facilitate the TCA cycle, is very
common in microorganisms and has been called ‘overflow metabolism’ or the Crabtree effect.
Overflow metabolism has been studied for decades in the model organism E. coli, which
ferments glucose to acetic acid when glucose is abundant, even in aerobic conditions (Millard et
al., 2021). Thus it is not surprising that E. amylovora also ferments carbon substrates in our
aerobic in vitro experimental conditions. Perhaps having multiple carbon metabolic pathways
operating simultaneously allows E. amylovora more versatility in deploying different virulence
strategies, such as rapid cell proliferation at the stigma tip, or biofilm formation where anabolic
biosynthesis pathways are important to produce EPSs, such as cellulose, in addition to cell
growth (Kharadi & Sundin, 2021; Koczan et al., 2009; Slack et al., 2022).

52



Cellulose has been observed microscopically in the xylem of apple tissue, and has been
most studied in the context of xylem biofilm formation (Castiblanco & Sundin, 2018; Yuan,
Eldred, Kharadi, et al., 2022). The xylem of plants is a confined, watery environment in which
oxygen levels can become depleted, as in the case of xylem-limited pathogen Xylella fastidiosa
(Shriner & Andersen, 2014). In the xylem, fermentation pathways could be favored, facilitating
cellulose production, autoaggregation, and biofilm formation in E. amylovora. Although xylem
fluid is usually considered nutrient-poor, without high carbohydrate levels as studied here, the
chemistry of apple xylem is highly variable with seasonal fluctuation of ions, proteins,
hormones, amino acids, and some carbohydrates (Biles & Abeles, 1991; Bradfield, 1976;
Davison & Young, 1973; Malaguti et al., 2001; Tromp, 1979). Furthermore, infection of some
plants by bacterial pathogens can greatly alter soluble carbon allocation and transport, and cell
lysis by bacterial pathogens can release additional nutrients (Breia et al., 2021). It is not
inconceivable that E. amylovora could at certain times encounter sufficient carbohydrates in the
xylem tissue for overflow metabolism via fermentation to occur.
In conclusion, we identified a DGC/PDE pair that regulates bacterial cellulose production
via c-di-GMP levels; DgcB and PdeC. We discovered that host sugar availability influences
dgcB expression, causing cellulose production and the associated behavior of autoaggregation.
PdeC is a negative regulator of these processes and is at least partially controlled by the Nterminal GAPES3 domain. Sugar fermentation is associated with cellulose production in E.
amylovora and could favor cellulose production in confined areas such as xylem biofilms, where
the physiochemical properties of cellulose may contribute to fire blight disease.

53



REFERENCES

Abidi, W., Torres-Sánchez, L., Siroy, A., & Krasteva, P. V. (2022). Weaving of bacterial
cellulose by the Bcs secretion systems. FEMS Microbiology Reviews, 46(2).
https://doi.org/10.1093/femsre/fuab051
Aldridge, P., Metzger, M., & Geider, K. (1997). Genetics of sorbitol metabolism in Erwinia
amylovora and its influence on bacterial virulence. Molecular and General Genetics, 256(6). doi:
10.1007/s004380050609.
Almblad, H., Randall, T. E., Liu, F., Leblanc, K., Groves, R. A., Kittichotirat, W., Winsor, G. L.,
Fournier, N., Au, E., Groizeleau, J., Rich, J. D., Lou, Y., Granton, E., Jennings, L. K., Singletary,
L. A., Winstone, T. M. L., Good, N. M., Bumgarner, R. E., Hynes, M. F., Harrison, J. J. (2021).
Bacterial cyclic diguanylate signaling networks sense temperature. Nature Communications,
12(1). https://doi.org/10.1038/s41467-021-22176-2
Ausmees, N., Mayer, R., Weinhouse, H., Volman, G., Amikam, D., Benziman, M., & Lindberg,
M. (2001). Genetic data indicate that proteins containing the GGDEF domain possess
diguanylate cyclase activity. FEMS Microbiology Letters, 204(1), 163–167.
https://doi.org/10.1111/j.1574-6968.2001.tb10880.x
Bellemann, P., & Geider, K. (1992). Localization of transposon insertions in pathogenicity
mutants of Erwinia amylovora and their biochemical characterization. Journal of General
Microbiology,138(5), 931-940.
Bennett, R. A., & Billing, E. (1980). Origin of the Polysaccharide Component of Ooze from
Plants Infected with Erwinia amylovora. Journal of General Microbiology, 116(2).
https://doi.org/10.1099/00221287-116-2-341
Biles, C. L., & Abeles, F. B. (1991). Xylem sap proteins. Plant Physiology, 96(2).
https://doi.org/10.1104/pp.96.2.597
Bobrov, A. G., Kirillina, O., & Perry, R. D. (2005). The phosphodiesterase activity of the HmsP
EAL domain is required for negative regulation of biofilm formation in Yersinia pestis. FEMS
Microbiology Letters, 247(2), 123–130. https://doi.org/10.1016/j.femsle.2005.04.036
Bogs, J., & Geider, K. (2000). Molecular Analysis of Sucrose Metabolism of Erwinia amylovora
and Influence on Bacterial Virulence. Journal of Bacteriology, 182(19).
http://www3.ncbi.nlm.nih.gov
Boureau, T., Elmaarouf-Bouteau, H., Garnier, A., Brisset, M.-N., Perino, C., Pucheu, I., &
Barny, M.-A. (2006). DspA/E, a Type III Effector Essential for Erwinia amylovora
Pathogenicity and Growth In Planta, Induces Cell Death in Host Apple and Nonhost Tobacco
Plants. Molecular Plant-Microbe Interactions MPMI, 19(1), 16–24.
https://doi.org/10.1094/MPMI
54



Bradfield, E. G. (1976). Calcium complexes in the xylem sap of apple shoots. Plant and Soil,
44(2). https://doi.org/10.1007/BF00015906
Breia, R., Conde, A., Badim, H., Fortes, A. M., Gerós, H., & Granell, A. (2021). Plant Sweets:
From sugar transport to plant–pathogen interaction and more unexpected physiological roles.
Plant Physiology, 186(2), 836-852. https://doi.org/10.1093/PLPHYS/KIAB127
Castiblanco, L. F., & Sundin, G. W. (2018). Cellulose production, activated by cyclic di-GMP
through BcsA and BcsZ, is a virulence factor and an essential determinant of the threedimensional architectures of biofilms formed by Erwinia amylovora Ea1189. Molecular Plant
Pathology, 19(1), 90–103. https://doi.org/10.1111/mpp.12501
Clark, D. P. (1989). The fermentation pathways of Escherichia coli . FEMS Microbiology
Letters, 63(3), 223–234. https://doi.org/10.1111/j.1574-6968.1989.tb03398.x
Datsenko, K. A., & Wanner, B. L. (2000). One-step inactivation of chromosomal genes in
Escherichia coli K-12 using PCR products. Biological Sciences, 97(12).
https://doi.org/10.1073/pnas.120163297
Davison, R. M., & Young, H. (1973). Abscisic-acid content of xylem sap. Planta, 109(1).
https://doi.org/10.1007/BF00385455
Dellagi, A., Brisset, M.-N., Paulin, J.-P., & Expert, D. (1998). Dual Role of Desferrioxamine in
Erwinia amylovora Pathogenicity. Molecular Plant-Microbe Interactions MPMI, 11(8).
Edmunds, A. C., Castiblanco, L. F., Sundin, G. W., & Waters, C. M. (2013). Cyclic Di-GMP
modulates the disease progression of Erwinia amylovora. Journal of Bacteriology, 195(10),
2155–2165. https://doi.org/10.1128/JB.02068-12
Enomoto, G., Wallner, T., & Wilde, A. (2023). Control of light-dependent behaviour in
cyanobacteria by the second messenger cyclic di-GMP. MicroLife, 4.
https://doi.org/10.1093/femsml/uqad019
Finn, R. D., Coggill, P., Eberhardt, R. Y., Eddy, S. R., Mistry, J., Mitchell, A. L., Potter, S. C.,
Punta, M., Qureshi, M., Sangrador-Vegas, A., Salazar, G. A., Tate, J., & Bateman, A. (2016).
The Pfam protein families database: Towards a more sustainable future. Nucleic Acids Research,
44(1), 279–285. https://doi.org/10.1093/nar/gkv1344
Galperin, M. Y., Natale, D. A., Aravind, L., & Koonin, E. V. (1999). A specialized version of the
HD hydrolase domain implicated in signal transduction. J Mol Microbiol Biotechnol, 1(2):303-5.
Gottesman, S., & Storz, G. (2011). Bacterial small RNA regulators: Versatile roles and rapidly
evolving variations. Cold Spring Harbor Perspectives in Biology, 3(12).
https://doi.org/10.1101/cshperspect.a003798

55



Haq, A., & Dawes, E. A. (1971). Pyruvic Acid Metabolism and Ethanol Formation in Erwinia
amylovora. Journal of General Microbiology, 68(3). https://doi.org/10.1099/00221287-68-3-295
Hengge, R., Galperin, M. Y., Ghigo, J. M., Gomelsky, M., Green, J., Hughes, K. T., Jenal, U., &
Landini, P. (2016). Systematic nomenclature for GGDEF and EAL domaincontaining cyclic diGMP turnover proteins of Escherichia coli. Journal of Bacteriology, 198(1), 7–11.
https://doi.org/10.1128/JB.00424-15
Herbst, S., Lorkowski, M., Sarenko, O., Nguyen, T. K. L., Jaenicke, T., & Hengge, R. (2018).
Transmembrane redox control and proteolysis of PdeC, a novel type of c‐di‐ GMP
phosphodiesterase. The EMBO Journal, 37(8). https://doi.org/10.15252/embj.201797825
Huang, C. J., Wang, Z. C., Huang, H. Y., Huang, H. Da, & Peng, H. L. (2013). YjcC, a c-diGMP Phosphodiesterase Protein, Regulates the Oxidative Stress Response and Virulence of
Klebsiella pneumoniae CG43. PLoS ONE, 8(7). https://doi.org/10.1371/journal.pone.0066740
Ionescu, M., & Belkin, S. (2009). Overproduction of exopolysaccharides by an Escherichia coli
K-12 rpoS mutant in response to osmotic stress. Applied and Environmental Microbiology,
75(2), 483–492. https://doi.org/10.1128/AEM.01616-08
Jahn, C. E., Selimi, D. A., Barak, J. D., & Charkowski, A. O. (2011). The Dickeya dadantii
biofilm matrix consists of cellulose nanofibres, and is an emergent property dependent upon the
type iii secretion system and the cellulose synthesis operon. Microbiology, 157(10), 2733–2744.
https://doi.org/10.1099/mic.0.051003-0
Kharadi, R. R., Castiblanco, L. F., Waters, C. M., & Sundin, G. W. (2019a). Phosphodiesterase
genes regulate amylovoran production, biofilm formation, and virulence in Erwinia amylovora.
Applied and Environmental Microbiology, 85(1). https://doi.org/10.1128/AEM.02233-18
Kharadi, R. R., Castiblanco, L. F., Waters, C. M., & Sundin, G. W. (2019b). Phosphodiesterase
genes regulate amylovoran production, biofilm formation, and virulence in Erwinia amylovora.
Applied and Environmental Microbiology, 85(1). https://doi.org/10.1128/AEM.02233-18
Kharadi, R. R., Selbmann, K., & Sundin, G. W. (2022a). A complete twelve-gene deletion null
mutant reveals that cyclic di-GMP is a global regulator of phase-transition and host colonization
in Erwinia amylovora. PLoS Pathogens, 18(8). https://doi.org/10.1371/journal.ppat.1010737
Kharadi, R. R., Selbmann, K., & Sundin, G. W. (2022b). A complete twelve-gene deletion null
mutant reveals that cyclic di-GMP is a global regulator of phase-transition and host colonization
in Erwinia amylovora. PLoS Pathogens, 18(8). https://doi.org/10.1371/journal.ppat.1010737
Kharadi, R. R., & Sundin, G. W. (2019). Physiological and microscopic characterization of
cyclic-di-GMP-mediated autoaggregation in Erwinia amylovora. Frontiers in Microbiology, 10.
https://doi.org/10.3389/fmicb.2019.00468

56



Kharadi, R. R., & Sundin, G. W. (2021). Dissecting the process of xylem colonization through
biofilm formation in Erwinia amylovora. In Journal of Plant Pathology, 103, 41–49.
https://doi.org/10.1007/s42161-020-00635-x
Koczan, J. M., McGrath, M. J., Zhao, Y., & Sundin, G. W. (2009). Contribution of Erwinia
amylovora exopolysaccharides amylovoran and levan to biofilm formation: Implications in
pathogenicity. Phytopathology, 99(11), 1237–1244. https://doi.org/10.1094/PHYTO-99-11-1237
Livak, K. J., & Schmittgen, T. D. (2001). Analysis of relative gene expression data using realtime quantitative PCR and the 2-ΔΔCT method. Methods, 25(4), 402–408.
https://doi.org/10.1006/meth.2001.1262
Malaguti, D., Millard, P., Wendler, R., Hepburn, A., & Tagliavini, M. (2001). Translocation of
amino acids in the xylem of apple (Malus domestica Borkh.) trees in spring as a consequence of
both N remobilization and root uptake. Journal of Experimental Botany, 52(361).
https://doi.org/10.1093/jxb/52.361.1665
Mansfield, J., Genin, S., Magori, S., Citovsky, V., Sriariyanum, M., Ronald, P., Dow, M.,
Verdier, V., Beer, S. V., Machado, M. A., Toth, I., Salmond, G., & Foster, G. D. (2012). Top 10
plant pathogenic bacteria in molecular plant pathology. Molecular Plant Pathology, 13(6), 614–
629. https://doi.org/10.1111/j.1364-3703.2012.00804.x
Massie, J. P., Reynolds, E. L., Koestler, B. J., Cong, J. P., Agostoni, M., & Waters, C. M. (2012).
Quantification of high-specificity cyclic diguanylate signaling. Proceedings of the National
Academy of Sciences of the United States of America, 109(31), 12746–12751.
https://doi.org/10.1073/pnas.1115663109
Millard, P., Enjalbert, B., Uttenweiler-Joseph, S., Portais, J. C., & Létisse, F. (2021). Control and
regulation of acetate overflow in Escherichia coli. ELife, 10. https://doi.org/10.7554/eLife.63661
Mohite, B. V., & Patil, S. V. (2014). A novel biomaterial: Bacterial cellulose and its new era
applications. Biotechnology and Applied Biochemistry, 61(2), 101–110.
https://doi.org/10.1002/bab.1148
Mookerjee, S. A., Gerencser, A. A., Nicholls, D. G., & Brand, M. D. (2017). Quantifying
intracellular rates of glycolytic and oxidative ATP production and consumption using
extracellular flux measurements. Journal of Biological Chemistry, 292(17), 7189–7207.
https://doi.org/10.1074/jbc.M116.774471
Morgan, J. L. W., McNamara, J. T., & Zimmer, J. (2014). Mechanism of activation of bacterial
cellulose synthase by cyclic di-GMP. Nature Structural and Molecular Biology, 21(5), 489–496.
https://doi.org/10.1038/nsmb.2803
Naomi, R., Idrus, R. B. H., & Fauzi, M. B. (2020). Plant-vs. Bacterial-derived cellulose for
wound healing: A review. International Journal of Environmental Research and Public Health,
17(18), 1–25. https://doi.org/10.3390/ijerph17186803
57



Norelli, J. L., Jones, A. L., & Aldwinckle, H. S. (2003). Fire Blight Management in the Twentyfirst Century Using New Technologies that Enhance Host Resistance in Apple. Plant Disease,
87(7).
Piqué, N., Miñana-Galbis, D., Merino, S., & Tomás, J. M. (2015). Virulence factors of Erwinia
amylovora: A review. International Journal of Molecular Sciences, 16(6), 12836–12854.
https://doi.org/10.3390/ijms160612836
Pusey, P. L., Rudell, D. R., Curry, E. A., & Mattheis, J. P. (2008). Characterization of Stigma
Exudates in Aqueous Extracts from Apple and Pear Flowers. Hortscience, 43(5).
Randall, T. E., Eckartt, K., Kakumanu, S., Price-Whelan, A., Dietrich, L. E. P., & Harrison, J. J.
(2022). Sensory Perception in Bacterial Cyclic Diguanylate Signal Transduction. In Journal of
Bacteriology (Vol. 204, Issue 2). https://doi.org/10.1128/jb.00433-21
Reinders, A., Hee, C. S., Ozaki, S., Mazur, A., Boehm, A., Schirmer, T., & Jenal, U. (2016).
Expression and genetic activation of cyclic di-GMP-specific phosphodiesterases in Escherichia
coli. Journal of Bacteriology, 198(3), 448–462. https://doi.org/10.1128/JB.00604-15
Richter, A. M., Possling, A., Malysheva, N., Yousef, K. P., Herbst, S., von Kleist, M., &
Hengge, R. (2020). Local c-di-GMP Signaling in the Control of Synthesis of the E. coli Biofilm
Exopolysaccharide pEtN-Cellulose. Journal of Molecular Biology, 432(16), 4576–4595.
https://doi.org/10.1016/j.jmb.2020.06.006
Römling, U., & Galperin, M. Y. (2015). Bacterial cellulose biosynthesis: Diversity of operons,
subunits, products, and functions. Trends in Microbiology, 23(9), 545–557.
https://doi.org/10.1016/j.tim.2015.05.005
Römling, U., Galperin, M. Y., & Gomelsky, M. (2013). Cyclic di-GMP: the First 25 Years of a
Universal Bacterial Second Messenger. Microbiology and Molecular Biology Reviews, 77(1), 1–
52. https://doi.org/10.1128/mmbr.00043-12
Ross, P., Weinhouse, H., Aloni, Y., Michaeli, D., Weinberger-Ohana, P., Mayer, R., Braun, S.,
de Vroomt, E., van der Marelt, G. A., van Boomt, J. H., & Benziman, M. (1973). Regulation of
cellulose synthesis in Acetobacter xylinum by cyclic diguanylic acid. M. J. PhysioL, 232(6).
Rossi, C. S., Barrientos-Moreno, L., Paone, A., Cutruzzolà, F., Paiardini, A., Espinosa-Urgel,
M., & Rinaldo, S. (2022). Nutrient Sensing and Biofilm Modulation: The Example of L-arginine
in Pseudomonas. International Journal of Molecular Sciences, 23(8).
https://doi.org/10.3390/ijms23084386
Ryjenkov, D. A., Simm, R., Römling, U., & Gomelsky, M. (2006). The PilZ domain is a receptor
for the second messenger c-di-GMP: The PilZ domain protein YcgR controls motility in
enterobacteria. Journal of Biological Chemistry, 281(41), 30310–30314.
https://doi.org/10.1074/jbc.C600179200
58



Salomone-Stagni, M., Bartho, J. D., Kalita, E., Rejzek, M., Field, R. A., Bellini, D., Walsh, M.
A., & Benini, S. (2018). Structural and functional analysis of Erwinia amylovora SrlD. The first
crystal structure of a sorbitol-6-phosphate 2-dehydrogenase. Journal of Structural Biology,
203(2), 109–119. https://doi.org/10.1016/j.jsb.2018.03.010
Schachterle, J. K., Onsay, D. M., & Sundin, G. W. (2019). Small RNA ArcZ Regulates
Oxidative Stress Response Genes and Regulons in Erwinia amylovora. Frontiers in
Microbiology, 10. https://doi.org/10.3389/fmicb.2019.02775
Schachterle, J. K., Zeng, Q., & Sundin, G. W. (2019). Three Hfq-dependent small RNAs regulate
flagellar motility in the fire blight pathogen Erwinia amylovora. Molecular Microbiology,
111(6), 1476–1492. https://doi.org/10.1111/mmi.14232
Sebaihia, M., Bocsanczy, A. M., Biehl, B. S., Quail, M. A., Perna, N. T., Glasner, J. D.,
DeClerck, G. A., Cartinhour, S., Schneider, D. J., Bentley, S. D., Parkhill, J., & Beer, S. V.
(2010). Complete genome sequence of the plant pathogen Erwinia amylovora strain ATCC
49946. Journal of Bacteriology, 192(7), 2020–2021. https://doi.org/10.1128/JB.00022-10
Serra, D. O., Richter, A. M., & Hengge, R. (2013). Cellulose as an architectural element in
spatially structured escherichia coli biofilms. Journal of Bacteriology, 195(24), 5540–5554.
https://doi.org/10.1128/JB.00946-13
Shriner, A. D., & Andersen, P. C. (2014). Effect of Oxygen on the Growth and Biofilm
Formation of Xylella fastidiosa in Liquid Media. Current Microbiology, 69(6), 866–873.
https://doi.org/10.1007/s00284-014-0660-2
Slack, S. M., Schachterle, J. K., Sweeney, E. M., Kharadi, R. R., Peng, J., Botti-Marino, M.,
Bardaji, L., Pochubay, E. A., & Sundin, G. W. (2022). In-Orchard Population Dynamics of
Erwinia amylovora on Apple Flower Stigmas. Phytopathology, 112(6), 1214–1225.
https://doi.org/10.1094/PHYTO-01-21-0018-R
Slack, S. M., Zeng, Q., Outwater, C. A., & Sundin, G. W. (2017). Microbiological examination
of erwinia amylovora exopolysaccharide ooze. Phytopathology, 107(4), 403–411.
https://doi.org/10.1094/PHYTO-09-16-0352-R
Sutton, D. D., & Starr, M. P. (1959). Anaerobic dissimilation of glucose by Erwinia amylovora.
Journal of Bacteriology, 78(3), 427-31. https://journals.asm.org/journal/jb
Takle, G. W., Toth, I. K., & Brurberg, M. B. (2007). Evaluation of reference genes for real-time
RT-PCR expression studies in the plant pathogen Pectobacterium atrosepticum. BMC Plant
Biology, 7. https://doi.org/10.1186/1471-2229-7-50
Tamura, K., Stecher, G., Peterson, D., Filipski, A., & Kumar, S. (2013). MEGA6: Molecular
evolutionary genetics analysis version 6.0. Molecular Biology and Evolution, 30(12), 2725–
2729. https://doi.org/10.1093/molbev/mst197
59



Tarutina, M., Ryjenkov, D. A., & Gomelsky, M. (2006). An unorthodox bacteriophytochrome
from Rhodobacter sphaeroides involved in turnover of the second messenger c-di-GMP. Journal
of Biological Chemistry, 281(46), 34751–34758. https://doi.org/10.1074/jbc.M604819200
Tromp, J. (1979). Seasonal Variations in the Composition of Xylem Sap of Apple with Respect
to K, Ca, Mg, and N. Zeitschrift Für Pflanzenphysiologie, 94(3). https://doi.org/10.1016/s0044328x(79)80157-9
Waldron, E. J., Snyder, D., Fernandez, N. L., Sileo, E., Inoyama, D., Freundlich, J. S., Waters, C.
M., Cooper, V. S., & Neiditch, M. B. (2019). Structural basis of DSF recognition by its receptor
rpfr and its regulatory interaction with the DSF synthase RpfF. PLoS Biology, 17(2).
https://doi.org/10.1371/journal.pbio.3000123
Wang, D., Korban, S. S., & Zhao, Y. (2009). The Rcs phosphorelay system is essential for
pathogenicity in Erwinia amylovora. Molecular Plant Pathology, 10(2), 277–290.
https://doi.org/10.1111/j.1364-3703.2008.00531.x
Wang, Z., & Stutte, G. W. (1992). The Role of Carbohydrates in Active Osmotic Adjustment in
Apple under Water Stress. J. AMER. SOC. HORT. SCI, 117(5).
Yap, M. N., Yang, C. H., Barak, J. D., Jahn, C. E., & Charkowski, A. O. (2005). The Erwinia
chrysanthemi type III secretion system is required for multicellular behavior. Journal of
Bacteriology, 187(2), 639–648. https://doi.org/10.1128/JB.187.2.639-648.2005
Yoo, B. K., & Chen, J. (2010). Role of Cellulose in Protecting Shiga Toxin-Producing
Escherichia coli against Osmotic and Chlorine Stress. J Food Pro, 73(11), 2084-2088. doi:
10.4315/0362-028x-73.11.2084.
Yuan, X., Eldred, L. I., Kharadi, R. R., Slack, S. M., & Sundin, G. W. (2022). The RNA-Binding
Protein ProQ Impacts Exopolysaccharide Biosynthesis and Second Messenger Cyclic di-GMP
Signaling in the Fire Blight Pathogen Erwinia amylovora. Applied and Environmental
Microbiology, 88(9). https://doi.org/10.1128/aem.00239-22
Yuan, X., Eldred, L. I., & Sundin, G. W. (2022). Exopolysaccharides amylovoran and levan
contribute to sliding motility in the fire blight pathogen Erwinia amylovora. Environmental
Microbiology, 24(10), 4738–4754. https://doi.org/10.1111/1462-2920.16193
Yuan, X., McGhee, G. C., Slack, S. M., & Sundin, G. W. (2021). A novel signaling pathway
connects thiamine biosynthesis, bacterial respiration, and production of the exopolysaccharide
amylovoran in Erwinia amylovora. Molecular Plant-Microbe Interactions, 34(10).
https://doi.org/10.1094/MPMI-04-21-0095-R
Zeng, Q., McNally, R. R., & Sundin, G. W. (2013). Global small RNA chaperone Hfq and
regulatory small RNAs are important virulence regulators in erwinia amylovora. Journal of
Bacteriology, 195(8), 1706–1717. https://doi.org/10.1128/JB.02056-12
60



Zeng, Q., & Sundin, G. W. (2014). Genome-wide identification of Hfq-regulated small RNAs in
the fire blight pathogen Erwinia amylovora discovered small RNAs with virulence regulatory
function. BMC Genomics, 15(1), 414. doi: 10.1186/1471-2164-15-414.
Zhao, Y., Sundin, G. W., & Wang, D. (2009). Construction and analysis of pathogenicity Island
deletion mutants of Erwinia amylovora. Canadian Journal of Microbiology, 55(4), 457–464.
https://doi.org/10.1139/W08-147
Zhao, Y., Wang, D., Nakka, S., Sundin, G. W., & Korban, S. S. (2009). Systems level analysis of
two-component signal transduction systems in Erwinia amylovora: Role in virulence, regulation
of amylovoran biosynthesis and swarming motility. BMC Genomics, 10.
https://doi.org/10.1186/1471-2164-10-245

61



APPENDIX A: BIOINFORMATIC OBSERVATIONS

Figure A1. Screenshot summary of BLAST and SMART database comparison of amino acid
sequences of Es. coli DgcC and PdeK and E. amylovora DgcB and PdeC.
In the model organism Es. coli, production of a specific EPS pEtN-cellulose is regulated
specifically by the DGC DgcC and PDE PdeK. This comparison of the amino acid sequence and
domain structure was conducted to evaluate any potential similarities between these enzymes and
the E. amylovora enzymes experimentally demonstrated to regulate cellulose production, DgcB
and PdeC. The PDEs appear very similar to each other in domain structure and sequence
similarity, whereas the DGCs had less percent identity in the amino acid sequence but still
showed the same domain structure.

62



E

Figure A2. Location of the pdeC gene next to the bcs operon in E. amylovora. Regulatory genes
are often encoded close to the genes they regulate; the location of the pdeC gene immediately
next to the bcs operon is evidence of the importance of pdeC as a regulator of cellulose in E.
amylovora.

63



Figure A3. Image at 30x magnification of washed bacterial cellulose produced by Ea1189
against a dark background for contrast. The fluffy, cloudlike aggregates showed affinity for each
other, often combining with each other when gently stirred, and appeared to be resistance to
dissolution in water, growth medium, or phosphate buffer solutions. Cellulose also collected in
the bottom of a centrifuge tube after centrifugation, unlike amylovoran, which remains in the
supernatant as evidenced by the amylovoran assay used in previous studies (Kharadi et al.,
2019). We make the simple observation that the insolubility of cellulose compared to other EPSs
could give this EPS a unique role in virulence, such as protecting cells from the flow of xylem
fluid.

64



APPENDIX B: DOMAIN-DIRECTED MUTAGENESIS

LB Arabinose
Sorbitol
GAPES3

AMP

GGDEF

EAL

T
pBAD
empty vector

GAPES3

AMP

AMP

GGDEF

GGDEF

EAL

EAL

GAPES3

Figure B1. Truncation of the pdeC gene on a plasmid vector abolishes PDE activity.
Because the experiments in this study demonstrated that PdeC negatively regulates cellulose
production in E. amylovora, and because many PDE enzymes are in turn regulated by N-terminal
signaling domains, we investigated whether the GAPES3 domain could alter PdeC enzymatic
activity. The function of the GAPES3 domain is previously unknown. We sought to create a
strain with a truncated gene that would eliminate the region containing this signaling domain.
We hypothesized that this truncation would greatly increase or decrease PdeC activity, which
could then be visualized by testing for cellulose production.
A truncated pdeC gene was designed by designing primers to start after the N-terminal GAPES3
domain region of the pdeC gene and insert a start codon, creating a shortened pdeC gene with a
start codon that was cloned into an arabinose-inducible plasmid vector. Graphics on the left
depict the domain structure of the PdeC protein, with red X’s indicate the missing gene or
domain in the plasmid. The empty vector, full length gene, and truncated gene were cloned into
the ΔpdeC strain and tested for cellulose production. While complementation of the full length
pdeC gene decreased cellulose production, as observed in other experiments in this study,
complementation with the truncated version missing the GAPES3 domain did not appear to alter
the cellulose phenotype, indicating that the GAPES3 domain could be critical to PDE function.
Note that one of the transmembrane domains (noted in blue) was also deleted in this process,
potentially affecting membrane localization. This limits the quality of the results, as truncated
proteins may not fold correctly, be degraded prematurely, or not be transported to the correct
location of the cell. The rest of the protein was intact, including the EAL active site, as
65



Figure B1. (cont’d)
confirmed by Sanger sequencing. These interesting results, despite potential drawbacks, inspired
us to try a different approach to verifying if the GAPES3 is essential for PdeC activity. The
technique chosen was domain-directed mutagenesis, where error-prone DNA polymerase
randomly mutagenizes only a select region of the gene. This maintains the gene at exactly the
same length and maintains the sequence and location of the transmembrane domains and the rest
of the gene except the GAPES3 domain, eliminating the drawbacks of the previous experiment.

LB Sorbitol

T
empty
vector

mutageni ed

Figure B2. Domain directed mutagenesis resulted in a mutagenized GAPES3 domain with a fully
intact gene – mutagenized domain denoted in blue.
Domain-directed mutagenesis was performed on the GAPES3 domain of pdeC using the EZ
Clone GeneMorph II Domain Mutagenesis Kit (Agilent) according to manufacturer instructions.
In brief, error-prone polymerases were used to amplify the GAPES3 region of the pdeC gene,
and the product of this PCR was used as a megaprimer to resynthesize the pBBR1-MCS5::pdeC
plasmid, with the result of an intact plasmid and full-length pdeC gene with a randomly
mutagenized GAPES3 domain. Results were analyzed via Sanger sequencing.
Mutant plasmids from PCR based mutagenesis were screened on Congo Red amended medium
for increased cellulose production, and the results were confirmed with Sanger sequencing. 600
colonies were screened, 11 of which showed enhanced cellulose production, indicating a lack of
PdeC activity. Of these 11, 10 were either frame shift or nonsense mutations which altered the
sequence of the entire gene. The selected strain showed a lack of PdeC activity as determined by
a cellulose assay. The strain shown possessed a completely intact pdeC gene except for 6
missense mutations in the GAPES3 domain, showing that this signaling domain controls
enzymatic activity of the EAL domain.
66



R39I

M46K

P75L

D65

GAPES3 residues 36-82

149L I151F

GAPES3 residues 141-157

Figure B3. Domain directed mutagenesis caused six amino acid changes in the GAPES3 domain.
Labeled arrows show the location and amino acid changes that occurred. Based on cellulose
assays, these residues appear to be critical to the function of PdeC.
The intriguing results from the domain directed mutagenesis experiment prompted us to examine
the importance of the 6 amino acid residues that were changed. We aligned the GAPES3 domain
amino acid sequence of E. amylovora to those from other related organisms.
Amino acid sequences from 29 bacterial GAPES3 domains were aligned as shown, showing the
high homology of these regions, especially in the amino acids that were randomly mutated. This
is especially interesting, as the conservation of these residues provides further evidence that they
are essential to proper protein function. In addition, this analysis reveals the conservation of the
GAPES3 domain overall. Perhaps this unknown signaling domain plays a similar role in a wide
variety of organisms.

67



APPENDIX C: AMINO ACID SCREENING

ala

arg

asn

asp

cys

gln

pro

ser

thr

trp

tyr

val

glu

gly

his

iso

leu

lys

met

phe

Figure C1. Cellulose production on MBMA minimal medium with sorbitol and 2 mM selected
amino acids. Top row: WT. Middle row: pdeC mutant. Bottom row: bcsA mutant.
WT Ea1189 produces cellulose on LB agar when amended with sorbitol, but not on plain LB
medium. We tested a minimal medium (MBMA) containing the same (20mg/mL) amount of
sorbitol, and found that the WT strain also failed to produce cellulose. Since LB medium is
composed mainly of amino acids, we hypothesized that the combination of sugars and amino
acids is necessary for the production of cellulose.
In addition, since the signal the GAPES3 domain of PdeC is unknown, we hypothesized that the
availability of one or several amino acids influences the activity of PdeC. 2 mM concentrations
68



Figure C1. (cont’d)
of the 20 proteinogenic amino acids were added to MBMA agar plates amended with Congo Red
dye and E. amylovora strains were tested for cellulose production.
Little or no effect was observed from the amino acid additions on the WT or bcsA mutant.
However, the pdeC mutant showed some striking differences, particularly methionine (met),
which produced less cellulose. This could suggest that the availability of amino acids impacts
dgcB expression, but the results are inconclusive. It does not appear that individual free amino
acids alter the activity of pdeC.
This experiment was not continued because none of the amino acids appeared to have an effect
on the WT strain cellulose production. While the methionine results on the pdeC mutant strain
certainly do appear interesting, addition of methionine to liquid LB and MBMA media did not
produce a consistently different phenotype from the control (data not shown). Further
experiments in this direction were not conducted.
The signal that impacts the periplasmic GAPES3 domain of PdeC remains to be discovered and
would present a fascinating future research topic with implications in plant pathology and
molecular signal transduction.

69