Xanthomonas in Florida

MULTI-LOCUS AND WHOLE-GENOME SEQUENCE ANALYSIS OF PSEUDOMONADS AND XANTHOMONADS IMPACTING TOMATO PRODUCTION IN FLORIDA

SUJAN TIMILSINA

A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2016

To my Family

ACKNOWLEDGMENTS

I would like to take this opportunity to express my gratitude towards Dr. Gary E.

Vallad, committee chair and Dr. Jeffrey B. Jones, co-chair, for their constant support, encouragement and guidance throughout my graduate studies. I couldn’t have done this without their scientific inputs and personal mentorship. I would also like to thank Dr.

Erica M. Goss for all her advice, recommendations and support. I would also like to extend my gratitude to my committee members, Dr. Bryan Kolaczkowski and Dr. Sam

Hutton for their valuable suggestions and guidance. Very special thanks to Gerald V.

Minsavage, for all his expertise, creativity, recommendations and constructive criticism.

We collaborated with Dr. Frank White, Dr. Brian Staskawicz and Dr. Jim Preston for some aspects of my research and to write articles and reviews. I would like to thank them all for providing me the opportunity.

During my PhD, I had the privilege to work with colleagues from the 2560 Jones lab in Gainesville and Vegetable Pathology lab in Balm and I thank the lab family. I appreciate the time and technical support from Dr. Neha Potnis. I thank my labmates

Amanda Strayer, Juliana Pereira, Serhat Kara, Eric A. Newberry, Alberto Gochez, Yang

Hu, and Deepak Shantaraj and numerous lab mates over the years for their co- operation and assistance. I would also like to thank all my Nepalese friends in

Gainesville for their lively company during my stay here.

I also want to thank my parents and brother, who helped me to shape my career.

I specially thank my wife, Naweena, for her undying support, encouragement and patience. Thank you all for making this possible.

TABLE OF CONTENTS

page

ACKNOWLEDGMENTS ...... 4

LIST OF TABLES ...... 7

LIST OF FIGURES ...... 8

ABSTRACT ...... 10

CHAPTER

1 LITERATURE REVIEW ...... 12

Pseudomonas in Florida ...... 12 Pseudomonas Characterization ...... 13 Taxonomy of Pseudomonas ...... 13 Xanthomonas in Florida ...... 14 Population Studies of Bacterial Spot Causing Xanthomonas Species ...... 16 Type III Secreted Effectors ...... 17 Project Goals and Objectives ...... 19

2 PHYLOGENETIC STUDY REVEALS A DISTINCT PHYLOGROUP OF PSEUDOMONAS CICHORII IN FLORIDA ...... 21

Introduction ...... 21 Materials and Methods...... 23 Bacterial Strains and Isolation ...... 23 Pathogenicity Tests ...... 23 PCR Amplification and Gene Sequencing ...... 24 Phylogenetic Analysis ...... 25 GenBank Accession Numbers ...... 26 Results ...... 26 Isolation and Phenotypic characterization ...... 26 Pathogenicity Test ...... 26 Genetic Characterization ...... 27 Discussion ...... 29

3 PSEUDOMONAS FLORIDAE SP. NOV., A NOVEL BACTERIAL PATHOGEN ISOLATED FROM TOMATO ...... 39

Introduction ...... 39 Materials and Methods...... 41 Bacterial Strains and Preliminary Characterization ...... 41 Biochemical and Physiological Tests...... 41 Pathogenicity Assay ...... 42

Gene Sequencing and Sequence Analysis ...... 42 Biolog Assay and Fatty Acid Profiles ...... 43 Whole Genome Sequence Comparison ...... 44 Results ...... 44 Bacterial Characterization ...... 44 Pathogenicity Assay ...... 45 Sequence Analysis ...... 45 Biolog Assay and FAME-ID ...... 46 Whole Genome Sequence Analysis ...... 47 Discussion ...... 48

4 ANALYSIS OF SEQUENCED GENOMES OF XANTHOMONAS PERFORANS IDENTIFIES CANDIDATE TARGETS FOR RESISTANCE BREEDING IN TOMATO ...... 60

Introduction ...... 60 Materials and Methods...... 65 Bacterial Strains Collection and Pathogenicity Testing ...... 65 Gene Sequencing ...... 65 Sequence Analysis ...... 66 Sequence Submission ...... 67 Results ...... 67 Pathogenicity and Race Determination ...... 67 Grouping of T3 Strains by gapA Sequence ...... 67 Effector Gene Sequences ...... 68 Presence of AvrBsT in T3 X. perforans ...... 69 Discussion ...... 70

5 CORE GENOME MULTILOCUS SEQUENCE TYPING OF XANTHOMONAS PERFORANS ...... 87

Introduction ...... 87 Materials and Methods...... 90 Bacterial Strains and Genome Sequencing ...... 90 Multilocus Sequence Analysis ...... 91 Results ...... 93 Bacterial Strains and Multilocus Sequence Analysis ...... 93 Core Gene Comparisons ...... 93 Phylogenetic Analysis ...... 95 Discussion ...... 96

6 SUMMARY AND CONCLUSION ...... 104

APPENDIX: SCRIPTS USED FOR CORE GENOME ANALYSIS ...... 111

LIST OF REFERENCES ...... 114

BIOGRAPHICAL SKETCH ...... 127

LIST OF TABLES

Table page

2-1 List of P. cichorii strains used for this study...... 33

3-1 Phenotypic characteristics of P. floridae GEV388T and closely related plant pathogenic Pseudomonas species...... 51

3-2 Cellular fatty acid composition of eight strains of P. floridae, and its closest relative based on fatty acid composition...... 52

3-3 Genome properties of P. floridae sp. nov strain GEV388 compared with selected reference strains ...... 53

3-4 Pairwise ANIb using P. floridae sp. nov. strain GEV388 and representative strains from different phylogroups of P. syringae ...... 54

4-1 Xanthomonas perforans strains used in this study and their genotypes ...... 76

4-2 List of primers used in this study ...... 78

4-3 Xanthomonas strains/population isolated over 21 years in Florida ...... 79

5-1 List of X. perforans strains isolated in Florida used in this study ...... 100

LIST OF FIGURES

Figure page

2-1 Bacterial disease symptom observed in tomato foliage caused by P. cichorii. ... 34

2-2 Disease symptoms on tomato and lettuce hosts (48 hpi) inoculated with representative P. cichorii strains isolated from Florida ...... 35

2-3 Maximum likelihood phylogeny of Pseudomonas strains isolated from Florida using concatenated sequences of gyrB and rpoD genes ...... 36

2-4 Maximum likelihood phylogenetic distribution of Pseudomonas strains isolated from Florida using concatenated sequences ...... 37

2-5 Maximum likelihood phylogenetic distribution of representative Pseudomonas strains isolated from Florida tomato using 16S rRNA sequences ...... 38

3-1 Foliar symptoms observed in tomato caused by the novel Pseudomonas strains...... 55

3-2 Specific binding of FITC labeled Concanavalin-A with levan ...... 56

3-3 Maximum likelihood phylogenetic analysis using 16S rRNA sequences...... 57

3-4 Maximum likelihood phylogenetic tree using concatenated sequences from four housekeeping genes: gap1, gyrB, gltA and fusA...... 58

3-5 Maximum likelihood phylogenetic comparison using concatenated sequences from four housekeeping genes: cts, gyrB, gap1 and rpoD ...... 59

4-1 Maximum likelihood of phylogenetic analysis of Xanthomonas perforans strains using gapA gene sequence...... 80

4-2 gapA partial sequence alignment for four different allelic forms observed in X. perforans...... 81

4-3 Complete coding region of avrXv3 as observed in the type strain of X. perforans, 91-118 ...... 82

4-4 Partial gene sequence alignment for xopQ in X. perforans. XopQ-allele-1 is found in the type strain of X. perforans ...... 83

4-5 Gel-electrophoresis of avrBsT gene including ~0.5kb flanking regions on either side of Xanthomonas perforans isolates collected in 2006 ...... 84

4-6 Maximum likelihood phylogenetic tree of avrBsT showing sequence identity between alleles in X. vesicatoria and X. perforans strains ...... 85

4-7 Infograph showing shifts in the Xanthomonas population on tomato in Florida over time. Prior to 1991, only X. euvesicatoria was identified ...... 86

5-1 Maximum likelihood phylogenetic distribution of Xanthomonas perforans strains isolated from Florida based on housekeeping gene sequences of fusA, gapA, gltA, gyrB, lacF and lepA ...... 102

5-2 Maximum likelihood phylogenetic distribution of Xanthomonas perforans strains using concatenated sequences of core genes (~1.2 Mb) ...... 103

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

MULTI-LOCUS AND WHOLE-GENOME SEQUENCE ANALYSIS OF PSEUDOMONADS AND XANTHOMONADS IMPACTING TOMATO PRODUCTION IN FLORIDA

Sujan Timilsina

December 2016

Chair: Gary E. Vallad Cochair: Jeffrey B. Jones Major: Plant Pathology

Bacterial disease is a major concern for tomato growers in Florida. There are two

major bacterial genera significantly affecting tomato production in Florida,

Pseudomonas and Xanthomonas. Among Pseudomonads, P. syringae pv. tomato, P.

viridiflava and P. corrugata were previously reported from tomato in Florida. Four

different Xanthomonas cause bacterial spot of tomato but X. perforans is the dominant

species in Florida.

We characterized Pseudomonas strains isolated in Florida during outbreak in

2010/11. We employed phenotypic and genotypic approaches to characterize the

bacteria. We identified pathogenic strains of Pseudomonas cichorii isolated from tomato

in Florida. The strains were identified based on standard LOPAT assay, pathogenicity

assay and genetic characterization based on multiple housekeeping genes.

Phylogenetic analysis identified a distinct phylogroup of P. cichorii strains present in

Florida that infected multiple hosts.

Furthermore, we identified a new pathogenic species of Pseudomonas infecting

tomato. We characterized the group of atypical strains using phenotypic characteristics

that included LOPAT assay, metabolic utilization and cellular fatty acid profiles along

with genotypic characterization and phylogenetic analysis using housekeeping genes and whole genome sequences. Our results strongly supported that the group of

Pseudomonas strains recently isolated from tomato in Florida belong to a new

pathogenic species.

Bacterial disease management is challenged by rapid and continuous evolution

of the pathogen. We observed significant changes in bacterial spot causing X. perforans

population over time. Variable effector profiles and race shifts were observed. We

examined the variation in putative targets of resistance and found two effectors that are

the best targets for resistance breeding against bacterial spot of tomato.

In addition, we conducted phylogenetic analysis based on core genes identified

between the X. perforans strains isolated from Florida. Whole genome sequences were

used to identify the current population structure and sources of genetic modifications in

the X. perforans population isolated in Florida. A core genome multilocus sequence

analysis approach was employed for allelic profiling of X. perforans genomes. Future

studies will be focused towards phylogenetic analyses to study pathogen population

dynamics and implementing disease management strategies.

CHAPTER 1 LITERATURE REVIEW

Tomato is one of the major crops cultivated in the United States. A total of 97,400 acres of fresh market tomatoes were planted and harvested in 2013 with an approximate value of $1.5 billion (USDA 2014). Florida is the largest producer of fresh market tomato in the United States with a total value of approximately $453 million (Freeman et al. 2016). Fresh market tomato production in Florida accounts for approximately 36% of production acreage and harvest in the

U.S. and is the largest agricultural industry after citrus production in the state. The tomato crop

is vulnerable to a number of pathogenic bacteria.

Pseudomonas in Florida

The genus Pseudomonas includes several widely distributed pathogenic species that

affect a wide range of hosts in diverse environments. Several pathogenic Pseudomonas species

have been observed in tomato. Diseases reported on tomato caused by Pseudomonas are

bacterial speck, syringae leaf spot, bacterial blight and tomato pith necrosis (Agrios 2005).

Bacterial speck of tomato caused by P. syringae pv. tomato (McCarter et al. 1983), bacterial

blight caused by P. viridiflava (Jones et al. 1984) and tomato pith necrosis caused by P.

corrugata (Jones et al. 1983) were previously reported from Florida. Symptoms of bacterial

speck appear as dark brown to black color lesions that vary in shape and size with a distinct

yellow halo surrounding the lesions (Agrios 2005). The lesions on above ground parts of tomato

are similar but smaller than bacterial spot symptoms and coalesce to form larger scabby areas

on the fruit surface. Bacterial blight causes extensive tissue necrosis and affects almost every

above ground part of the plant, and pith necrosis causes extensive vascular and pith browning.

In Florida, regular cold fronts passing through the state during the winter lead to heavy rainfall

along with cooler temperatures, favoring Pseudomonas survival and spread. Apart from tomato,

a recent increase in other pathogenic Pseudomonas cichorii has been reported from hosts

including Stevia rebudiana and Duranta erecta (Strayer et al. 2012, Gumtow et al. 2013). The P.

cichorii species is considered a warm weather pathogen as compared to P. syringae and can

grow in conditions with high temperature and humidity (Jones et al. 1984).

Pseudomonas Characterization

Pseudomonas strains are phenotypically characterized using the standard

LOPAT tests based on Levan production, Oxidase reaction, Pectinolytic activity,

Arginine dihydrolase activity and their ability to cause a hypersensitive reaction on

Tobacco (Lelliott et al. 1966). In addition, with the inclusion of host range, there are 40

well-characterized pathovars of plant pathogenic Pseudomonas syringae (Palleroni

2008, Palleroni 2010). Although LOPAT profiles of Pseudomonas strains are highly

effective for preliminary pathogen identification, high diversity within the Pseudomonas

complex limits the use of LOPAT profiles for further characterization and speciation

(Hwang et al. 2005). Sequencing of conserved housekeeping genes and 16S rRNA

genes are recommended for further identification and classification of phytopathogenic

Pseudomonas. Hwang et al. (2005) suggested using the conserved gene sequences of

gap1, gltA, gyrB and rpoD to effectively characterize Pseudomonas syringae strains.

Multilocus sequence typing/analysis based on these housekeeping and 16S rRNA

genes are commonly used to evaluate the phylogenetic relationships between strains

and their evolution (Gevers et al. 2005).

Taxonomy of Pseudomonas

Bacteria of the genus Pseudomonas are gram-negative and rod shaped. The members are present in a wide range of natural environments and include number of species that are pathogenic on plants and animals (Anwar et al. 2016). The taxonomy of

Pseudomonas is frequently updated with the identification of novel characteristics and

advances in molecular techniques. For taxonomic purposes, DNA-DNA hybridization

(DDH) is often used to differentiate bacterial species (Bull and Koike 2015). DNA re- association above 70% between two bacterial strains suggests that the strains belong to the same species. Gardan et al. (1999) used DDH to differentiate various strains of the plant pathogenic P. syringae complex into genomospecies groups. However, DDH

approach is not repeatable, it requires isotope use, the process is prone to errors and

can be applied only for closely related species or subspecies (Sentausa and Fournier

2013). Recent characterization and taxonomic approaches are sequence based.

Multilocus sequence analysis that uses sequences from multiple conserved

housekeeping genes are used for phylogenetic and species characterization. The P.

syringae complex is grouped in phylogroups based on multilocus sequence analysis

(Bartoli et al. 2014, Bull et al. 2011, Berge et al. 2014). With increased availability of

high-throughput sequencing and increased computational ability, recent studies are

using complete genomes comparisons for characterization, epidemiology and

population studies. Pseudomonas taxonomy based on whole genome sequences is

proposed (Gomila et al. 2015). Sequence comparison of specific housekeeping genes,

analyzing pathogenicity islands and various approaches of whole genome sequence

comparisons like the average nucleotide identity (ANI) and the genome-genome

distance calculator (GGDC) are used for taxonomic purposes (Konstantinidis et al.

2006, Auch et al. 2010). More than 140 validated species of Pseudomonas currently

exist, and with rapid genome sequencing and comparison, the number of Pseudomonas

species reported over the years has increased significantly (Winsor et al. 2010).

Xanthomonas in Florida

Similar to bacterial speck, bacterial spot, caused by Xanthomonas, is another

common and economically significant disease of tomato. Xanthomonads are

characterized by their distinct yellow color in nutrient agar media. They are Gram-

negative gammaproteobacteria, aerobic, catalase- and aminopeptidase-positive, but

oxidase and urease negative (Garrity et al. 2006). Four different species of

Xanthomonas, to include X. euvesicatoria, X. vesicatoria, X. perforans and X. gardneri,

cause bacterial spot disease (Jones et al. 2004). Strains of X. perforans and X. vesicatoria are associated with tomato disease only but X. euvesicatoria and X. gardneri are associated with tomato and pepper disease. Disease symptoms can develop on all above ground parts of plants including leaves, stems and fruits. Initial symptoms consist of water soaked lesions that turn dark brown to black as they enlarge and dry. Fruit lesions are found to have distinct raised margins with a scabby appearance and can be larger than the foliar lesions. The center of fruit lesions usually begin to sink as the lesion expands and become prone to secondary infection by other potential pathogens.

Flower infection can result in severe blossom drop (Agrios 2005). In the case of X. perforans, foliar lesions can develop a shot-hole appearance when conditions are favorable (Potnis et al. 2015). Likewise, X. gardneri symptoms are relatively large with a water-soaked appearance on the foliage and with raised and scabby appearing spots in the fruits. Significant crop loss resulting from seedling damage is reported from areas with warm temperature and high humidity (Momol et al. 2008). Fresh market tomato production in Florida is highly affected by this disease due to constant presence of the casual bacterium and weather conditions that are often conducive to the disease (Sun et al. 2002). Fruit lesions reduce fruit quality and are also prone to infection by secondary post-harvest pathogens. Using costs and values from 2007 to 2008, it was

estimated that bacterial spot cost Southwest Florida approximately $3000 per acre

(VanSickle and Weldon 2009).

Population Studies of Bacterial Spot Causing Xanthomonas Species

Among the four species responsible for bacterial spot of tomato, X. euvesicatoria

and X. gardneri occur on both tomato and pepper, but X. perforans and X. vesicatoria

are primarily pathogens only of tomato. Different populations of Xanthomonas are

reported from various geographic locations. For example, X. gardneri is more prevalent

in tomato producing areas with relatively cooler temperatures (Potnis et al. 2015).

Interestingly, a recent comparison of global X. gardneri strains determined that the

population is clonal, suggesting a recent global spread of the X. gardneri strains

(Timilsina et al. 2015). Similar diversity and population studies have been conducted in

Southwest Indian Ocean islands (Hamza et al. 2010). These reported Xanthomonas

population were described based on the sequences of conserved housekeeping genes.

Along with these strains, various atypical strains have been reported that share

sequence similarities between different species from other locations like Nigeria,

Ethiopia, Grenada and India (Hamza et al. 2010, Jibrin et al. 2015, Kebede et al. 2014),

which has limited the use of multilocus sequence analysis for population studies.

Additionally, multiple species of Xanthomonas were reported from a same tomato

production area, which increases the likelihood of genetic exchanges between the

strains.

In Florida, several population changes have been observed for Xanthomonas.

Strains of T1 X. euvesicatoria were the only strains isolated from symptomatic tomato

before 1991. The T3 X. perforans strains were isolated since 1991. Three different

bacteriocins produced by T3 X. perforans were shown to be important in the population

shift from T1 to T3 (Tudor-Nelson et al. 2003, Hert et al. 2005). Later surveys conducted in 2006 and 2012 showed another shift from T3 to T4. These surveys identified 75% of the strains collected in 2006 as T4 (Horvath et al. 2012), and 100% T4 X. perforans in

Florida by 2012 (Vallad et al. 2014). During these population changes, only one strains of X. perforans was isolated from pepper (Schwatz et al. 2015). As part of this population shift in X. perforans, evidence of recombination between X. perforans and X.

euvesicatoria species was observed among a subgroup of X. perforans strains

(Schwartz et al. 2015, Timilsina et al. 2015).

The international production of commercial seed has likely affected the

distribution of these Xanthomonas species and their global movement. However,

multilocus sequence analysis using only a limited number of conserved housekeeping

genes is not sufficient to monitor movement of bacterial strains or clonal lineages. A

thorough understanding of Xanthomonas evolution would be more explicit by utilizing

the next-generation sequencing approaches (Chan et al. 2012). Core genome

multilocus sequence typing (cgMLST) approaches were developed for different bacteria

to standardize analyses based on whole genome sequences (de Been et al. 2015, Kohl

et al. 2014). The cgMLST approach utilizes computational methods to identify core

genes within genomes used for comparison, which expands the number of genes used

in the study. This approach provides higher discriminatory power for genetic studies,

unlike the conventional multilocus sequence typing approach using a few housekeeping

genes.

Type III Secreted Effectors

The type III secretion system encoded by the Hrp cluster in Xanthomonas

translocates secreted protein effectors into the plant hosts, and these determine the

pathogenicity and host range of the pathogen (Potnis et al. 2015). Approximately 45 groups of type III secreted effectors (T3Es) are associated with Xanthomonas (Potnis et al. 2015, Potnis et al. 2011, White et al. 2009). The T3Es secreted by Xanthomonas also moderate pathogen adaptation to specific host species and genotypes. Core effectors play essential roles at different stages of disease development. The core effectors XopD, XopL and XopN were reported to interfere pathogen associated molecular pattern (PAMP)-triggered immunity (PTI) responses in tomato plants (Potnis et al. 2015). The core effectors XopD, XopL and AvrBs3 mimic host proteins to interfere with host cellular processes (Üstün and Börnke 2014).

Characterization of effector profiles among X. perforans in Florida identified two

different phylogenetic groups of T4 X. perforans strains (Schwartz et al. 2015). A recent

acquisition of avrBsT effector was observed in the tomato pathogenic X. perforans

strains. AvrBsT is known to cause incompatible reaction against pepper cultivars

resulting in hypersensitive reactions (Kim et al. 2010). Mutations in avrBsT from the

Group 1 of race T4 X. perforans strains were unable to cause lesions in pepper, but the

strains of a group 2 were virulent on pepper (Schwartz et al. 2015). Interestingly, the

group 1 T4 X. perforans shares identical XopQ allele with race T3 X. perforans, while

the XopQ allele in second group was identical to T1 X. euvesicatoria strains. Based on

XopQ sequences, pepper pathogenic X. perforans has effectors similar to the group 2

T4 X. perforans, but lacks avrBsT. Thus, the role of avrBsT is likely associated with

bacterial fitness in tomato and a host-limiting factor in pepper.

A common form of plant resistance against pathogen is through gene for gene

interaction (Heath et al. 1990). Resistance (R) gene from plants recognizes the

avirulence (avr) gene from pathogens resulting in programmed cell death and limits the pathogen growth in the plants (Greenberg 1997). Resistance against avrBs2 gene is already available in transgenic tomato but mutations are already reported in avrBs2 effector that can break the resistance (Horvath et al. 2012). Thus assessing X. perforans effector profiles is relevant for population studies and provides insights for resistance breeding approaches.

Project Goals and Objectives

Bacterial diseases in Florida often cause significant economic losses due to the states sub-tropical environment that favors rapid disease development. Recent changes within the pathogen population have led to an increase in host range and resistance against bactericides. Bacterial taxonomy provides identification of organisms in a meaningful way that be used to make predictions about the organisms’ characterstics.

Reliable characterization of plant pathogenic bacteria is important for identification and for taxonomic studies (Staley et al. 2010). Bacterial taxonomy is dynamic and has its own limitation due to mutations, recombination events, and the identification of novel bacterial strains. The availability of new comparative tools based on whole genome sequencing will improve bacterial taxonomy.

The objective of this project will be to characterize tomato pathogenic

Pseudomonas and Xanthomonas strains isolated in Florida using sequence based approaches. Multilocus and whole genome based sequence analyses provide essential information regarding the phylogenetic relationship of Pseudomonas. Changes in the

Xanthomonas perforans population isolated from Florida will also be studied.

Subsequent comparative studies of type III effectors among pathogenic X. perforans

strains will also be conducted to identify effective targets for resistance breeding efforts

in tomato. Whole genomes from X. perforans strains will be used for population studies to identify genetic changes in the bacterial population over time.

The four specific objectives of this dissertation are to: I) Characterize

Pseudomonas cichorii strains isolated from tomato in Florida; II) Characterize a novel

Pseudomonas species also isolated from tomato in Florida; III) Identify candidate

targets for resistance breeding to Xanthomonas perforans in tomatoes; and IV) Develop

a core genome multi-locus sequence analysis method to effectively conduct

phylogenetic studies in Xanthomonas perforans.

CHAPTER 2 PHYLOGENETIC STUDY REVEALS A DISTINCT PHYLOGROUP OF PSEUDOMONAS CICHORII IN FLORIDA

Introduction

Among the pathogenic pseudomonads commonly associated with tomato, bacterial speck caused by Pseudomonas syringae pv. tomato, bacterial leaf blight

caused by P. viridiflava and pith necrosis caused by P. corrugata were reported from

tomato in Florida (Jones et al. 1984). Foliar disease symptoms caused by bacterial

speck consist of small, dark brown raised lesions with a yellow halo (LeBoeuf et al.

2005), whereas bacterial leaf blight symptoms consist of tissue necrosis (Jones et al.

1983, Jones et al. 1984). Regardless of symptomatology, these diseases are caused by

pseudomonads that favor cool, wet weather conditions.

In 2010 and 2012, tomato foliage and fruit exhibiting symptoms similar to

bacterial speck were observed during weather conditions uncharacteristic for P.

syringae pv. tomato or P. viridiflava. Foliar symptoms consisted of irregular lesions with

distinct yellow margins (Figure 2-1) and fruit exhibiting pinpoint sunken lesions. These

symptoms had not been observed previously on tomato in Florida. Preliminary

characterization identified the strains as P. cichorii, which is considered a warm weather

pathogen able to grow in conditions with high temperature and humidity (Jones et al.

1984), but had never been reported on tomato in Florida in the past.

Pseudomonas cichorii has been reported on a number of economically important

hosts. Pseudomonas cichorii strains were first reported in endive (Swingle 1925) and

later it was recovered from rots on head Lettuce in New York (Burkholder 1954). P.

cichorii is reported to have a wide host range (Stead et al. 2003) and are widely

distributed with reports in Belgium (Cottyn et al. 2009), Greece (Trantas et al. 2013)

Italy (Scortichini et al. 2002) and Turkey (Aysan et al. 2003). The pathogen was previously isolated in Florida from chrysanthemum, celery, geranium, and hibiscus

(Chase, 1986; Jones et al. 1983, Pernezny et al. 1994), and more recently from Stevia

rebudiana (Strayer et al. 2012) and Duranta erecta (Gumtow et al. 2013). Tomato as a

host for P. cichorii was first reported in 1974 in New Zealand (Wilkie and Dye 1974).

Symptoms of P. cichorii on tomato were described as being similar to late blight in

tomato (Wilkie and Dye 1974) and pith-necrosis like symptoms (Testen et al. 2015).

Along with tomato, plants from the Asteraceae family including, Lactuca sativa

(Lettuce), Cichorium endivia (Endive) and Helianthus annus (Sunflower) are regarded

as natural hosts of P. cichorii.

Based on phylogenetic analysis using gyrB and rpoD, P. cichorii groups with the

P. syringae pathogen complex (Yamamoto et al. 2000) and more recently was placed in

Phylogroup 11 based on multilocus sequence analysis of housekeeping genes (Berge

et al. 2014). Standard LOPAT tests can be used to differentiate P. cichorii strains from

other fluorescent Pseudomonads (Lelliott et al. 1966). Variation within P. cichorii strains

has been identified. Three distinct morphotypes, based on colony formation and

fluorescent pigmentation of P. cichorii were defined (Cottyn et al. 2009). Moreover, a

new genomovar of P. cichorii causing tomato pith necrosis was recently reported, based

on BOX-PCR and MLSA, indicating wide variation within multi-host pathogenic P.

cichorii strains (Trantas et al. 2013).

Although, P. cichorii strains were frequently reported in Florida on various crops,

the recent outbreaks in tomato have increased the necessity for in depth

characterization of the P. cichorii strains isolated from tomato and for a comparison with

strains from different hosts in Florida to determine if this was a recent introduction or perhaps an expansion of host range by endemic strains. We hypothesize that the P. cichorii strains isolated in Florida are similar and that the presence of hosts year round may have been crucial in establishing the pathogen on tomato. The purpose of this study was to characterize beyond the LOPAT tests the recently isolated Pseudomonas

strains isolated from tomato in Florida using pathogenicity tests and multilocus

sequence analysis. In addition, we studied genetic diversity of P. cichorii isolated from

various hosts in Florida over more than 30 years in order to clarify the genetic

relatedness of P. cichorii strains in comparison with strains identified in other areas of

the world.

Materials and Methods

Bacterial Strains and Isolation

Bacterial strains used in this study were isolated from various plant hosts in

Florida since the early 1980s (Table 2-1). The majority of the strains used in this study

were recently isolated from tomato. Based on the LOPAT tests the tomato stains were

most closely related to P. cichorii. Thus, the type strain of P. cichorii, PC22, and

representative P. cichorii strains previously reported from different hosts in Florida were

used in this study. The bacteria strains were plated on nutrient agar and on Kings B

medium (KB) to detect fluorescence (King et al. 1954). A total of 18 strains isolated from

tomato and 7 strains isolated from 7 different hosts in Florida were also characterized in

this study.

Pathogenicity Tests

Six-week-old plants of the susceptible tomato cultivar Bonnie Best, were spray

inoculated with 50 mL of bacterial suspension from a 24-h culture to evaluate

pathogenicity. Each strain was adjusted to approximately 108 CFU/ml for inoculation.

The plants were sealed in plastic bags and maintained in a growth chamber at 28 °C for

48 hours. Given that the type strain of P. cichorii was originally isolated from lettuce

(Mirik et al. 2011), romaine lettuce plants were also spray inoculated in the same manner. Following the appearance of disease symptoms after 48 hr, bacteria were isolated from lesions on KB to confirm Pseudomonas identity.

PCR Amplification and Gene Sequencing

Bacterial strains were grown overnight at 28 °C on nutrient agar. Four

housekeeping genes (gap1, gltA, gyrB and rpoD) and 16S rRNA gene were amplified

using their respective primers as described previously (Hwang et al. 2005 and Marchesi

et al. 1998). The partial coding regions of four housekeeping genes were amplified for

all the Pseudomonas strains used in this study. Based on the results from multilocus

gene sequences, only 3 strains isolated from tomato were sequenced for 16S rRNA

gene along with the P. cichorii strains isolated from other hosts for comparison to

reference strains. The resulting PCR products were sequenced at the Interdisciplinary

Center for Biotechnology Research at University of Florida using Sanger sequencing.

Pseudomonas strains are often reported with multiple copies of 16S rRNA gene

(Ueda et al. 1999). Multiple copies of 16S rRNA gene were also detected in Florida

tomato isolates that resulted in sequencing errors. The PCR products from strains with

multiple copies of 16S rRNA gene were purified using a Qiagen PCR purification kit

(Qiagen N.V., Hilden, Germany). Single copies for each of the amplified products were

then cloned into Escherichia coli using the pGEM-T plasmid vector (Promega

Corporation, Madison, WI). Multiple colonies of E. coli carrying 16S rRNA copies from

each individual P. cichorii strain were sequenced using specific primers to confirm the

16S rRNA sequence for each strain.

Phylogenetic Analysis

The resulting sequences from individual strains were compared with sequences

available from the National Center for Biotechnology Information (NCBI) and Plant

Associated and Environmental Microbe Database (PAMDB) to find the closest relatives.

Based on sequence similarity, sequences of closely related P. cichorii and P. syringae

strains were retrieved from the databases for phylogenetic analyses. The sequences for

respective genes were aligned individually using muscle tool in MEGA v. 6 software

(Tamura et al. 2013). Sequences from the four housekeeping genes were concatenated

to generate a combined phylogenetic tree. A maximum likelihood phylogenetic tree was

generated using the general time reversible model with gamma distributed invariant

sites within MEGA. Multiple MLSA schemes are proposed for characterizing

Pseudomonas. Although different housekeeping genes are used for characterizing

Pseudomonas in various MLSA schemes (Hwang et al. 2005, Mulet et al. 2008, Trantas

et al. 2013), gyrB and rpoD genes are included in more than one MLSA scheme. In

addition, Yamamoto et al. (2000) reported that significant phylogenetic information on

Pseudomonas can be etrieved using gyrB and rpoD gene sequences. Thus, along with

the MLSA using four housekeeping genes, Pseudomonas strains were phylogenetically

analyzed using concatenated sequences of gyrB and rpoD genes. Phylogenetic

analysis of 16S rRNA gene sequences was carried out using the same model used for

analyzing housekeeping genes in MEGA.

GenBank Accession Numbers

Genes sequenced during this study were submitted to NCBI database. The

accession numbers for each gene sequenced during this study are: gap1, KX673643 –

KX673668; gltA, KX673669 – KX673694; gyrB, KX673695 – KX673720, rpoD, KX67321

– KX673746 and 16S, KX670410 – KX670420.

Results

Isolation and Phenotypic characterization

Bacteria were isolated from symptomatic tissue, which was similar to bacterial

speck of tomato caused by Pseudomonas syringae pv. tomato (Figure 2-1). Unlike,

bacterial speck, the symptoms in this case included irregular leaf lesions with distinct

yellow margins along with fruit exhibiting sunken, pin-point necrotic lesions.

Interestingly, these symptoms were observed when temperatures in Central Florida

were unfavorable for bacterial speck development (Figure 2-1). All bacteria strains

isolated from tomato were fluorescent on King’s medium B. The strains were negative

for levan production, arginine dihydrolase activity and pectolytic activity, but positive for

oxidase activity and caused a hypersensitive reaction when infiltrated into tobacco

leaves. As the tomato strains were characterized as P. cichorii based on the LOPAT

assay (Lelliot et al. 1966), various P. cichorii strains previously reported from Florida

were also included in the study. Standard LOPAT assay was repeated to reconfirm the

previously reported strains as P. cichorii.

Pathogenicity Test

All Florida tomato strains were pathogenic on tomato after inoculation and

produced symptoms similar to natural infection (Figure 2-2). Interestingly, the P. cichorii

type strain PC22 was pathogenic on tomato. Furthermore, Florida strains isolated from

tomato were also pathogenic on lettuce producing disease symptoms 48 hr post

inoculation (Figure 2-2). Typical P. cichorii was re-isolated from symptomatic tissue in

high concentrations.

Among other P. cichorii strains from Florida, PC2, PC24, PC25, PC37, PC49 and

PPST 50936 isolated from Geranium, Hibiscus, Basil, Ficus lyrata, crown of thorns and

Stevia, respectively, were all pathogenic on tomato.

Genetic Characterization

Results from MLSA and 16S rRNA sequences were in agreement with LOPAT

characterization as the strains were closely related to P. cichorii. The Pseudomonas

strains shared similar sequences for housekeeping genes (Figure 2-3 and Figure 2-4)

and 16S gene (Figure 2-5) with type strain P. cichorii PC22. However, strains isolated in

Florida formed a distinct cluster even though they were collected from multiple hosts.

The strains isolated from tomato were identical to each other with one exception

(GEV417). The tomato strains varied from the type strain P. cichorii PC22 in gap1, gltA,

gyrB and rpoD by a total of 106 nucleotides out 2043. The unique tomato strain,

GEV417, differed by a total of 20 nucleotides from the other tomato strains. The P.

cichorii strains isolated from hosts other than tomato were within the same clade as the

other P. cichorii strains from tomato. Among the non-tomato strains from Florida, all

strains carried different allelic forms of housekeeping genes compared to tomato P.

cichorii strains and only PC25 shared sequence similarity with tomato strains in gap1

and gltA. Strain PC25 was isolated from basil and formed a distinct clade with rpoD

sequences. The sequence types of different genes for each strain used in this study are

listed in Table 2-1.

Concatenated sequences of two housekeeping genes, gyrB and rpoD, were used separately for phylogenetic analysis to compare the Florida P. cichorii strains to a large number of Pseudomonas strains characterized using various MLSA schemes (Figure 2-

3). P. cichorii strains from Florida formed a clade separate from the P. cichorii type strain. Two P. cichorii strains isolated from pumpkin in Tennessee also clustered with the Florida P. cichorii strains. Sequence comparison identified the P. cichorii strains

from this study varied from previously described morphotypes and genomovar of P.

cichorii.

Phylogenetic analysis of 16S rRNA sequences produced similar results to the

housekeeping genes (Figure 2-5). Two representative tomato strains differed at 4

nucleotides to the 16S gene from the P. cichorii type strain PC22. Strain GEV417

showed a single nucleotide difference from the other Florida tomato strains. Florida

strain PC4 isolated from chrysanthemum in 1984 had 16S sequence identical to the two

Florida tomato strains. Florida strains, PC25, PC27, PC37 and PC49 isolated in the

1980s and 90s from Basil, Ficus pandurata, Ficus lyrata and Euphorbia milii, respectively, had 16S sequences identical to the GEV417 strain from tomato. P. cichorii strain PC24 isolated from hibiscus had 16S rRNA sequence identical to P. cichorii strain

PPST 50936 previously reported from Stevia rebaudiana in Florida. 16S rRNA

sequences from two representative strains of P. cichorii isolated from pumpkin in

Tennessee were available in NCBI (Accession: KU174188 and KU174189) (Newberry et al. 2016). Phylogenetic analysis using 16S rRNA sequences clustered these strains and distinct from the type strain (Figure 2-5).

The geranium strain, PC2, was identified as P. cichorii by the LOPAT tests, but was more similar to Pseudomonas strains pathogenic on potato (Accession: AB512620) based on 16S rRNA sequences. The housekeeping genes also grouped PC2 with

Pseudomonas strain StFLB209 (Accession: AP014637) isolated from the potato phyllosphere (Someya et al. 2009).

Discussion

The bacterial strains isolated from symptomatic tomato in Florida were confirmed to be P. cichorii on the basis of standard LOPAT assay and phylogenetic analyses of housekeeping genes and 16S rRNA squences. Our study indicates the P. cichorii strains from Florida form a genetic subgroup distinct from the type strain and other reported P. cichorii strains. The strains of P. cichorii isolated over the last 30 years in

Florida all clustered together in the same group. Thus, it seems unlikely that the P. cichorii strains recently isolated from tomato in Florida represent a new introduction, but possibly are representative of P. cichorii already present in Florida that are pathogenic on tomato under favorable environmental conditions and host availability. Although, strains of P. cichorii were reported from number of hosts in Florida since the early 1980s

(Chase 1986), they were not genetically characterized until recently. P. cichorii was previously reported to cause pith necrosis disease of tomato in Greece, New Zealand and Turkey (Mirik et al. 2011, Trantas et al. 2013 and Wilkie and Dye 1974), but sequence data for these strains are not available.

In Florida, P. cichorii caused foliar symptoms on tomato with symptoms similar to bacterial speck disease caused by P. syringae pv. tomato. Angular lesions with distinct

yellow margins were observed during P. cichorii infection in tomato. Pathogenicity

results show that P. cichorii strains isolated from Florida were pathogenic on more than

one host. The P. cichorii strains isolated from tomato in Florida were pathogenic on lettuce. Likewise, P. cichorii strains isolated from various hosts in Florida, including the type strain from lettuce, were also pathogenic on tomato. The ability of P. cichorii to cause disease on multiple hosts suggests that the pathogen can spread rapidly to a number of economically important hosts.

Sequence comparison of P. cichorii strains clustered the Florida strains together into a closely related group. The tomato P. cichorii strains form a distinct cluster based

on MLSA and 16S rDNA sequences. Two allele types of 16S rDNA sequence were

detected among tomato strains. One allele was identical to the 16S rDNA of

Chrysanthemum strain PC4, while only tomato strain GEV417 carried the second 16S

rDNA allele type. Interestingly, this second 16S rDNA allele type was detected in other

P. cichorii strains isolated from hosts like Basil, Crown of Thorns, Hibiscus and two

Ficus species. Both 16S rDNA allele types varied from the type strain (Table 2-1).

Additionally, a P. cichorii strain PC2 isolated from Geranium formed a distant clade with

other Pseudomonas sp. isolated from the potato phyllosphere based on 16S

sequences. Similar results were observed from the housekeeping gene sequences.

Two allelic forms of housekeeping genes were observed in tomato pathogenic strains of

P. cichorii. The strain GEV417 was isolated in 2011 and other tomato strains were

isolated in 2012. As these strains were isolated in different years, the sequence

variation between these strains suggests that there could be multiple introduction

events for this pathogen to tomato in Florida. In addition to the Florida strains, two

representative strains of P. cichorii pathogenic on pumpkin isolated in Tennessee also

clustered with the Florida strains based on 16S rRNA and the housekeeping genes gyrB

and rpoD. All P. cichorii strains clustered together with the Florida tomato isolates except for strain PC2 isolated from Geranium. The strain PC2 is found similar to

Pseudomonas sp. StFLB209, isolated from potato, based on 16S rRNA and four

housekeeping genes. Whole genome sequence analysis showed this strain from potato,

StFLB209, was closely related to P. cichorii strain JBC1 isolated from Soybean

(Someya et al. 2009, Morohoshi et al. 2014). The sequence variations within the Florida

P. cichorii strains may be due to their association with different hosts.

Three different morphotypes and a distinct genomovar pathogenic on tomato are

reported within P. cichorii (Cottyn et al. 2009, Trantas et al. 2013). Florida P. cichorii

strains clustered separately from previously described P. cichorii strains. Phylogenetic

analyses using housekeeping genes suggest that the population in Florida represents a

distinct clade of P. cichorii. These strains span collections from the early 1980s until

2012. Florida P. cichorii strains are phylogenetically similar and their ability to cross-

infect hosts suggests that the same phylogenetic group of P. cichorii has been affecting

multiple hosts within Florida for over 30 years. This distinct phylogenetic population of

P. cichorii could be widespread in various hosts throughout the United States or

restricted to a particular environmental niche but further identification and

characterization of additional Pseudomonas strains would be necessary.

The ability of P. cichorii strains to infect multiple hosts may cause significant loss

in economically important crops in Florida. In addition to various isolations made in

Florida, the reported frequency of disease caused by P. cichorii has increased

substantially over the last decade from different geographical locations and from a

number of economically important hosts, including tomato (Myung et al. 2013, Mirik et

al. 2011 and Newberry et al. 2016). Since the pathogen has a wide host range, the increasing spread of P. cichorii may pose a potential threat to several agricultural industries in Florida. McCarter et al. 1983 reported that Pseudomonas strains could

spread through seeds and weed hosts. Presence of volunteer hosts in Florida fields

increases the pathogen’s potential to spread over different hosts including tomato.

In conclusion, a distinct phylogenetic group of P. cichorii was observed in Florida

and a phylogenetic sub-group specific to tomato was observed among P. cichorii strains

isolated from Florida. It seems likely that the P. cichorii strains already present in

various hosts in Florida were transferred to tomato in multiple occasions. Presence of

tomato fields in close proximity to other hosts of P. cichorii and favorable environmental

conditions in Florida may have triggered the change in host range. Conducive weather

conditions and the presence of multiple hosts throughout the year in Florida are ideal for

disease development and survival of this pathogen.

Table 2-1. List of P. cichorii strains used for this study. Strain Host Year Location 16S gap1 gltA gyrB rpoD Source GEV417 Tomato 2011 Florida 2* 2 2 2 2 This study GEV1093 Tomato 2012 Florida 3 3 3 3 3 This study GEV1105 Tomato 2012 Florida NS 3 3 3 3 This study GEV1112 Tomato 2012 Florida NS 3 3 3 3 This study GEV1124 Tomato 2012 Florida NS 3 3 3 3 This study GEV1127 Tomato 2012 Florida NS 3 3 3 3 This study GEV1129 Tomato 2012 Florida NS 3 3 3 3 This study GEV1141 Tomato 2012 Florida NS 3 3 3 3 This study GEV1144 Tomato 2012 Florida NS 3 3 3 3 This study GEV1149 Tomato 2012 Florida NS 3 3 3 3 This study GEV1151 Tomato 2012 Florida NS 3 3 3 3 This study GEV1154 Tomato 2012 Florida NS 3 3 3 3 This study GEV1158 Tomato 2012 Florida NS 3 3 3 3 This study GEV1164 Tomato 2012 Florida NS 3 3 3 3 This study GEV1169 Tomato 2012 Florida NS 3 3 3 3 This study GEV1171 Tomato 2012 Florida NS 3 3 3 3 This study GEV1173 Tomato 2012 Florida NS 3 3 3 3 This study GEV1178 Tomato 2012 Florida 3 3 3 3 3 This study PC2 Geranium 1985 Florida 5 6 8 9 11 This study PC4 Chrysanthemum 1984 Florida 3 5 5 5 5 This study PC22T Lettuce 1954 USA 1 1 1 1 1 This study PC24 Hibiscus 1984 Florida 4 4 4 4 4 This study PC25 Basil 1986 Florida 2 2 3 7 9 This study PC27 Ficus pandurata 1985 Florida 2 5 6 6 6 This study PC37 Ficus lyrata 1985 Florida 2 5 7 8 8 This study PC49 Crown of Thorns 1995 Florida 2 5 7 8 10 This study PPST Stevia 2012 Florida 4 NS# NS 7 7 Strayer et 50936 al. 2012 P. sp. Potato NA Japan 6 6 9 10 12 Morohoshi StFLB209 et al. 2014 P. cichorii Soybean 2008 S. 1 1 1 1 1 Ramkumar JBC1 Korea et al. 2015 P. cichorii Pumpkin 2014 Tennes 7 NS NS 11 13 Newberry TN-E4 see et al. 2016 P. cichorii Pumpkin 2014 Tennes 8 NS NS 11 13 Newberry 14- see et al. 2016 WTREC T Type strain * Numbers represent the allele type of a gene present in each strain. Type strain of P. cichroii PC22 was used as reference allele. # NS – Genes not sequenced

Figure 2-1. Bacterial disease symptom observed in tomato foliage caused by P. cichorii. (Photo courtesy: Dr. Gary E. Vallad)

Figure 2-2. Disease symptoms on tomato and lettuce hosts (48 hpi) inoculated with representative P. cichorii strains isolated from Florida along with the type strain on (A) Tomato and (B) Lettuce. Representative strains were: (i) GEV417 (ii) GEV1127 and (iii) P. cichorii PC22. (Photo courtesy of author)

Figure 2-3. Maximum likelihood phylogeny of Pseudomonas strains isolated from Florida using concatenated sequences of gyrB and rpoD genes. Host of isolation for strains isolated from Florida and Tennessee are provided next to the strain name (T: Tomato, P: Pumpkin, B: Basil, FP: Ficus pandurata, FLy: Ficus lyrata, CT: Crown of Thorns, S: Stevia, C: Chrysanthemum, H: Hibiscus and G: Geranium). Bootstrap values for each branch are expressed as percentage and scale bar represents substitutions per site.

Figure 2-4. Maximum likelihood phylogenetic distribution of Pseudomonas strains isolated from Florida using concatenated sequences of four housekeeping genes (gap1, gltA, gyrB and rpoD) compared with reference sequences downloaded from PAMDB database. Bootstrap values for each branch are expressed as percentage and scale bar represents substitutions per site.

Figure 2-5. Maximum likelihood phylogenetic distribution of representative Pseudomonas strains isolated from Florida tomato using 16S rRNA sequences compared with reference sequences. Bootstrap values for each branch are expressed as percentage and scale bar represents substitutions per site.

CHAPTER 3 PSEUDOMONAS FLORIDAE SP. NOV., A NOVEL BACTERIAL PATHOGEN ISOLATED FROM TOMATO

Introduction

The genus Pseudomonas is a large group of Gram-negative, rod-shaped bacteria that are motile and are highly diverse in nature. The bacterial members of the genus are widely prevalent in soil, water, plants and animals (Palleroni 1981; Anwar et al. 2016) and include numerous pathogenic species capable of causing disease in various hosts. Approximately, 150 species of Pseudomonas are validly published (Parte

2014). Although the taxonomy of Pseuodmonas has been described multiple times

(Parte 2014), speciation has traditionally been based on only a few phenotypic characteristics. DNA-DNA hybridization (DDH) is regarded as the gold standard for taxonomic classification (Bull and Koike 2015). Although DDH provides higher resolution for taxonomy, the approach is laborious and is not feasible to establish a central database (Cho and Tiedje 2001). The availability of DNA sequencing technologies and databases provided the foundation for sequence based taxonomy

(Gomila et al. 2015). Sequences from 16S rRNA and conserved housekeeping genes are frequently used for Pseudomonas characterization and classification. Several schemes are proposed for multilocus sequence analysis and characterization of

Pseudomonas strains (Bartoli et al. 2014; Berge et al. 2014; Hwang et al. 2005).

Recently, the whole genome sequences (WGS) have been used for comparative and taxonomic purposes. Approaches towards species differentiation in plant pathogenic

Pseudomonas utilizing whole genome sequences are underway (Marcelletti and

Scortichini 2014; Bull and Koike 2015). Multiple pairwise comparative measures like average nucleotide identity and genome-genome-distance calculation are implemented

for species differentiation using WGS (Auch et al. 2010; Konstantinidis et al. 2006, Chan

et al. 2012, von Neubeck et al. 2016). Unlike, conventional DDH approaches, WGS can

be used to compare a wide range of bacteria from multiple sources by pairwise

comparison and to establish a central database for future comparison.

Pseudomonas syringae represents a complex group of strains that cause various

diseases in a wide range of plant hosts (Lamichhane et al. 2015). In the case of plant

pathogenic pseudomonads; infra-subspecific taxonomy has been described by using

pathovar designations based on host range. A total of 13 phylogroups and 8

genomospecies of P. syringae are described based on DNA relatedness (Gardan et al.

1999) and multilocus sequence analysis (Bull et al. 2011; Berge et al. 2014).

Recent characterizations and population studies of plant pathogenic

Pseudomonas species are based on multilocus sequence analysis (Bull and Koike

2015). Strains of P. syringae, P. viridiflava and P. corrugata were previously reported in

Florida but P. cichorii strains are only reported recently from tomato in Florida (Timilsina

et al. in preparation, McCarter et al. 1983, Jones et al. 1986). An outbreak of

Pseudomonas was observed in Florida tomato in 2010/11. Symptoms similar to disease

caused by Pseudomonas syringae were observed. The lesions were raised and dark

brown to black in color with a yellow halo. Among the strains isolated during this

outbreak, atypical LOPAT profile was observed in 8 Pseudomonas strains. The novel

strains produced yellow mucoid biofilm in 5% sucrose media. Pseudomonas strains with

a similar atypical reaction in hypersucrose medium were previously isolated from

common bean, kiwi and lettuce in Spain (Gonzalez et al. 2003).

Further comparisons using representative16S rRNA sequences differentiated these novel strains from previously reported Pseudomonas species. However, these

novel strains from Florida were closely related to P. viridiflava but varied significantly

from all known Pseudomonas at the species level. The bacterial strains were shown to

be pathogenic on tomato and the strains were characterized using a number of

phenotypic assays and genomic comparisons. We present data indicating that these

strains belong to a new Pseudomonas species that is most closely related to P.

viridiflava.

Materials and Methods

Bacterial Strains and Preliminary Characterization

Eight atypical bacterial strains of Pseudomonas were isolated from tomato in

Florida (Table 3-1). The bacterial strains were fluorescent on King’s Medium B (KMB)

(King et al. 1954).

Biochemical and Physiological Tests.

All bacterial strains were characterized using standard LOPAT procedures

(Lelliott et al. 1966). Based on LOPAT results, the strains were further grown in Ayers’

minimum media (Deshmukh 1997) with 1% sucrose to determine sucrose utilization.

Due to the atypical biofilm production during the levan assay, Concanavalin A

(ConA) assay was used to study the biofilm composition. ConA is a carbohydrate

binding protein that binds to levan and similar carbohydrate compounds (Laue et al.

2006). Fluorescent FITC-labelled ConA was mixed with the biofilm produced during

levan assay and observed for fluorescence. A levan positive P. syringae strain and

levan negative P. viridiflava and P. cichorii were used as controls.

Pathogenicity Assay

The strains were inoculated into tomato hosts to fulfill Koch’s postulates. One- day-old bacterial cultures grown in NA were suspended in 50 mL of sterile tap water and the bacteria concentration was adjusted to approximately 108 CFU/mL. The suspension

was sprayed until run off on susceptible tomato plants (cv. Bonnie Best). Control plants

were sprayed with sterile tap water. Three plants per strain were spray inoculated and

bagged for 48 hours at 28 °C. The plants were observed for symptom development and

the assay was repeated twice. Bacteria were isolated from inoculated plants and were

streaked on KMB to evaluate fluorescence activity. The standard LOPAT tests were

conducted on the reisolated strains to confirm the isolated bacterium behaved like the

original strains.

Gene Sequencing and Sequence Analysis

Genomic DNA was extracted from the eight strains using a modified CTAB

method (Chen and Kuo 1993). The 16S rDNA gene was amplified using Pseudomonas

specific primers 63f and 1387r (Marchesi et al. 1998). Initially, only three representative

strains were amplified. The resulting products were sequenced using Sanger

sequencing at the Interdisciplinary Center for Biotechnology Research (ICBR) at

University of Florida. Consensus sequences were derived based on forward and

reverse sequence reads. The 16S rDNA gene sequences were compared with

sequences within the National Center for Biotechnology Information (NCBI) database.

Phylogenetic analysis of 16S rDNA gene sequences was conducted using MEGA v.6

(Tamura et al. 2013) after multiple sequence alignments using muscle tool (Edgar 2004)

available within MEGA. A maximum likelihood phylogenetic tree was constructed using

the general time reversible model with gamma distributed invariant sites and the topology was estimated using 1000 bootstrap replicates.

To further elucidate the phylogenetic relationship; four housekeeping genes gap1, gyrB, gltA and rpoD were sequenced from all eight strains along with the type strain of P. viridiflava ATCC 13223 using respective primers described by Hwang et al.

(2005). Reference sequences were extracted from NCBI and Plant associated and environmental microbes database (PAMDB) (www.pamdb.org) based on BLAST sequence similarity. Phylogenetic analysis was conducted as described previously using single and concatenated gene sequences in MEGA v.6.

Additionally, a representative strain, GEV388, was phylogenetically compared with reference strains representing 13 phylogroups of the P. syringae complex using cts, rpoD, gap1 and gyrB gene sequences (Berge et al. 2014) available in the Plant

Associated and Environmental Microbes Database.

Biolog Assay and Fatty Acid Profiles

Based on the preliminary results, the atypical strains likely grouped as a novel species group of Pseudomonas genus. Thus, all strains of this novel Pseudomonas

group were characterized phenotypically using Biolog GENIII Microplate System

(Hayward, CA) to identify the closest relative of Pseudomonas with similar metabolic

reactions. Overnight bacterial cultures grown in NA were suspended in Biolog

inoculation fluid at recommended concentration and 100 µl of solution was suspended

in each well of the GEN III MicroPlates and incubated at 28 °C for 48 hours. Metabolic

reactions were read using a Biolog plate reader.

The strains of the novel Pseudomonas group were sent to Microbial Identification

(MIDI) labs (Newark, DE) to determine the cellular fatty acid composition.

Whole Genome Sequence Comparison

Along with single and multilocus sequence analyses, the whole genome of a representative strain GEV388 was sequenced. A genomic library was prepared using the Nextera DNA library prepare kit and was sequenced using an Illumina MiSeq platform. The raw sequences were assembled using the de novo genome assembly method in CLC Genomics workbench v. 5.0 (CLC Bio-Qiagen, Aarhus, Denmark). The draft genome was uploaded to Integrated microbial genomes and microbiome samples- joint genome institute (IMG/M-JGI) platform for genome annotation. The annotated genome characteristics were compared with other closely related Pseudomonas species.

The draft genome of GEV388 was used to calculate the average nucleotide identity (ANI) and genome distance by computing in silico DNA-DNA hybridization with several reference strains. For comparative purposes, pairwise ANI between GEV388 and all Pseudomonas genomes publicly available in the IMG database was calculated using the IMG/M-JGI web tool. Additionally, pairwise ANI based on BLAST (ANIb) and mummer (ANIm) was calculated with representative Pseudomonas species using jSpecies v1.2.1 (Richter and Rosselló-Móra 2009). A web-based tool, genome-genome distance calculator v2.1 was used to compute the in silico DNA-DNA hybridization between the novel Pseudomonas strain GEV388 and other reference strains (Meier-

Kolthoff et al. 2013).

Results

Bacterial Characterization

The bacteria isolated from tomato were fluorescent on KMB suggesting the strains belonged to Pseudomonas. All eight strains resulted in atypical LOPAT reactions

(Table 3-1). The strains were negative for arginine dihydrolase and oxidase reaction but positive for potato rot and tobacco hypersensitivity. However, the strains produced atypical yellowish mucoid growth on a hyper sucrose medium used for testing levan

production. Interestingly, the bacteria did not produce any mucoid growth when plated in

minimum media with 1% sucrose similar to P. viridiflava. However, P. syringae strains

produced levan like biofilm when streaked in minimum media with 1% sucrose.

Interestingly, ConA was able to bind with the mucoid biofilm produced by atypcial

Pseudomonas strains during levan assay and fluorescence was observed similar to a

levan positive Pseudomonas syringae strain GEV1421 (Figure 3-2).

Pathogenicity Assay

All atypical bacteria strains were inoculated on tomato plants and found to be

pathogenic on this host. Symptoms similar to natural infections were detected 48 hours

post inoculation. Using the LOPAT tests the reisolated strains were confirmed to behave

as the atypical strains.

Sequence Analysis

Sequence results from 16S rRNA genes and housekeeping genes showed

significant differences between the novel Pseudomonas strains and reference

Pseudomonas strains. The 16S rRNA sequences from 3 strains were identical to each

other (Figure 3-3). These sequences were compared with publicly available sequences

in the National Center for Biotechnology Information (NCBI). The sequence identities

between these novel Pseudomonas strains with the type strain of P. viridiflava ATCC

13223 (NCBI Accession number: NR_114482), the levan positive P. viridiflava strains

LPPA 144 (NCBI accession number AY180968.1) and LPPA 366 (AM182934.1)

reported from Spain were 98.81%, 98.89% and 99.05% respectively.

Likewise, all eight atypical Pseudomonas strains had identical housekeeping gene sequences. The strains were phylogenetically distinct from previously described

Pseudomonas species including P. viridiflava, P. syringae pv. syringae, and P. syringae pv. tomato (Figure 3-4). Similar to the 16S rRNA sequences, the housekeeping genes were also closely related to P. viridiflava strains. The novel Pseudomonas strains varied from type strain of P. viridiflava by 17, 21, 25 and 37 nucleotides in gap1, gltA, rpoD and gyrB respectively. Berge et al. (2014) utilized housekeeping gene sequences to characterize the plant pathogenic P. syringae complex into 13 different phylogroups. A representative novel Pseudomonas strain from this study, GEV388, was compared with reference strains representing 13 phylogroups of the P. syringae complex using the

Morris MLST schema available in PAMDB. Phylogenetic analysis showed that the representative Pseudomonas strain do not belong to any of the predefined phylogroups of Pseudomonas (Figure 3-5). However, the strain was phylogenetically closely related to phylogroup 8 of P. viridiflava.

Biolog Assay and FAME-ID

Results from Biolog substrate utilization suggested the bacterial strains were closely related to P. viridiflava but differed from P. viridiflava and P. syringae in a number of reactions (Table 3-1). The novel Pseudomonas strains utilized α-D-glucose,

D-mannose, D-fructose, D-galactose, D-sorbitol, D-mannitol, D-arabitol, myo-inositol, D- galacturonic acid, L-galactonic acid lactone, D-gluconic acid and D-glucoronic acid and grew at pH 5. The bacterial strains did not react with glycerol, glucose-6-PO4 and fructose-6-PO4. As opposed to reaction observed during the levan assay and consistent with the reaction in Ayers’ minimum medium with 1% sucrose, none of the strains reacted with sucrose in the Biolog assay.

Results from fatty acid profiles for the atypical Pseudomonas strains indicated

the strains were closely related to averaged reaction results from P. viridiflava and P.

syringae pv. tagetis (Table 3-2). The majority fatty acids of all strains were summed

feature 3 (C16:1w7c and/or C16:1w6c ) and summed feature 8 (C18:1w7c and/or

C18:1w6c).

Whole Genome Sequence Analysis

The raw sequences for GEV388 were assembled using the de novo genome assembly method in CLC Genomics workbench v.5.0 (CLC Bio-Qiagen, Aarhus,

Denmark). The assembly yielded a genome size of 6,103,769 base pairs with 17.72x coverage. The GC content was 59.2%; within the range observed for Pseudomonas species (Palleroni 2005) (Table 3-3).

Collectively, the highest ANI of GEV388 was detected with the P. viridiflava species group. Based on current classification, P. viridiflava is grouped in phylogroups 7 and 8 of plant pathogenic P. syringae complex. The ANIs between GEV388 and a

representative strain of phylogroup 7, P. viridiflava UASW0038, and phylogroup 8, P.

viridiflava ICMP2848, were both 86.6% (Table 3-4). The value is significantly lower than

95%, which is the minimum threshold value required for the strains to be considered as

same species (Richter and Rosselló-Móra, 2009, Rodrigues-R and Konstantinidis,

2014). In silico DNA-DNA hybridization of GEV388 was estimated between 31-64% and

26.5-41.2% with P. viridiflava and P. syringae pv. tomato respectively. In both cases,

the DNA hybridization values were significantly below the 70% threshold value

recommended for species description.

Discussion

In this study, we confirmed that the bacterial strains isolated from symptomatic tomato belong to a novel Pseudomonas species. Several phenotypic tests including

utilization of carbon sources, LOPAT tests, and fatty acid analyses, and genotypic

approaches including phylogenetic analysis of housekeeping genes and whole genome

sequence analysis were implemented for specific characterization of atypical

Pseudomonas strains. The LOPAT tests revealed these strains to be rather unique.

Strains with similar LOPAT profile were isolated in Spain (Gonzalez et al. 2003) given

that they produced mucoid growth on hyper-sucrose medium and were strongly

pectolytic in the potato soft rot assay.

We observe unusual production of what has been considered levan when the

tomato pseudomonad strains were tested for levan production. They produced copious

amounts of a viscous yellowish polysaccharide on nutrient agar amended with 5%

sucrose unlike a typical P. syringae strain that forms white, mucoid growth. The reaction

of ConA suggested the presence of levan or levan-like carbohydrate produced by these

tomato pseudomonad strains. Surprisingly, the bacteria did not produce similar biofilm

when grown in minimum media with 1% sucrose. Results from Biolog GenIII microplate

assay also showed that the strains did not react with sucrose. Thus, our results suggest

that the bacterial strains are levan negative with a LOPAT profile similar to P. viridiflava

but produce a levan-like yellowish biofilm in the presence of nutrient agar and sucrose.

Pseudomonas strains producing similar yellowish biofilm in high sucrose medium were

reported from Spain and were characterized as P. viridiflava based on 16S rRNA

sequence identity. Additional characterization based on utilization of metabolites using

Biolog GENIII microplates and fatty acid profiling also determined the strains were

closely related to P. viridiflava but could not be associated on the species level as in both cases, metabolic reactions and fatty acid profiles of atypical strains varied significantly from P. viridiflava.

The atypical Pseudomonas strains were phylogenetically close to P. viridiflava

strains but formed a separate phylogroup based on both 16S rRNA and housekeeping

gene sequences supported by a strong bootstrap value. Pseudomonas is the largest

gram-negative genus with 144 valid species (Gomila et al. 2015, Parte 2014), and

unique phenotypes are reported in various species (Gonzalez et al. 2003, Paulsen et al.

2005). But, unlike the atypical strains isolated from Spain that grouped with P. viridiflava

based on 16S sequences (Gonzalez et al. 2003), these tomato pseudomonad strains

from Florida varied based on multiple gene sequences from other Pseudomonas

species including P. viridiflava. Multilocus sequence analysis using four housekeeping

genes further differentiated the tomato Pseudomonas strains from P. viridiflava.

Currently, strains of P. viridiflava encompass the phylogroup 7 and 8 of P. syringae

complex. Based on current phylogroup description, GEV388 appears to be most closely

related to members of phylogroup 8, however it is phylogenetically distinct.

Based on these phenotypic and genetic results, these tomato pseudomonad

strains fulfill the criteria to be categorized as a new bacterial species. In addition to the

single and multilocus sequence comparison, the whole genome of a representative

strain, GEV388, was sequenced. ANIb was used to compare the genomic sequence of

the representative strain, GEV388, with genomes of other pseudomonads, including

sequenced P. viridiflava strains. Strain GEV388 had less than 87% ANI with all of the

sequenced pseudomonads with the highest ANI occurring with P. viridflava strains

(~86.6% nucleotide identity). The draft genome assembly of the strain was used to calculate ANI based on both blast and mummer. A strain of P. asturiensis LMG 26898 was identified as another closest relative with 86.11% pairwise nucleotide identity. The species, P. asturiensis sp. nov., was recently proposed as a new pathogenic

Pseudomonas species isolated from soybean and weeds in Spain (González et al.

2013). Based on DNA-hybridization, P. asturiensis was also reported as a close relative of P. viridiflava species. Interestingly, the tomato pseudomonad strains varied significantly from both species group. Along with ANI, in-silico DNA-hybridization values clearly determined the strains belong to a novel Pseudomonas species.

Our results strongly support that these tomato pseudomonads represent a novel

Pseudomonas species pathogenic on tomato. A new species, Pseudomonas floridae sp. nov. is proposed for these strains and that strain GEV388 is designated as the type strain.

Table 3-1. Phenotypic characteristics of P. floridae sp. nov. GEV388T and closely related plant pathogenic Pseudomonas species. Biolog Assay Lev Oxid Potato Argin Tobacco D- D- L- D- Strains Growth D- D- Glyc Trolean Sucro an ase rot ine HR Sor Manni Galacto Treh at pH 5 Fucose arabitol erol domycin se bitol tol nic alose GEV388 - - + - + + + + + w + + + - - GEV358 - - + - + + + + + w + + + - - GEV359 - - + - + + + + + w + + + - - GEV364 - - + - + + + + + w + + + - - GEV366 - - + - + + + + + w + + + - - GEV380 - - + - + + + + + w + + + - - GEV382 - - + - + + + + + w + + + - - GEV392 - - + - + + + + + w + + + - - P.

51 viridiflava - - + - + + + + - + + + - - -

GEV354 P. + - - - + + - - - w - w - - + syringae

Table 3-2. Cellular fatty acid composition of eight strains of P. floridae sp. nov., and its closest relative based on fatty acid composition. Library entry profile of the average value of fatty acid profiles for P. viridiflava and P. syringae pv. tagetis was provided by MIDI labs. Index Fatty acid Pseudomonas strains# 1 2 3 4 5 6 7 8 9 10 10 C10:0 3-OH 3.63 3.72 3.48 2.84 2.83 2.97 3.04 2.92 3.04 2.41 14 C12:0 4.73 4.92 4.59 4.73 4.64 4.75 4.52 4.72 5.10 4.6 23 C12:0 2-OH 2.83 2.69 2.53 2.48 2.48 2.56 2.54 2.5 2.63 2.95 26 C12:0 3-OH 4.20 4.15 3.78 3.98 3.88 4.00 3.94 3.98 4.3 3.73 53 C16:0 28.12 26.26 27.2 26.74 27.34 26.91 28.06 26.5 25.03 27.12 80 C18:0 0.61 0.81 0.87 0.84 0.88 0.88 0.91 0.90 1.26 1.07 81 C18:1w7c 11- 0.83 0.92 0.90 0.79 0.86 0.80 0.96 0.77 0.45 0.85 methyl 103 Summed 36.98 36.10 35.62 36.16 35.87 35.81 35.40 35.88 35.85 37.35 Feature 3* 108 Summed 16.46 18.68 19.14 20.08 19.84 19.97 18.68 20.56 21.47 18.92 Feature 8 #Strains: 1, P. floridae GEV388T; 2, P. floridae GEV358; 3, P. floridae GEV359; 4, P. floridae GEV364; 5, P. floridae GEV366; 6, P. floridae GEV380; 7, P. floridae GEV382; 8, P. floridae GEV392; Library entry profile of: 9, P. viridiflava; 10, P. syringae pv. tagetis are listed in the following table. *Summed feature represents groups of two or more fatty acids that could not be separated by the Microbial identification system. Summed feature 3, C16:1w7c and/or C16:1w6c; Summed feature 8, C18:1w7c and/or C18:1w6c.

Table 3-3. Genome properties of P. floridae sp. nov strain GEV388 compared with selected reference strains Properties GEV388T P. viridiflava P. syringae ICMP 2848 DSM 10604 Size (Mb) 6,056,958 5,902,031 6,072,447 GC content (%) 59.17 59.37 58.96 Number of genes 5728 5399 5329 Number of coding sequences 5614 5289 5207

Table 3-4. Pairwise ANIb using P. floridae sp. nov. strain GEV388 and representative strains from different phylogroups of P. syringae

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 1 - 2 84 - 3 83.73 85.93 - 4 82.97 83.05 82.72 - 5 83.8 86.49 89.23 83.09 - 6 83.96 86.09 88.34 83.17 89.15 - 7 84.1 86.28 88.47 83.31 89.28 93.82 - 8 86.16 84.02 83.72 83.29 84.08 83.95 84.1 - 9 83.92 86.06 88.24 83.15 89.07 94.05 94.68 83.98 - 10 83.75 89.16 85.7 82.94 86.18 86.41 86.55 83.86 86.98 - 11 84.07 87.24 86.69 83.24 87.68 86.84 86.94 84.18 86.78 86.78 - 12 86.58 84.09 83.75 83.28 84.09 84.08 84.2 87.31 84.04 83.96 84.18 - 13 83.98 89.59 85.92 83.19 86.43 86.52 86.66 84.06 86.46 94.11 86.93 84.1 - 14 83.51 86.64 85.63 82.84 86.23 85.8 86 83.55 85.66 86.28 86.6 83.62 86.4 - 15 84.22 87.15 86.65 83.2 87.51 86.81 86.9 84.13 86.71 86.68 95.86 84.14 86.86 86.56 - 54 16 83.66 86.16 97.14 82.85 89.4 88.31 88.5 83.7 88.27 85.8 86.69 83.72 85.92 85.81 86.68 - 17 83.83 86.2 89.26 83.08 97.56 89.02 89.16 83.89 88.9 86.28 87.58 83.95 86.34 85.93 87.48 89.23 - 18 81.81 81.89 81.64 82.21 81.99 81.94 82.01 81.93 81.94 82.17 81.95 81.93 81.94 81.62 81.92 81.61 81.98 - 19 83.99 87.26 86.73 83.21 87.62 86.79 86.9 84.14 86.69 86.78 96.54 84.13 86.91 86.6 95.59 86.73 87.61 81.94 - 20 86.59 84.14 83.75 83.29 84.14 84.08 84.18 87.33 84.02 84.03 84.23 96.26 84.14 83.63 84.19 83.73 84.08 81.93 84.14 - 21 86.56 84.12 83.74 83.28 84.11 84.09 84.19 87.33 84.02 84.01 84.21 96.28 84.12 83.61 84.16 83.73 84.03 81.89 84.13 99.52 - Strains: 1. P. floridae GEV388; 2. P. canabina ICMP 2823 (5); 3. P. caricapapayae ICMP 2855 (6); 4. P. cichorii JBC1 (11); 5. P. savastanoi NCPPB 3335 (3); 6, P. syringae 642 (2); 7, P. syringae aceris M302273 (2); 8, P. syingae CC1524 (9); 9, P. syringae CC1543 (2); 10, P. syringae CC1557 (10); 11, P. syringae CC1559 (1); 12, P. syringae CC1582 (7); 13, P. syringae CC1583 (10); 14, P. syringae CC1629 (4); 15, P. syringae CC1630 (1); 16, P. syringae pv. helianthi ICMP4531 (6); 17, P. syringae pv. phaseolicola 1448a (3); 18, P. syringae UB246 (13); 19, P. syringae USA007 (1); 20, P. viridiflava UASW0038 (7); 21, P. viridiflava ICMP2848 (8). Phylogroups of each strain are indicated in parenthesis.

Figure 3-1. Foliar symptoms observed in tomato caused by the novel Pseudomonas strains. (Photo courtesy: Dr. Gary E. Vallad)

Figure 3-2 Specific binding of FITC labeled Concanavalin-A with levan as observed in (A) GEV388 (B) Pseudomonas syringae pv. tomato GEV1421 (C) Pseudomonas viridiflava ATCC 13223T and (D) Pseudomonas cichorii GEV417. (Photo courtesy of author)

Figure 3-3. Maximum likelihood phylogenetic analysis using 16S rRNA sequences. Representative sequences from the Pseudomonas floridae strains were compared with closely related Pseudomonas species. Atypical strains of P. viridiflava isolated in Spain (LPPA 366 and LPPA 139) grouped together with typical P. viridiflava strains, while the atypical strains from Florida formed a different cluster that varied from any previously identified Pseudomonas species.

Figure 3-4. Maximum likelihood phylogenetic tree using concatenated sequences from four housekeeping genes: gap1, gyrB, gltA and fusA. Sequences from the Pseudomonas floridae strains were compared with typical Pseudomonas viridiflava and P. syringae strains. Only P. viridiflava and P. syringae pv. tomato were reported from tomato in Florida previously. Scale bar indicates nucleotide substitutions per site and bootstrap values are indicated in percentages.

Figure 3-5. Maximum likelihood phylogenetic comparison using concatenated sequences from four housekeeping genes: cts, gyrB, gap1 and rpoD. Sequences from the Pseudomonas floridae sp. nov. strain GEV388 was compared with strains representing different phylogroups (only phylogroups 7,8 and 9 are shown in the picture). GEV 388 formed a branch between but separate from phylogroups 8 and 9 indicating the strain is different from previously known phylogroups of Pseudomonas syringae complex.

CHAPTER 4 ANALYSIS OF SEQUENCED GENOMES OF XANTHOMONAS PERFORANS IDENTIFIES CANDIDATE TARGETS FOR RESISTANCE BREEDING IN TOMATO

Introduction

Bacterial spot of tomato and pepper is caused by four distinct species within the genus Xanthomonas: X. vesicatoria, X. euvesicatoria, X. perforans and X. gardneri

(Jones et al. 2004). The disease is found in all areas where tomato and pepper are grown and is most severe in areas with tropical and sub-tropical climates including the southeastern United States. High temperatures, humidity and rainfall favor disease development. The disease can cause yield losses of up to 50% in tomato (Scott et al.

1989). Current disease management strategies include resistant cultivars, use of

disease-free seeds and transplants, and the preventative application of antibiotics,

copper-based bactericides, and the plant defense elicitor acibenzolar-S-methyl (Huang et al. 2012). Disease management has been challenged by changes in the pathogen population, including increased resistance to bactericides and shifts to different

Xanthomonas races and species.

Four tomato races have been described based on the presence of specific effector genes that interact with the corresponding resistance genes in tomato cultivars

(Stall et al. 2009). Strains that elicit a hypersensitive response (HR) only in the tomato differential genotype Hawaii 7998, are designated as tomato race 1 (T1). Strains that produce a susceptible reaction in all of the tomato differentials are designated race 2

(T2). All T1 strains to date have been X. euvesicatoria (Stall et al. 2009). Initially, all T2

strains were X. vesicatoria (Stall et al. 2009), however, X. gardneri strains were

determined to also behave as race 2 (J.B. Jones unpublished data). A third pathogenic

race, identified in the early 1990s, was designated as race 3 (T3). These strains were

placed in a new species, X. perforans, which is closely related to but distinct from X. euvesicatoria. X. perforans T3 strains cause disease on Hawaii 7998, like T2 strains, but carry the avrXv3 effector gene that elicits an incompatible reaction on tomato genotypes with the corresponding Xv3 resistance gene. Later, T3 strains were determined to contain another avirulence gene, xopJ4 (avrXv4), which elicits a hypersensitive reaction in Solanum pennellii LA716 and other tomato genotypes with the RxopJ4 (Xv4) resistant locus (Astua-Monge et al. 2000, Sharlach et al. 2013; Stall et al. 2009). X. perforans strains with xopJ4, but lacking a functional avrXv3, were later identified and designated as race 4 (T4). Surveys conducted in Florida in 2006 and

2012 showed a shift from T3 to T4, with 75% of the collected strains identified as T4 in

2006 (Horvath et al. 2012) and 100% T4 by 2012 (Timilsina et al. 2015).

Most Xanthomonas species use the type III secretion system to transfer pathogen effectors into their hosts. The acquisition or mutation of various effectors governs Xanthomonas host range and virulence (Mhedbi-Hajri et al. 2013).

Approximately forty-five type III effector (T3E) groups are associated with Xanthomonas

(Potnis et al. 2015, White et al. 2009). T3Es are known or hypothesized to facilitate disease. Effectors that are perceived by host resistance proteins (R genes) induce a hypersensitive reaction (Nimchuk et al. 2003) that confers race-specific resistance (Stall et al. 2009).

The durability of plant resistance, regardless of how it was introduced, depends on the ability of the pathogen to evolve over time. For bacterial pathogens, durability depends on rates of mutation, the ability to horizontally exchange genetic material, the overall level of selection pressure, and the relative fitness of virulent strains (Stall et al.

2009). Effectors that are conserved in pathogen populations can be evaluated as

potential targets for resistance breeding. Resistance breeding strategies that target

evolutionarily conserved effectors coupled with the pyramiding of R genes could lead to

durable resistance against pathogen populations.

The Bs2 resistance gene recognizes the core effector AvrBs2, which is present in

all four Xanthomonas species that cause bacterial spot of tomato (Horvath et al. 2012,

Potnis et al. 2011). However, this pepper gene is not currently available in commercial

tomato cultivars. A transgenic approach was used to develop resistant tomatoes

expressing the Bs2 transgene (Tai et al. 1999). The effectiveness of Bs2 in commercial

peppers can be greatly diminished by the appearance of mutations within avrBs2, and

was observed in X. euvesicatoria populations (Swords et al. 1996, Gassmann et al.

2000). In order to improve the durability of Bs2 resistance in tomato, additional effectors should be targeted for pyramiding of resistance genes, line mixing, or alternating resistance lines (McDonald and Linde 2002). Resistance targeting effector AvrBsT, has been reported in Solanum lycopersicoides, but the resistance gene has not been identified and introgressed into tomato (Wang 1992). Likewise, resistance against the

XopJ4 effector in X. perforans has been mapped to a 4.2-Mb cluster in S. pennellii

(Sharlach et al. 2013). AvrBsT and XopJ4 are not considered core effectors because they are not present in all four bacterial spot species (Potnis et al. 2011). However,

because resistance against these effectors is known, they are potential targets.

Whole genome sequence analyses of field strains of the four bacterial spot

species revealed a high degree of variation in T3Es within species (Schwartz et al.

2015). Even among the 11 core effectors, allelic variation was observed in XopAD,

XopD and XopQ in T4 X. perforans strains from Florida. Previously, multilocus analysis revealed recombination between strains of X. perforans and X. euvesicatoria that generated two different genetic groups of T4 X. perforans strains in Florida (Timilsina et al. 2015). Strains designated as Group 1 were similar to the taxonomic type strain for X. perforans, strain 91-118, a T3 strain collected in Florida in 1991; while Group 2 strains had sequence similarities with X. euvesicatoria (Timilsina et al. 2015). This pattern was also observed for the T3E xopQ, such that X. perforans strains with the X. euvesicatoria gapA sequence also had the X. euvesicatoria xopQ sequence (Schwartz et al. 2015,

Timilsina et al. 2015). The specific molecular function of XopQ in pathogenicity is unknown (Teper et al. 2014). Subsequent whole genome comparisons of T4 strains identified two subgroups within Group 1, Group 1A and Group 1B (Schwartz et al.

2015). In addition to allelic variation in core genes, the majority of T4 X. perforans strains collected in 2006 and 2012 contained avrBsT. The X. perforans type strain 91-

118, did not have avrBsT. This effector, when expressed in X. perforans and translocated into pepper mesophyll cells, elicits a hypersensitive response, while when expressed in X. vesicatoria and translocated into tomato mesophyll cells this effector protein suppresses defense responses and increases virulence (Kim et al. 2010). A

Group 2 T4 strain of X. perforans from bell pepper was isolated that lacked AvrBsT

(Schwartz et al. 2015). When avrBsT was mutated in Group 2 strains originally isolated from tomato, they also caused lesions in pepper. However, mutation of avrBsT in Group

1A X. perforans strains did not confer pathogenicity in pepper, suggesting that host range of X. perforans strains is determined in part by genetic background (Schwartz et al. 2015). Both XopQ and AvrBsT appear to have a role in host range, because double

deletion of these two effectors also allowed host expansion of X. perforans to Nicotiana benthamiana (Schwartz et al. 2015).

The long-term effectiveness of a R-gene within a breeding program is generally dependent on the presence and fitness value of the corresponding effector gene within a defined pathogen population. However, in the case of tomato in Florida, unforeseen changes in the pathogen population due to the introduction of exotic species, mutations in effector genes, and even recombination among Xanthomonas species have necessitated the re-evaluation of resistance breeding strategies. The future of resistance breeding in tomato may fail if we do not understand the drivers of past changes within the pathogen population. The genomic variation found in Florida T4 X. perforans strains raised questions regarding variation in T3 X. perforans before and during the race shift. Is the avrBsT gene exclusive to T4 strains, indicating that it was acquired simultaneous to the loss of AvrXv3? Did recombination with X. euvesicatoria occur at the same time as loss of AvrXv3 or as gain of AvrBsT? In previous studies, only T4 strains had been genotyped (Schwartz et al. 2015, Timilsina et al. 2015). We hypothesized that the loss of AvrXv3, gain of AvrBsT, and recombination with X. euvesicatoria were related events, such that AvrBsT is exclusive to T4 recombinant strains. To test this hypothesis, we genotyped T3 strains from the 2006 and earlier surveys and compared them to T4 strains collected in 2006, 2010 and 2012. We sequenced the housekeeping gene gapA and the effector xopQ as indicators of the recombination event with X. euvesicatoria to determine the timing of recombination relative to the race shift. We examined sequences of avrXv3 to determine the genetic

basis of the race shift, and we sequenced avrBsT and xopJ4 to determine their stability as potential targets for resistance breeding against X. perforans.

Materials and Methods

Bacterial Strains Collection and Pathogenicity Testing

Xanthomonas perforans strains used in this study were isolated over a period of

20 years from symptomatic plants (Table 4-1). In Florida, T3 and T4 strains of X. perforans were first identified in 1991 and 1998, respectively. Representative strains from each year between 1991-1998 and T3 strains from 2006 were selected for study.

Additional T3 strains from Mexico and Alabama were also included for comparison.

Tomato and pepper differentials were inoculated to confirm pathogenicity and race (Stall et al. 2009). To determine race, bacterial suspensions of approximately 5 X

108 CFU/mL (0.3 A at 600 nm) were infiltrated into tomato genotypes that included

Hawaii 7998 with the rxv resistance gene, FL216 with the Xv3 resistance gene, a tomato genotype carrying the RXopJ4 resistance locus from Solanum pennellii LA716, and the susceptible cultivar Bonnie Best. Similarly, bacterial suspensions were infiltrated into four pepper genotypes, Early Calwonder (ECW), ECW-10R, ECW-20R and ECW-30R (Bouzar et al. 1994). Xanthomonas euvesicatoria 85-10 and X. perforans

91-118 were used as control strains.

Gene Sequencing

Bacterial DNA was extracted using a selective precipitation method with CTAB-

NaCl extraction protocol (Ausubel et al. 1994). Previous studies using multilocus sequence analysis of six housekeeping genes showed different allelic forms of gapA and gyrB that differentiated X. perforans Group 1 from Group 2 (Timilsina et al. 2015).

To assign X. perforans strains used in this study to Group, we sequenced a portion of

gapA. To examine allelic variation in effectors, partial sequences of avrXv3 and xopJ4

(avrXv4), the effector genes responsible for determining tomato races in X. perforans, and xopQ were sequenced. We tested for the presence of avrBsT in T3 X. perforans strains using internal and flanking primers. When found, the complete gene was amplified and sequenced. For sequence comparison, avrBsT from two X. vesicatoria

strains, Xv144 and ETH17, isolated from Argentina and Ethiopia respectively, was

sequenced. Sanger sequencing was conducted at the Interdisciplinary Center for

Biotechnology Research (ICBR) at the University of Florida. The target genes

sequenced in this study are listed in Table 4-2 along with their respective primers.

Sequence Analysis

Housekeeping and effector genes sequenced were compared to other reference

Xanthomonas, including X. euvesicatoria, X. gardneri, X. vesicatoria and X. perforans

sequences available from the plant associated and environmental microbes database

(PAMDB) and National Center for Biotechnology information (NCBI) databases.

Sequence variation in the effector genes was screened comparing nucleotide

polymorphisms using Basic Local Alignment Search Tool (BLAST) (Altschul et al. 1990).

Allele designations were assigned manually based on sequence similarity. All

sequences were aligned using the Muscle tool in MEGA software ver. 6 (Tamura et al.

2013). Sequences for the gapA housekeeping gene were used to construct a maximum

likelihood phylogenetic tree using general time reversible model with gamma distributed

invariant sites in MEGA (Tamura et al. 2013). Model selection for phylogenetic analysis

was determined using Akaike information criterion (Burnham and Anderson 2004) in

jModeltest v.2.1.4 (Darriba et al. 2012).

Sequence Submission

The genes sequenced for this study were submitted to the NCBI database. The

accession numbers are: KU759430 - KU759471 for partial gene sequences of gapA,

KU759370 – KU759372 for complete coding regions of avrBsT, KU759373 – KU759401

for partial sequence of avrXv3, KU759402 – KU759429 for partial sequence of xopJ4

and KU759472 – KU759498 for partial sequence of xopQ. The reference gene

sequences used during this study are available on NCBI database and PAMDB

database.

Results

Pathogenicity and Race Determination

Strains previously assigned to tomato race 3 (T3) were confirmed as T3. These

strains were pathogenic on tomato differentials Bonnie best and Hawaii 7998 but

elicited a hypersensitive response on the tomato cultivars carrying the Xv3 resistance

gene and the RXopJ4 resistance gene (Table 4-1). One X. perforans strain isolated in

1998, Scott-1, was confirmed as T4. It did not elicit a hypersensitive response on

FL216, which carries the Xv3 gene. PCR amplification of the avrXv3 and xopJ4 effector genes produced results consistent with the assigned tomato race.

Grouping of T3 Strains by gapA Sequence

The housekeeping gene gapA differentiates the two genetic groups of T4 X.

perforans strains (Timilsina et al. 2015). Group 2 strains of T4 X. perforans have a gapA

sequence identical to the X. euvesicatoria gapA gene, which differs by 4 or 5

nucleotides from the Group 1 alleles. Sequencing of gapA in T3 X. perforans strains

isolated between 1991 and 2006 revealed three different alleles (Figure 4-1 and Figure

4-2). Sequence comparisons showed a single nucleotide difference in gapA between

strain 91-118 and strains Xp1268, Xp1275, Xp1550 and Xp1564, isolated in 1993 or

1994 (designated as allele 1.2 in Table 4-1). Strain Xp1913 differed from strain 91-118 by 19 nucleotides for gapA. This allele was most similar to a non-pathogenic strain,

ETH12, isolated from tomato in Ethiopia (Kedebe et al. 2014). All of the T3 strains collected in 2006 were identical to each other and to strain Xp91-118 for gapA. None of the T3 strains contained a gapA allele identical to the Group 2 recombinant T4 X. perforans strains.

Effector Gene Sequences

Four effector genes avrXv3, xopJ4, xopQ and avrBsT were sequenced for the study. Effector gene sequences, avrXv3 and xopJ4, were found to be identical in all examined T3 X. perforans strains. Whole genome sequences of T4 X. perforans strains were examined for presence and allelic diversity of the avrXv3 gene. Three different mutations were observed in avrXv3 (Figure 4-3). In Group 2 strains, there was a contig break in avrXv3, with remnants of an insertion sequence. Primers specific to avrXv3 were not able to amplify the gene and inoculations in tomato differentials indicated that

AvrXv3 was nonfunctional in these strains. Within Group 1 strains, the two subgroups that were previously defined by whole genome sequencing (Schwartz et al. 2015), showed different forms of mutations in avrXv3. In Group 1A, an early stop codon was detected due to a point mutation. In Group 1B, avrXv3 was either completely absent or truncated, rendering AvrXv3 non-functional in both cases. In contrast, xopJ4 had identical sequence in all T3 and T4 strains of X. perforans.

Like gapA, T4 X. perforans strains have two allelic forms of XopQ consistent with the two groups (Schwartz et al. 2015). The xopQ sequence of Group 1 strains is identical to type strain Xp91-118, but Group 2 strains have a xopQ sequence identical to

X. euvesicatoria. The difference between these alleles of XopQ is 10 nucleotides, or 6

amino acids (Figure 4-4). For all T3 X. perforans strains examined, xopQ sequences

were identical to strain Xp91-118 and Group 1 T4 X. perforans strains, consistent with

the gapA results.

Presence of AvrBsT in T3 X. perforans

Primers were designed to amplify the entire type III secreted effector gene,

avrBsT along with ~0.5 kb flanking regions on either side of the gene. These primers identified the presence of avrBsT gene in T3 strains of X. perforans isolated in 2006 but

not prior to 2006 (Table 4-1). Although primers were designed for PCR amplification of

avrBsT, two different band sizes were observed during gel electrophoresis (Figure 4-5).

The expected band size that of the functional avrBsT gene in T4 X. perforans strains

was approximately 2 kb. Sequencing of the 2 kb band amplified from T3 X. perforans

strains showed an intact avrBsT gene that was identical to that in T4 X. perforans

strains. A smaller ~1.2 kb band was observed in strain Xp1861 isolated in 1996 from

Alabama and in strain Xp1-13 isolated in 2006 from Florida. These PCR products were

sequenced and determined to have no homology to avrBsT, but a fragment of

transposon sequence was found. The primers for avrBsT were designed to anneal the

flanking sequence of the gene but were also found to be identical to the putative

transposon region. Primers internal to the coding region of the avrBsT gene were used

to further confirm the presence or absence of avrBsT in all strains. The X. perforans

strains collected from 1991 to 1998 lacked avrBsT, while all but two of the T3 strains

collected in 2006 contained the gene. The avrBsT gene sequences from both T3 and T4

strains of X. perforans were identical to the avrBsT gene from X. vesicatoria strain Bv5-

4a (GQ266402.1), but varied from the avrBsT gene from X. euvesicatoria strain 85-10

by 2 nucleotides (AF156163.1). We found the X. perforans-type avrBsT allele in other X. vesicatoria strains (Figure 4-6). An identical avrBsT allele is also present in the cassava pathogen X. axonopodis pv. manihotis strain CIO151 (AKDA01000004).

Discussion

Bacterial spot disease of tomato and pepper is caused by four different

Xanthomonas species, but tomato in Florida is currently affected only by X. perforans, and in recent years only tomato race 4 strains of X. perforans have been found

(Schwartz et al. 2015). Prior to 1991, only tomato race T1 X. euvesicatoria was isolated from bacterial spot affected tomato in Florida. The population shift from T1 X. euvesicatoria to T3 X. perforans was attributed to bacteriocin activity of T3 X. perforans

(Hert et al. 2005). The shift from tomato race 3 (T3) to race 4 (T4) started in 1998, when the first T4 strain was found (strain Scott-1), and was complete by 2012 (Figure 4-7).

The reason for the population shift from T3 to T4 X. perforans remains unknown, as no known sources of resistance towards X. perforans are commercially available. Whole genome sequence analysis of X. perforans identified substantial variation in the effector profiles among T4 strains collected in 2006 and 2012, and a recombination event affecting gapA and xopQ among other genes (Timilsina et al. 2014, Schwartz et al.

2015). The most notable change in effector profiles in comparison to the T3 strain Xp91-

118 was the presence of the avrBsT gene in most of the T4 X. perforans strains. The avrBsT effector is universally recognized by most varieties of pepper. Kim et al. (2010) demonstrated that X. vesicatoria strain Bv5-4a with avrBsT exhibited higher in planta titers post inoculation on a susceptible tomato line compared to strains lacking avrBsT.

Therefore, X. perforans strains with avrBsT may have gained increased virulence and fitness on tomato. Only three strains of Group 2 T4 X. perforans have been found

without the gene for AvrBsT to date. Although, X. perforans is usually only pathogenic

on tomato, an X. perforans strain was isolated from symptomatic pepper that lacked

AvrBsT (Schwartz et al. 2015; Figure 4-7, Table 4-3). To determine if changes were occurring in the Florida population prior to the race shift, we characterized T3 strains of

X. perforans that were collected from 1991 to 2006. We did not find recombinant gapA or xopQ genes in T3 X. perforans. In addition, all strains exhibited identical avrXv3 and xopJ4 gene sequences. Given the apparent homogeneity of the T3 X. perforans population, we were surprised to find avrBsT in T3 strains isolated in 2006. Thus, our results show that AvrBsT was acquired before the race shift, or during the race shift by both T3 and T4 X. perforans strains.

All T3 X. perforans strains isolated in 2006 had the same avrBsT allele as T4 X. perforans. Interestingly, this allele has 2 adjacent nucleotides that are different from the avrBsT sequence in the X. euvesicatoria type strain 85-10. The two sites result in substitutions of two adjacent amino acids in the AvrBsT protein. The avrBsT gene from

X. perforans was identical to the avrBsT gene of X. vesicatoria strain LMG 919 isolated

from Zimbabwe (GCA_001469445.1). Identical avrBsT sequence was detected in other

X. vesicatoria strains isolated from Argentina and Ethiopia that were sequenced during

this study. The effector is not universally present in X. vesicatoria, as the type strain of

X. vesicatoria ATCC 35937 appears to lack the avrBsT gene. X. vesicatoria is known to

be resistant to bacteriocins produced by X. perforans, and given the sequence identity

of avrBsT between these two species, avrBsT may have been acquired through

horizontal gene transfer from strains of X. vesicatoria carrying avrBsT. This is a

surprising find, because there are no reports of X. vesicatoria in Florida. However, the

presence of multiple Xanthomonas species was associated with bacterial spot in Africa

(Kebede et al. 2014, Jibrin et al. 2015, Timilsina et al. 2015). The presence of multiple

Xanthomonas species within the same tomato growing region increases the likelihood of genetic exchange between species. Further studies examining the global distribution of avrBsT in the Xanthomonas bacterial spot species could help to clarify the importance of this effector in Xanthomonas populations on tomato, determine whether there is selection for strains carrying avrBsT, and identify the avrBsT donor to X. perforans.

We previously found genetic exchange among bacterial species affecting housekeeping genes and the xopQ effector (Timilsina et al. 2015, Schwartz et al. 2015).

Whole genome sequencing confirmed that there are at least two genetic groups of T4 X. perforans with Group 2 identified as recombinants with X. euvesicatoria. Within Group

1, two subgroups were observed. Group 1A was identified exclusively in 2012. Group

1B is a more genetically diverse group that includes the type strain Xp91-118 but does not include any strains collected in 2012. The groups are defined based on core genome phylogeny, yet each group has distinct unique mutations in the avrXv3 gene resulting in a shift from T3 to T4. Group 1A had a point mutation resulting in an early stop codon, Group 1B either lacked the entire avrXv3 coding region or a section of the gene making it a likely pseudogene, and Group 2 contained insertion sequences in avrXv3 (Figure 4-7, Table 4-3). The unique mutations within the genetically homogenous Groups 1A and 2 suggest that the distinct mutations in avrXv3 occurred in a strain ancestral to each of these groups and any further variation in effector profiles occurred after the loss of AvrXv3.

We looked for evidence of recombination with X. euvesicatoria in T3 strains by sequencing the gapA and xopQ genes. We found no variation in xopQ within T3 strains.

T3 strains had three gapA alleles, but none of them were the recombinant allele from X. euvesicatoria. The most common allele was identical to that of X. perforans type strain

91-118, while the second allele varied by only one nucleotide and was isolated in 1993 and 1994 only. The third allele found in strain Xp1913, isolated in 1996 in Alabama, exhibited a gapA gene sequence that differed by 19 nucleotides from the other sequenced X. perforans strains. This gapA sequence is closely related to an unknown

Xanthomonas species isolated from Ethiopia and originally identified as X. euvesicatoria, but found to be nonpathogenic on tomato (Kebede et al. 2014). Here,

Xp1913 was identified as a T3 X. perforans by inoculation on differential tomato lines and based on sequence similarity of the avrXv3 and XopJ4 genes to other sequenced

X. perforans strains.

The development of tomato cultivars with individual dominant R genes to bacterial spot has been undermined by changes in bacterial spot species and their effectors. A long-term breeding program to introgress Xv3 into tomato genotypes began in 1991 after the displacement of T1 X. euvesicatoria by T3 X. perforans in Florida. In

1998, X. perforans overcame T3 resistance as a result of a mutation in avrXv3, prior to commercial deployment of Xv3. More concerning was that by 2006, 75% of the Florida population contained mutations in avrXv3 without any apparent selection pressure.

Similarly, in pepper, Bs2 was introgressed into commercial pepper genotypes in an effort to produce cultivars with durable resistance to the bacterial spot pathogen.

Although pepper cultivars containing Bs2 provided high levels of resistance for a few

years, bacterial spot began to develop on the cultivars resulting in significant disease.

Gassmann et al. (2000) characterized Xanthomonas strains associated with these

outbreaks and determined that point mutations occurred in avrBs2 that resulted in the

protein not being recognized by Bs2, but still allowed AvrBs2 to function as a virulence

factor in pepper. Although Bs2 has been transferred to tomato (Horvath et al. 2012), it is

unlikely that this gene alone will provide durable resistance. Therefore identifying

multiple R genes that target Xanthomonas effectors will be required.

Breeding for resistance against pathogens depends on selecting appropriate

effector gene targets. A single nucleotide variation in an effector may undermine the

usefulness of the resistance gene. Thus, for breeding purposes, effectors that have

conserved sequences within a diverse pathogen population would be preferred. The

factor(s) driving the race shift from T3 to T4 in the X. perforans population in Florida remains unknown. However, our data suggest that mutations in avrXv3 were associated with other simultaneous changes in the population that resulted in additional variation in effector profiles. Nevertheless, we confirmed that the xopJ4 gene is conserved in all X. perforans strains regardless of tomato race or genetic group and in strains collected over a period of 20 years. Effector XopJ4 is currently a stable effector in X. perforans.

Likewise, the acquisition of the avrBsT in X. perforans makes this gene another potential target for resistance breeding. Preliminary field data indicate that AvrBsT is a virulence factor in X. perforans (Abrahamian et al., unpublished data), and this effector is largely present in recent X. perforans collections. Additionally, corresponding host resistance has been studied against both XopJ4 and AvrBsT effectors. A resistance locus RXopJ4 that recognizes xopJ4 was mapped (Sharlach et al. 2013) and identified

in the wild tomato accession Solanum pennellii LA716. The locus is a 4.2 mb segment conferring resistance to Xanthomonas strains carrying xopJ4 and the resistance gene was fine mapped to a 190 kb region. Likewise, heat shock protein 70a was recently

reported to interact with the avrBsT effector to induce programmed cell death in pepper

(Kim and Hwang 2015). An incompatible reaction in the presence of avrBsT gene was

observed in S. lycopersicoides and its hybrid S. lycopersicum (Wang 1992), as well as

in most pepper cultivars (Potnis et al. 2011), although the resistance locus has not been

identified. Additional studies are needed to identify the specific R genes against xopJ4

and avrBsT. In conclusion, these genes could potentially be pyramided with the Bs2

resistance gene to develop cultivars with more durable resistance against X. perforans

in Florida.

Table 4-1. Xanthomonas perforans strains used in this study and their genotypes Strain Source Year Host Tomato avrXv3 avrBsT xopJ4 xopQ gapA Race allele allele§ X. perforans Xp894 Florida 1991 T T3 + - + 1 1 Xp909 Florida T T3 + - + 1 1 Xp1183 Florida T T3 + - + 1 1 Xp1118 Florida 1992 T T3 + - + 1 1 Xp1144 Florida T T3 + - + 1 1 Xp1126 Florida T T3 + - + 1 1 Xp1221 Florida 1993 T T3 + - + 1 1 Xp1224 Florida T T3 + - + 1 1 Xp1241 Florida T T3 + - + 1 1 Xp1247 Florida T T3 + - + 1 1 Xp1262 Florida T T3 + - + 1 1 Xp1268 Florida T T3 + - + 1 1.2 Xp1275 Florida T T3 + - + 1 1.2 Xp1508 Florida T T3 + - + 1 1 Xp1345 Mexico T T3 + - + 1 1 Xp1485 Mexico T T3 + - + 1 1 Xp1550 Florida 1994 T T3 + - + 1 1.2 Xp1564 Florida T T3 + - + 1 1.2 Xp1574 Florida T T3 + - + 1 1 Xp1757 Mexico T T3 + - + 1 1 Xp1775 NA 1995 T T3 + - + 1 1 Xp1797 Florida T T3 + - + 1 1 Xp1805 Florida T T3 + - + 1 1 Xp1835 Florida 1996 T T3 + - + 1 1 Xp1856 Florida T T3 + - + 1 1 Xp1861 Alabama T T3 + - + 1 1 Xp1911 Florida 1997 T T3 + - + 1 1 Xp1912 Florida T T3 + - + 1 1 Xp1913 Florida T T3 + - + 1 3 Xp1920 Florida 1998 T T3 + - + 1 1 Scott-1 Florida T T4 Ø + + 1 1 Xp1-5 Florida 2006 T T3 + + + 1 1 Xp1-6 Florida T T3 + + + 1 1 Xp1-13 Florida T T3 + - + 1 1 Xp3-8 Florida T T3 + + + 1 1 Xp3-12 Florida T T3 + + + 1 1 Xp3-16 Florida T T3 + + + 1 1 Xp5-8 Florida T T3 + - + 1 1 Xp5-9 Florida T T3 + - + 1 1 Xp5-14 Florida T T3 + + + 1 1 Xp5-15 Florida T T3 + + + 1 1 Xp6-19 Florida T T3 + + + 1 1 Xp10-20 Florida T T3 + + + 1 1 Xp20-5 Florida T T3 + + + 1 1 Xp20-7 Florida T T3 + + + 1 1 Xp3-15 Florida T T4* Ø + + 2 2 Xp4-20 Florida 2006 T T4 + + 1 1 Xp5-6 Florida T T4 - + 1 1 Xp7-12 Florida T T4 Ø∂ + + 2 2 Xp8-16 Florida T T4 Ø∂ + + 2 2 Xp9-5 Florida T T4 Ø + + 2 2 Xp10-13 Florida T T4 Ø + + 2 2 Xp11-2 Florida T T4 - + + 1 1 Xp15-11 Florida T T4 Ø + + 1 1 Xp17-12 Florida T T4 - - + 1 1 Xp18-15 Florida T T4 - + + 1 1 Xp2010 Florida 2010 P T4 Ø - + 2 2

Table 4-1. Continued Strain Source Year Host Tomato avrXv3 avrBsT xopJ4 xopQ gapA Race allele allele§ GEV839 Florida 2012 T T4 Ø + + 2 2 GEV872 Florida T T4 ES + + 1 1 GEV893 Florida T T4 ES + + 1 1 GEV904 Florida T T4 ES + + 1 1 GEV909 Florida T T4 ES + + 1 1 GEV915 Florida T T4 ES + + 1 1 GEV917 Florida T T4 ES + + 1 1 GEV936 Florida T T4 ES + + 1 1 GEV940 Florida T T4 ES + + 1 1 GEV968 Florida T T4 ES + + 1 1 GEV993 Florida T T4 ES + + 1 1 GEV1001 Florida T T4 Ø + + 2 2 GEV1026 Florida T T4 ES + + 1 1 GEV1044 Florida T T4 Ø + + 2 2 GEV1054 Florida T T4 Ø + + 2 2 GEV1063 Florida T T4 Ø + + 2 2 X. vesicatoria† ETH17 Ethiopia 2011 T T2 - + - - 4 Xv144 Argentina NA T T2 - + - - 4 *, Sequence information of all T4 strains collected in 2006, 2010 and 2012 were previously published in Timilsina et al. 2015 and Schwartz et al. 2015. †, Housekeeping gene sequences of X. vesicatoria strains were published in Timilsina et al. 2015. Ø, insertion sequence. Could not be PCR amplified. potential pseudogene. ES, Early stop codon detected in the sequence. §∂,, gapA allele 1 differed from allele 1.2 by only one nucleotide and are within the same genetic group as allele 1.

Table 4-2. List of primers used in this study Gene Direction Sequence Reference xopQ* Forward AAGCTCGGCATGGTGGTCAT This study Reverse AGCAGGGTCAGCGGATCGTA avrBsT^ Forward TTGGAACAACTCCTGGCG This study Reverse ATGCACCGCACTATGCACAT avrBsT* Forward TGTTGGTAATTGAGCCAGCACTTG This study Reverse AGTTGGATGACCGCCATGTCTC avrXv3* Forward AGCAACTGTCTGCCAGCCAGAACC This study Reverse TGAGCGAGAGCTACTATCGCCTCC xopJ4* Forward ACTCGCCAAATTGTCATGC This study Reverse TACGACCTTTTCAGGGTTGG gapA* Forward GGCAATCAAGGTTGGYATCAACG Almeida et al. Reverse ATCTCCAGGCACTTGTTSGARTAG 2010 *Internal primers for genes xopQ, avrBsT, avrXv3, xopJ4 and gapA. ^ Complete coding region for avrBsT gene.

Table 4-3. Xanthomonas strains/population isolated over 21 years in Florida. Population shifts along with variation in effector profile were observed during this period. Number in brackets next to T4 X. perforans indicate specific group within T4 races as published in Schwartz et al. 2015 Year Strain/Population Host XopJ4 AvrBsT AvrXv3 XopQ gapA 1991 T3 X. perforans Tomato 1 - 1 1 1 1998 T4 X. perforans Tomato 1 1 Ø 1 1 2006 T3 X. perforans Tomato 1 - 1 1 1 T3 X. perforans 1 1 1 1 1 T4 X. perforans (1B) 1 1 ∂ / - 1 1 T4 X. perforans (1B) 1 1 Ø 1 1 T4 X. perforans (2) 1 1 Ø 2 2 2010 T4 X. perforans (2) Pepper 1 - Ø 2 2 2012 T4 X. perforans (1A) Tomato 1 1 ES 1 1 T4 X. perforans (2) 1 1 Ø 2 2 Ø, Contig breaks unable to be confirmed by Sanger sequencing; Potential pseudogene -, Absence of gene ES,∂, Early stop codon detected in the sequence.

Figure 4-1. Maximum likelihood of phylogenetic analysis of Xanthomonas perforans strains using gapA gene sequence. Strains are color-coded based on the years they were isolated. T4 Group 1 and T4 Group 3 represent the gapA sequences for these respective group of strains collected in 2006 and 2012 (listed in Table 4-1). Reference strains of X. euvesicatoria, X. gardneri, X. perforans and X. vesicatoria were downloaded from PAMDB and NCBI database. Values on the branches indicate bootstrap values for each branch expressed as percentages. The scale bar indicates the number of substitution per site.

Figure 4-2. gapA partial sequence alignment for four different allelic forms observed in X. perforans. gapA-allele-1 is found in type strain of X. perforans, allele-1.2 has one nucleotide variation from allele 1. Allele 2 is found in the type strain of X. euvesicatoria and Group 2 T4 X. perforans strains. Allele 3 was observed in only one T3 strain, Xp1913. Variable nucleotides are highlighted.

Figure 4-3. Complete coding region of avrXv3 as observed in the type strain of X. perforans, 91-118. Three different mutations are observed in T4 strains. Two T4 strains, Xp4-20 and Xp5-6, do not have the first 30 nucleotides of the gene thus making the gene non-functional. An early stop codon is detected in the Group 1A T4 strains of X. perforans as a result of single point mutation in the 205th nucleotide from cytosine to thymine. A contig break after the 434th nucleotide, with traces of transposon sequences, was observed in Group 2 T4 strains of X. perforans. Other Group 1B T4 strains are missing the gene completely.

Figure 4-4. Partial gene sequence alignment for xopQ in X. perforans. XopQ-allele-1 is found in the type strain of X. perforans, Group 1 T4 strains, and all T3 strains sequenced to date. XopQ-allele-2 is found in the type strain of X. euvesicatoria and Group 2 T4 X. perforans. Variable nucleotides are highlighted.

Figure 4-5. Gel-electrophoresis of avrBsT gene including ~0.5kb flanking regions on either side of Xanthomonas perforans isolates collected in 2006. Two different band sizes were observed. The first lane was a reference strain for avrBsT gene, Scott-1. Strain 1-13 had a smaller band size, strains 5-9 and 5-14 lack the gene and the remaining 2006 T3 strains contain the avrBsT gene. From left to right: ‘1’- Scott-1, ‘2’ – Xp1-5, ‘3’ – Xp1-6, ‘4’ – Xp1-13, ‘5’ – Xp3-8, ‘6’ – Xp3-12, ‘7’ – Xp3-16, ‘8’ – Xp5-8, ‘9’ – Xp5-9, ‘10’ – Xp5-14, ‘11’ – Xp5-15, ‘12’ – Xp6-19, ‘13’ – Xp10-20, ‘14’ – Xp20-5, ‘15’ – Xp20-7 with ‘L’ – Ladder on either sides. (Photo courtesy of author)

Figure 4-6. Maximum likelihood phylogenetic tree of avrBsT showing sequence identity between alleles in X. vesicatoria and X. perforans strains. Additionally, the avrBsT sequence from X. axonopodis pv. Manihotis was identical to the avrBsT in bacterial spot causing Xanthomonas bacteria.

Figure 4-7. Infograph showing shifts in the Xanthomonas population on tomato in Florida over time. Prior to 1991, only X. euvesicatoria was identified. Later, the population changed from T3 X. perforans to T4 X. perforans. Three different mutations are observed in the AvrXv3 gene of T4 strains. The population acquired AvrBsT, which was first observed in 1998, in most strains in 2006, and in all strains examined from 2012.

CHAPTER 5 CORE GENOME MULTILOCUS SEQUENCE TYPING OF XANTHOMONAS PERFORANS

Introduction

Xanthomonads causing bacterial spot of tomato are continuously evolving in

Florida. The yield loss from bacterial spot was estimated up to $3,000 per acre (Van

Sickel et al. 2009). During the study of bacterial spot disease of tomato over the past 25 years, X. perforans emerged as the major causal bacterium in Florida (Horvath et al.

2012) and number of recombination events and mutational changes were observed

(Timilsina et al. 2016). The first strain of X. perforans in Florida was isolated in 1991, prior to which X. euvesicatoria was the only known species of Xanthomonas present in

Florida.

Bacterial spot causing xanthomonads have been categorized into four groups based on a number of phenotypic tests including fatty acid profiles, protein profiles and amylolytic and pectolytic activity (Jones et al. 2004); furthermore they have been characterized into at least four tomato races based on the presence of avirulence gene(s) in various bacterial strains that gives hypersensitive reaction on plants carrying a corresponding resistance gene(s) (Stall et al. 2009). Strains of X. perforans belong to

group C and encompass two tomato races: race 3 (T3) and race 4 (T4). Gene mutations

resulting in race shifts and recombination events shaped the present X. perforans

population in Florida. Initially, X. perforans isolated from Florida were identified as T3

strains and carried both avrXv3 and xopJ4 effectors. Additionally, all T3 strains of X.

perforans produced bacteriocins antagonistic against tomato race 1 (T1) X.

euvesicatoria strains (Hert et al. 2005). Later in 1998, a T4 strain was isolated in Florida

with a non-functional avrXv3 gene (Astua-Monge et al. 2000). Surveys conducted later

in 2006 and 2012 determined that T4 had become the dominant tomato race in Florida.

Extensive multilocus sequence analysis/typing (MLSA/T) studies have been

conducted in the Xanthomonas genus (Young et al. 2008, Hamza et al. 2012, Timilsina

et al. 2015). Various MLST schemes using four to six housekeeping genes were implemented to study bacterial spot causing xanthomonads (Young et al. 2008, Almeida

et al. 2010). Genetic characterization using six housekeeping genes from a worldwide collection of bacterial spot causing xanthomonads reported recombination between X. perforans and X. euvesicatoria (Timilsina et al. 2015). Two phylogenetic groups, group

1 and 2, of T4 X. perforans were identified. Group 1 strains were similar to T3 strains of

X. perforans, while group 2 strains shared sequence similarity for two housekeeping

genes with X. euvesicatoria (Timilsina et al. 2015). Interestingly, variations in bacteriocin

activity and effector profiles were observed between strains of X. perforans (Schwartz et

al. 2015). Although X. perforans is the dominant pathogen on tomato in Florida, only

one X. perforans strain, Xp2010, has been isolated from pepper (Schwartz et al. 2015).

Based on MLSA of six housekeeping genes, the pepper pathogen grouped together

with X. euvesicatoria for two of the genes suggesting recombination between X.

euvesicatoria and X. perforans (Timilsina et al. 2015). MLSA/T can be useful in creating

allelic profiles and studying molecular epidemiology and bacterial evolution. However,

MLSA approaches are limited as only a few genes are considered for the study.

Whole genome sequencing (WGS) of bacterial strains provides a higher

resolution in phylogenetic studies (Kwong et al. 2015). The decreasing cost of high-

throughput sequencing now allows efficient bacterial strain characterization and

comparisons. Previous studies using WGS in X. perforans focused on type III secretion effectors and changes in effector profiles that affected host specificity (Schwartz et al.

2015). Phylogenetic analysis based on multilocus sequence analysis using six

housekeeping genes identified recombination between X. perforans and X.

euvesicatoria. However, a recent study on effector analysis showed that X. perforans

likely received effector genes, like avrBsT, from other closely related xanthomonads

besides X. euvesicatoria (Timilsina et al. 2016). Phylogenetic studies reveal pattern of

evolution within closely related organisms that provides basis for comparative studies.

Whole genomes provide additional niches for phylogenetic and genomics studies (Soltis

and Soltis 2003). Thus, extensive phylogenetic analysis based on WGS is important in studying the impact of recombination during X. perforans evolution. WGS provides

variation for studying the phylogenetic relationships and epidemiological linkages

between sequenced strains (Alföldi and Lindblad-Toh 2013, Leekicharoenphon et al.

2014). Various mechanisms and tools for whole genome sequence comparisons can be

implemented by comparing single nucleotide polymorphisms (SNPs) or individual

genes. Gene-by-gene comparison can be used along with traditional MLST, as an

extended MLSA/T for evolutionary and epidemiological studies. Core genes are

identified among the genomes used for comparison that will basically extend MLSA/T

analyses using several hundred to over a thousand genes. Phylogenetic studies can be

conducted using those conserved genes among the genomes, commonly termed as

core genome MLST (cgMLST), whole genome MLST or MLST+ (Maiden et al. 2013).

The approach has been used for population and functional studies on a number of

bacterial pathogens including Nisseria meningitis, Escherichia coli, and Mycobacterium

tuberculosis (Bratcher et al. 2014, Leopold et al. 2014, Kohl et al. 2014). With the diversity in X. perforans population identified previously (Schwartz et al. 2015, Timilsina et al. 2015), the objective of this study focused on using WGS to evaluate phylogenetic relationships and patterns of evolution within X. perforans strains. A gene-to-gene comparison using core genes of X. perforans strains collected over time in Florida was used in this study to visualize the relationships between bacterial strains. This study will likely serve as a model for using WGS to identify core genes of plant pathogenic bacteria for an improved determination of phylogenetic relationships.

Materials and Methods

Bacterial Strains and Genome Sequencing

The genomes of 53 T4 X. perforans (Xp) strains collected in Florida in 2006,

2012, and 2015 were used in this study along with the type strain Xp91-118 (Table 5-1).

Whole genome sequences of 33 strains, including the type strain Xp91-118 were published prior to this study (Schwartz et al. 2015, Potnis et al. 2011). The remaining 20

X. perforans strains collected in 2015 were sequenced and genomic libraries were

prepared using a Nextera library preparation kit (Illumina Inc., San Diego, CA). The

genomes were sequenced using Illumina MiSeq platform at the Interdisciplinary Center

for Biotechnology Research, University of Florida. The sequences yielded 609,582 to

1,274,455 reads per strain. Draft genomes were de novo assembled using CLC

Genomics Workbench v5, with an average of 37.31X coverage. The assembled

sequences were annotated using IMG/JGI platform (Markowitz et al. 2012). Pairwise

Average Nucleotide Identity (ANI) based on blast was calculated using jSpecies v1.2.1

(Richter and Rosselló-Móra 2009) for the 53 strains used in this study.

Multilocus Sequence Analysis

MLSA based on six housekeeping genes (fusA, gapA, gltA, gyrB, lacF and lepA), was used for preliminary phylogenetic characterization of X. perforans strains collected

recently in 2015 along with strains isolated previously. The sequences of recently

collected X. perforans strains were compared with sequences of four Xanthomonas

species causing bacterial spot of tomato and pepper (Potnis et al. 2011). Sequences

were aligned using muscle tool (Edgar 2004) and a maximum likelihood phylogenetic

tree was constructed using general time reversible model with gamma distributed

invariant sites in MEGA v.6 (Tamura et al. 2013).

Core Genome Multilocus Sequence Typing

The draft assemblies of bacterial genomes were annotated using the IMG-JGI

annotation pipeline (Markowitz et al. 2012). The annotated sequences were used to

identify the core genes among the 53 genomes using get_homologues v2.0.19

(Contreras-Moreira and Vinuesa 2013). Genes present in at least 95% of the input

genomes with at least 75% pairwise alignment coverage were considered as core

genes. The output core gene sequences were parsed using a python script to identify

core genes that were present in 100% of the genomes that contained a start codon

(Appendix A). Multiple copies of the same gene within a single genome were removed

from the analysis. The sequences were then aligned by Mafft alignment using a python

script (Cock et al. 2009).

A total of 1,316 genes were identified as core genes from all Xp strains that were

concatenated to create a 1.2 megabases long sequence for each strain using Sequence

matrix v.1.8 (Vaidya et al. 2010). Sequence matrix v.1.8 was used to calculate

uncorrected pairwise distance for each gene with Xp91-118 as reference. The pairwise 91

distance was then used for sequence typing of all the core genes. Allele types of genes identical with genes from Xp91-118 were labeled ‘0’ and the allele type for genes was assigned gradually based on the pairwise distance from Xp91-118. Additionally, a heat map was generated in R using allele type of all core genes from X. perforans strains used in this study. The heat map was color coded to see genetic distribution of core genes based on phylogenetic groups.

Phylogenetic Analysis

The single gene sequences and concatenated sequences were phylogenetically analyzed using PhyML v.3.1 (Guindon et al. 2010). The nucleotide substitution model for phylogenetic analysis was selected using jModelTest (Posada 2008, Darriba et al.

2012). General time reversible with gamma distributed invariable sites was identified as the best substitution model for concatenated nucleotide sequences. For single gene sequences, the general time reversible model was used when all models supported one tree topology but the model with highest log-likelihood (lnL) was used for phylogenetic analysis if multiple tree topologies were suggested during model testing. Maximum likelihood phylogenetic trees were constructed with 100 bootstraps (Guindon et al.

2010) for both concatenated and single gene sequences using the suggested genetic model. ClonalFrameML was used to evaluate the phylogenetic tree (Didelot and Wilson

2015) to generate a consensus tree and infer the evolution of X. perforans strains from

Florida. Factors affecting the current population distribution via recombination or mutation were calculated. Pairwise homoplasy index (PHI) was calculated to identify recombination effect using SplitsTree version 4 (Huson and Bryant 2006). PHI calculates the incompatibility between closely linked sites to estimate the degree of

genealogical correlation that is negatively correlated with the rate of recombination

(Bruen et al. 2006, Hudson and Kaplan 1985).

Phylogenetic distance was computed between the single gene trees and the maximum likelihood tree based on concatenated sequences using Robinson-Foulds symmetric distance computed using a python framework available from the ETE toolkit

(Huerta-Cepas et al. 2016). Robinsion-Foulds symmetry calculates distance between two unrooted phylogenetic trees and evaluates the number of nodes or partitions supported by a phylogenetic tree in comparison to the reference tree (Robinson and

Foulds 1981).

Results

Bacterial Strains and Multilocus Sequence Analysis

Six housekeeping genes from the bacterial strains were concatenated to obtain the preliminary phylogenetic tree (Figure 5-1). Among the 20 new strains collected in

2015 and used in this study, 19 strains shared 100% nucleotide sequence identity with the group 2 of T4 X. perforans strains. The strain GEV2010 varied by a single nucleotide in lacF and gapA gene with the type strain of Xp91-118 but was grouped together as phylogenetic group 1. However, the gapA gene sequence of GEV2010 was identical with Xp17-12 isolated in 2006.

Core Gene Comparisons

Schwartz et al. (2015) described two phylogenetic groups of X. perforans along with two subgroups within group 1 strains of T4 X. perforans. Phylogenetic analyses showed the genetic composition of X. perforans strains corresponds to phylogenetic groups previously described by Schwartz et al. 2015 (Figure 5-2). All the strains from the recent collection in 2015 clustered together with group 2 T4 X. perforans strains

except for one strain GEV2010. However, the GEV2010 strain was phylogenetically closely related to Xp17-12 and shared a similar genetic pattern. The strains GEV2010 and Xp17-12 grouped with group 1 strains, however a distinct genetic pattern was observed for both strains. Recently collected strains from Florida were identified as group 2 recombinant strains and shared a majority of sequence types with the representative strains from group 2 strains isolated in 2006 and 2012. Similar phylogenetic groups and subgroups as described by Schwartz et al. 2015 were observed from the core gene analysis and the genetic profiles for the strains were consistent with the phylogenetic groups. Among the 1,316 genes used in this study, 783 genes were found identical among all X. perforans strains. Within the remaining 533 genes, 134 genes were identical between all group 1 strains including Xp17-12 and

GEV2010. The subgroup 1A varied from the type strain Xp91-118, grouped as 1B, in only 65 genes. Distinct allele types were observed in 121 genes between group 1 and group 2 strains and interestingly, 92 among the 121 genes from group 2 X. perforans were identical to the gene sequence of X. euvesicatoria strain 85-10.

Average nucleotide identity of X. perforans strains from same phylogenetic group was more than 99.9%. All group 2 strains had sequence identity ≥99.9% among themselves. With the exception of Xp17-12 and GEV2010 strains, sequence identity among Group 1 strains was also ≥99.9%. The sequence identity between Group 2 strains and Xp17-12 was as low as 99.2% but with other Group 1 strains, the nucleotide identities ranged between 99.7-99.9%. As observed in the heat map, the sequence of

Xp17-12 was very similar to GEV2010 strain isolated in 2015 with 99.89% nucleotide identity and the two strains varied from each other in only 18 genes among the core

genes compared in this study. The only X. perforans strain pathogenic on pepper,

Xp2010, had ~99.9% sequence similarity with the group 2 recombinant strains but the sequence identity when compared with group 1 strains was between 99.6-99.8%.

Phylogenetic Analysis

The tree topology for the majority of the single gene phylogenetic trees supported the consensus maximum likelihood phylogenetic tree based on concatenated gene sequences. Among the 1,316 genes used in this study, 1,107 genes had similar tree topology compared with the consensus tree based on the Robinson-Foulds symmetric distance. Although the 209 single gene trees varied from the consensus tree, 204 of the single gene trees supported the consensus tree topology by 98% and the 5 remaining single gene trees supported the tree by 96%. When observing the single gene trees that supported the consensus tree at 96%, a separate cluster of Group 1A strains was

detected. Allelic variations in different strains, probably resulting from mutation, lead to

different tree topologies for single genes that supported the consensus tree by 98%.

ClonalFrameML was used to infer the maximum likelihood phylogenetic tree

constructed in PhyML using the concatenated sequences. The ratio of rate of

recombination to mutation events (R/theta) was calculated as 0.188292. Likewise, the

average size of recombination fragment (delta, ∂) and probability of recombination per

site (nu) were calculated as 140 basepairs and 0.236 respectively. The relative effect of

recombination to mutation events (R/theta X ∂ X nu) in pathogen population was

calculated to be 6.22 suggesting the role of recombination in X. perforans population structure is significantly higher than mutation. Similarly, the observed PHI value was calculated as 0.0011 (p=0.0001), using splitstree software, indicating significant evidence for recombination. 95

Discussion

The population of bacterial spot causing X. perforans in Florida is rapidly evolving. Timilsina et al. (2016) reported the changes in X. perforans population since its first identification in 1991 (also included in Chapter 4). Recombination events, point mutations and fitness costs associated with various genes may have played a significant role in population shifts.

In this study, we developed a cgMLST approach for the phylogenetic analysis for current population of X. perforans. We utilized sequences from core genes of 53 X. perforans genomes collected during various surveys in Florida to apply the cgMLST approach. Allelic variations in genes within the core genes were phylogenetically distributed. Based on MLSA, using six housekeeping genes, all strains were identified as recombinant group 2 strains except for strain GEV2010. As the recent collection of strains from 2015 included strains from only one tomato production area in Florida, the current population status of X. perforans is not interpreted by this study. However, three strains: TB6, TB9 and TB15 collected in early 2013 were also identified as group 2 strains and the genetic composition of group 2 strains, as described by Schwartz et al.

(2015), were similar regardless of the isolation years. Similar cgMLST approach can be applied for future population studies using additional X. perforans strains representing

Florida population.

Corresponding to the previous study, two subgroups within group 1 were observed. Group 1A included the T4 X. perforans strains isolated from 2012 only.

Significant allelic variations were observed between group 1 strains and apart from a few individual strain variations, group 1A strains varied from the type strain Xp91-118 in at least 65 genes among the core genes of 1,316 genes. Group 1B encompassed 96

relatively diverse strains; as it includes the type strain Xp91-118, a T3 strain, and T4 strains isolated in 2006.

Even without the presence of any known selective pressure on X. perforans population, the pathogen is constantly evolving. Population shift was observed from T3

X. perforans to T4 X. perforans (Horvath et al. 2012, Astua-Monge et al. 2000, Timilsina

et al. 2016). Although the strains collected in 2015 were isolated in the same area,

Group 2 T4 strains were more prominent than group 1 strains, which may be due to the fitness associated with one or more genetic factor(s). Phenotypic variations, such as changes in bacteriocin activity, and genotypic variations, such as mutations in avrXv3 and the acquisition of avrBsT, were reported for T4 X. perforans strains associated with

recent population changes (Horvath et al. 2012, Schwartz et al. 2015, Timilsina et al.

2016). Interestingly, the only pepper pathogenic strain of X. perforans isolated in 2010 was identified as a potential recombinant strain and did not carry the avrBsT gene unlike the recent X. perforans isolated from tomato (Timilsina et al. 2016).

Previous results identified that the recombinant group 2 X. perforans strains

isolated from tomato were able to cause disease on pepper after the deletion of avrBsT

gene unlike strains that belong to Group 1 (Schwartz et al. 2015). Along with the Group

2 strains isolated in 2006 and 2012, the X. perforans strains isolated in 2015 carried an

identical avrBsT gene. Therefore, X. perforans strains, isolated in 2015, may be

pathogenic on pepper after deletion of the avrBsT gene.

Recombination events have played a significant role in shaping the current

population of X. perforans. Previous studies based on multilocus sequence analysis and

whole genomes compared with X. euvesicatoria confirmed the recombination events

(Schwartz et al. 2015, Timilsina et al. 2015). cgMLST showed that the recently isolated

strains have similar allelic distribution with the group 2 strains isolated in 2012 and

2006. Significant recombination was observed in our study with an average predicted

length of recombinant sites as 140 bp. The results showed that the group 2 recombinant

strains carried at least 92 genes that were identical to X. euvesicatoria, but vary from

the group 1 X. perforans strains, suggesting the genes were received during

recombination events. Along with recombination events, a number of allelic variations

were observed like the sequence types for a number of genes of strains GEV2010 and

Xp17-12 were significantly different from other X. perforans strains. Although highly

variable, the strains were phylogenetically categorized as group 1B strains.

The emerging concept of cgMLST is widely used in medically important microbes

to study bacterial characterization, population genomics and epidemiology (Willems et

al. 2016, Forsythe et al. 2014, Maiden and Harrison 2016). The cgMLST approach

provides a plethora of information that could provide the key to understand the on-going

population shifts in X. perforans. The allele types for all the core genes used in the

study showed the conserved and variable regions within various X. perforans groups.

The phylogenetic approach could provide information on pathogen characteristics and

host association and also could be used to study disease epidemiology.

Genetic movement in X. perforans population along with the source of variation

and population shifts could be studied on a large scale using the cgMLST approach.

Our study provides phylogenetic insights based on the core genes of X. perforans.

Future studies can involve additional X. perforans strains along with other closely

related Xanthomonas species to identify genes that were transferred during

recombination and their potential impact on current and future pathogen populations.

The potential of cgMLST studies lies in the capacity to use multiple genomes to understand pathogen diversity, epidemiology and population studies of a bacterial population. The approach will also be useful to provide extensive results to study plant pathogenic bacteria.

Table 5-1. List of X. perforans strains isolated in Florida used in this study Strain Host Year Phylogenetic group Remarks Xp91-118 T 1991 1B Potnis et al. 2011 Xp3-15 T 2006 2 Schwartz et al. 2015 Xp4-20 T 2006 1B Schwartz et al. 2015 Xp4B T 2006 1B Schwartz et al. 2015 Xp5-6 T 2006 1B Schwartz et al. 2015 Xp7-12 T 2006 2 Schwartz et al. 2015 Xp8-16 T 2006 2 Schwartz et al. 2015 Xp9-5 T 2006 2 Schwartz et al. 2015 Xp10-13 T 2006 2 Schwartz et al. 2015 Xp11-2 T 2006 1B Schwartz et al. 2015 Xp15-11 T 2006 1B Schwartz et al. 2015 Xp17-12 T 2006 1B Schwartz et al. 2015 Xp18-15 T 2006 1B Schwartz et al. 2015 Xp2010 P 2010 2 Schwartz et al. 2015 GEV839 T 2012 2 Schwartz et al. 2015 GEV872 T 2012 1A Schwartz et al. 2015 GEV893 T 2012 1A Schwartz et al. 2015 GEV904 T 2012 1A Schwartz et al. 2015 GEV909 T 2012 1A Schwartz et al. 2015 GEV915 T 2012 1A Schwartz et al. 2015 GEV917 T 2012 1A Schwartz et al. 2015 GEV936 T 2012 1A Schwartz et al. 2015 GEV940 T 2012 1A Schwartz et al. 2015 GEV968 T 2012 1A Schwartz et al. 2015 GEV993 T 2012 1A Schwartz et al. 2015 GEV1001 T 2012 2 Schwartz et al. 2015 GEV1026 T 2012 1A Schwartz et al. 2015 GEV1044 T 2012 2 Schwartz et al. 2015 GEV1054 T 2012 2 Schwartz et al. 2015 GEV1063 T 2012 2 Schwartz et al. 2015 TB6 T 2013 2 Schwartz et al. 2015 TB9 T 2013 2 Schwartz et al. 2015 TB15 T 2013 2 Schwartz et al. 2015 GEV1989 T 2015 2 This study GEV1991 T 2015 2 This study GEV1992 T 2015 2 This study GEV1993 T 2015 2 This study GEV2013 T 2015 2 This study GEV2011 T 2015 2 This study GEV2004 T 2015 2 This study GEV2063 T 2015 2 This study GEV2058 T 2015 2 This study GEV2055 T 2015 2 This study GEV2060 T 2015 2 This study

100

Table 5-1. Continued Strain Host Year Phylogenetic group Remarks GEV2047 T 2015 2 This study GEV2059 T 2015 2 This study GEV2050 T 2015 2 This study GEV2015 T 2015 2 This study GEV2048 T 2015 2 This study GEV2052 T 2015 2 This study GEV2049 T 2015 2 This study GEV2009 T 2015 2 This study

101

Figure 5-1. Maximum likelihood phylogenetic distribution of Xanthomonas perforans strains isolated from Florida based on housekeeping gene sequences of fusA, gapA, gltA, gyrB, lacF and lepA. Two distinct groups are detected based on sequence variation in gapA and gyrB genes. Group 2 strains have gapA and gyrB genes identical to X. euvesicatoria 85-10 and includes the pepper pathogenic strain Xp2010.

102

103

Figure 5-2. Maximum likelihood phylogenetic distribution of Xanthomonas perforans strains using concatenated sequences of core genes (~1.2 Mb), encompassing all 1,316 genes along with the heat map based on the allelic profiles of individual genes, within the core genes, that were not identical among all X. perforans strains (533 genes).

CHAPTER 6 SUMMARY AND CONCLUSION

Fresh market tomato production in Florida is affected by a number of bacterial diseases. The humid subtropical climate in Florida is conducive for bacterial disease development. Pseudomonas and Xanthomonas are the two major bacterial pathogens isolated from tomato in Florida. Frequent outbreaks of bacterial speck caused by P. syringae pv. tomato have been reported from

Florida since 1970s (McCarter et al. 1983). Pseudomonas viridiflava was reported to cause bacterial blight of tomato in Florida (Jones et al. 1984).

Likewise, X. perforans is the current dominant species that cause bacterial spot of tomato (Jones et al. 2004) after replacing X. euvesicatoria.

Pseudomonads are typically favored by cool and moist weather causing frequent disease epidemic. Although a several of Pseudomonas species have been reported from different hosts in Florida, only three species of

Pseudomonas: P. syrinage pv. tomato, P. viridiflava and P. corrugata were reported from tomato. We detected some atypical and previously unreported

Pseudomonas during the preliminary characterization of strains isolated from epidemics that occurred on tomato from 2010-12 in Florida tomato. Standard

LOPAT characterization identified a group of strains similar to P. cichorii and novel group of Pseudomonas strains that produced an atypical lvean reaction.

Although P. cichorii strains have been isolated from Florida since early

1980s, it was not previously reported from tomato in the state. Thus, we used phenotypic and genotypic assays to characterize P. cichorii strains isolated from multiple hosts in Florida. The P. cichorii isolated from Florida were cross inoculated in other hosts to observe the pathogenicity. As lettuce is regarded as 104

the natural host of P. cichorii (Mirik et al. 2011), the strains isolated from tomato were inoculated in lettuce. In addition, we inoculated P. cichorii strain from various hosts in tomato. We observed the ability of P. cichorii strains to cross infect hosts. Strains isolated from tomato were pathogenic on lettuce and majority of P. cichorii strains isolated from various hosts were pathogenic on tomato. We sequenced representative strains of P. cichorii for 16S rRNA and four housekeeping genes. Similar results were observed from the phylogenetic analysis using 16S and housekeeping gene sequences. The tomato strains had identical sequences except for one strain, GEV417. Variable sequences were observed for P. cichorii strains isolated from different hosts. However, phylogenetic analysis clustered P. cichorii strains from Florida together in one phylogroup, but separate from other P. cichorii strains. Besides Florida, two P. cichorii strains from Tennessee isolated from pumpkin were included in this study. Phylogenetic comparison grouped the P. cichorii strains from Tennessee together with the Florida strains.

The study used sequences from P. cichorii strains isolated from early

1980s to 2012. We observed the same phylogenetic cluster of P. cichorii present in Florida regardless of the strains being isolated from multiple hosts. The results suggest the P. cichorii strains isolated from tomato in Florida are more likely the representatives of P. cichorii previously present in Florida that may have a broader host range given environmental conditions and host availability. Based on housekeeping gene sequences, the same P. cichorii strains were also observed in strains isolated from pumpkin in Tennessee, indicating that this P. cichorii population may be present throughout the United States.

105

During the Pseudomonas outbreak of 2010 in Florida tomato, we also observed several strains exhibiting an atypical LOPAT profile. The strains produced a yellow mucoid viscous biofilm on nutrient agar amended with 5% sucrose, unlike the typical white mucoid biofilm production by levan positive strains. Similar atypical strains were reported from strains isolated from Spain that were identified as P. viridiflava based on 16S rRNA sequence comparison

(Gonzalez et al. 2003). We were interested in characterizing the novel strains and sequenced 16S gene along with four housekeeping genes. Sequence comparison and phylogenetic analysis could not associate the strains with any previously known Pseudomonas species but showed the strains were closely related with P. viridiflava. Similar results were observed from BiologTM assay used to characterize the bacteria strains based on metabolic activity. Fatty acid profiles suggested the novel strains were closely related with P. viridiflava and P. syringae pv. tagetis.

Average Nucleotide Identity (ANI) and Genome-Genome distance

Calculator (GGDC) for in-silico DNA-DNA hybridization are two common methods based on whole genome sequences of bacteria strains used for species differentiation (Kostantinidis et al. 2006, Auch et al. 2010). An ANI of 95% or greater would suggest the two genomes used for comparison belong to the same species (Kostantinidis et al. 2006). Whole genome sequence from a representative strain of the novel Pseudomonas group (GEV388) was used to calculate pairwise ANI against all publicly available genomes of Pseudomonas.

The highest ANI was observed with P. viridiflava but was only 86.61%. Insilico

DDH between GEV388 and P. viridiflava was only 31%. Both ANI and GGDC

106

values were significantly smaller than the threshold required for genomes to be considered as the same species. In conclusion, the novel strains of

Pseudomonas isolated from tomato do not belong to any previously identified

Pseudomonas species and fulfills the criteria to be described as a novel

Pseudomonas species. Thus we proposed, Pseudomonas floridae sp. nov. as a new species name for the newly identified Pseudomonas pathogenic on tomato with GEV388 as the type strain.

Along with the characterization of Pseudomonas strains isolated from

Florida, the research project also focused on X. perforans that cause bacterial spot of tomato. Xanthomonas perforans was first isolated in Florida in 1991. Prior to 1994, all bacterial spot causing Xanthomonas were characterized as X. campestris pv. vesicatoria. Later, Jones et al. (2004) classified the bacterial spot causing xanthomonads into four Xanthomonas species based on multiphasic characterization. The bacterial spot causing Xanthomonas are further categorized into four tomato races based on the presence of avirulence genes that give incompatible reaction resulting in hypersensitive response in presence of corresponding resistance gene(s) in tomato hosts.

Multiple surveys and independent isolations of X. perforans were made since its first identification in 1991 to 2015. Before 1991, X. euvesicatoria was the sole bacterial species causing bacterial spot of tomato. A tomato race 3 (T3) X. perforans was isolated in 1991 and carried antagonistic bacteriocin against tomato race 1 (T1) X. euvesicatoria (Hert et al. 2005). The T3 X. perforans carried both avrXv3 and XopJ4 effectors. Later in 1998, a tomato race 4 (T4) strain was isolated in Florida lacking a functional avrXv3 gene (Astua-Monge et

107

al. 2000). Strains of X. perforans continuously evolved without any known selection pressure and in a 2006 survey, we observed a 3:1 ratio of T4 to T3 X. perforans population in Florida (Horvath et al. 2012). Eventually, another survey conducted in 2012 found the T4 strains completely replaced T3 strains. As X. perforans in Florida are continuously evolving, we were interested to identify conserved avirulence gene(s) that could be used as resistance targets against X. perforans population for effective disease management. Based on the information gathered from whole genome sequence comparison of T4 and T3 X. perforans strains, we examined variation in putative resistance targets (XopQ,

XopJ4, avrXv3, avrBs2 and avrBsT). Effector genes of X. perforans strains collected in Florida from 1991 to 2015 were compared. Consistent with race shift, avrXv3 was present in race 3 strains but nonfunctional in race 4 strains due to multiple independent mutations. Effectors xopJ4 was unchanged in all strains.

The effector avrBsT was absent in T3 strains collected in the 1990s but present in T3 strains collected in 2006 and nearly all T4 strains. These changes in effector profiles suggest that the effectors xopJ4 and avrBsT are currently the best targets along with avrBs2 for resistance breeding against bacterial spot of tomato. Gene-for-gene resistance against avrBs2 is already incorporated in a transgenic tomato breeding line. A resistance locus against xopJ4, of size 4.2 mb, was identified in wild tomato accession Solanum pennellii (Sharlach et al.

2013). Likewise, resistance against avrBsT can be observed in almost all pepper cultivars (Potnis et al. 2011) and is also reported from a hybrid of S. lycopersicoides and S. lycopersicum (Wang 1992).

108

Furthermore, we were interested to phylogenetically analyze the population shift of X. perforans in Florida. A study based on multilocus sequence analysis using six housekeeping genes identified recombination events between

X. euvesicatoria and T4 X. perforans (Timilsina et al. 2015). Strains of T4 X. perforans similar to T3 X. perforans were designated as group 1 and the recombinant type strains were identified as group 2. This phylogenomic study using whole genome sequences identified two subgroups within group 1, reported as group 1A and 1B. We further extended the phylogenetic study using a cgMLST approach. We compared the allele types of all core genes to observe the genetic profile for each phylogenetic group. We identified a total of 1,316 genes present in all the X. perforans strains. The core genes were concatenated and a maximum likelihood phylogenetic tree was constructed. Allele types for genes were designated to generate a MLST matrix based on the pairwise distance from Xp91-118. A heat map was constructed using the MLST matrix.

We were able to reconstruct the X. perforans phylogeny using cgMLST and genes associated with different phylogenetic groups could be observed simultaneously. The potential of the cgMLST approach lies in the use of >1000 genes with extensive gene profiling. Limited informative sites are available in conventional MLST/A approaches using few housekeeping genes. However, whole genomes based approaches can be used for large scale phylogenetic and epidemiological studies. Genes conserved in each phylogenetic group can be used for tracking strains and generating a phylogenetic group specific genetic fingerprints in future studies. The cgMLST approach has already been applied extensively to study human pathogenic bacteria (Forsythe et al. 2014, Maiden

109

and Harrison 2016). We can use this similar approach/model to determine phylogenetic distribution based on cgMLST in other plant pathogenic bacteria for high-resolution population studies. Information based on cgMLST analyses can be extensively used for epidemiological and population studies to identify sources of genetic variation, evolution of a bacterial pathogen and can be applied to understand host parasite interaction, host association and global pathogen movement to design sustainable disease management strategies.

110

APPENDIX SCRIPTS USED FOR CORE GENOME ANALYSIS

Following python scripts were used for sequence parsing:

1. The single gene sequences were analyzed to find if they have a start codon and the sequence id of the genes from each genome were changed to the strain name.

from glob import glob for filename in glob("*.fna"): with open(filename) as f: import sys orig_stdout = sys.stdout output = str(filename) output += '-test.fna' #replacing the output file name final_name = file(output, 'w') #open output to write sys.stdout = final_name line_list = [] for line in f: if line[0] == '>': line = ">"+ line[line.find('[')+1:line.find(']')] line = line[:line.find(':')] line_list.append(line) elif line[0:3] == 'ATG': line_list.append(line) elif line[0:3] != 'ATG': del line_list[-1]

new_lines = '\n'.join(line_list) print new_lines sys.stdout = orig_stdout final_name.close()

2. The output files from the first step were analyzed to see if multiple copies were detected.

from glob import glob for filename in glob('*.fna'): with open(filename) as f: import sys orig_stdout=sys.stdout output = str(filename)

111

output += '.edited.fna' final_name = file(output,'w') sys.stdout = final_name dict_list = {} names = [] seq=[] for line in f: if line.startswith('>'): names.append(line) if line[0] == 'A': seq.append(line) if line[0] == '\n': if len(names) >=2: if names[-1] == names[-2]: if len(seq[-1]) > len(seq[-2]): names.remove(names[-2]) seq.remove(seq[-2]) else: names.remove(names[-1]) seq.remove(seq[-1])

else: continue

dict_list = dict(zip(names, seq)) dict_lines = ''.join(['%s%s' % (k,v) for k,v in dict_list.iteritems()]) print dict_lines sys.stdout = orig_stdout final_name.close()

3. A total of 53 genomes were compared. The output files from step 2 were used for this procedure. The genes with multiple/missing gene copies were removed from further analysis.

from glob import glob for filename in glob('*.fna'): with open(filename) as f: import sys dict_list = {} names = [] seq = [] for line in f: if line.startswith('>'): names.append(line) elif line.startswith('A'): seq.append(line)

112

if len(names) == 53: orig_stdout=sys.stdout output = str(filename) output += '.final.fna' final_name = file(output,'w') sys.stdout = final_name dict_list = dict(zip(names, seq)) dict_lines = ''.join(['%s%s' % (k,v) for k,v in dict_list.iteritems()]) print dict_lines sys.stdout=orig_stdout final_name.close()

4. The single gene sequences were aligned using Mafft Alignment. Single gene sequence files created from step 3 were used for this analysis as an input sequence.

from glob import glob for filename in glob('*.fna'): with open(filename) as f: output = str(filename) output += '-aligned.fna' for the filename you want to use in_file = str(filename) from Bio.Align.Applications import MafftCommandline mafft_cline = MafftCommandline(input=in_file) stdout, stderr = mafft_cline() with open(output, 'w') as handle: handle.write(stdout)

Following perl script was used to run jModelTest and PhyML for single gene sequences in a batch. a. Phylogenetic model were tested using jModelTest1.2.1 for filename in ./*.fna do java –jar jModelTest.jar –d “${filename}” –g 4 –i –f –AIC –a –o “${filename}.log” done b. Although GTR model is used in the following script, respective models were used for phylogeny construction.

for filename in ./*.fna do phyml –I “${filename}” –m GTR –b 100 –s NNI > ./phylip/tree done

113

LIST OF REFERENCES

Agrios, G.N. 2005. Plant Pathology. Academic Press.

Alföldi, J., and Lindblad-Toh, K. 2013. Comparative genomics as a tool to understand evolution and disease. Genome Res. 23, 1063–1068. doi:10.1101/gr.157503.113

Almeida, N.F., Yan, S., Cai, R., Clarke, C.R., Morris, C.E., Schaad, N.W., Schuenzel, E.L., Lacy, G.H., Sun, X., Jones, J.B., Castillo, J.A., Bull, C.T., Leman, S., Guttman, D.S., Setubal, J.C., Vinatzer, B.A. 2010. PAMDB, A Multilocus Sequence Typing and Analysis Database and Website for Plant-Associated Microbes. Phytopathology 100, 208–215. doi:10.1094/PHYTO-100-3-0208

Anwar, N., Abaydulla, G., Zayadan, B., Abdurahman, M., Hamood, B., Erkin, R., Ismayil, N., Rozahon, M., Mamtimin, H., Rahman, E. 2016. Pseudomonas populi sp. nov., an endophytic bacterium isolated from Populus euphratica. International Journal of Systematic and Evolutionary Microbiology 66, 1419–1425. doi:10.1099/ijsem.0.000896

Astua-Monge, G., Minsavage, G.V., Stall, R.E., Vallejos, C.E., Davis, M.J., Jones, J.B., 2000. Xv4-vrxv4: A New Gene-for-Gene Interaction Identified Between Xanthomonas campestris pv. vesicatoria Race T3 and the Wild Tomato Relative Lycopersicon pennellii. MPMI 13, 1346–1355. doi:10.1094/MPMI.2000.13.12.1346

Auch, A.F., von Jan, M., Klenk, H.-P., Göker, M. 2010. Digital DNA-DNA hybridization for microbial species delineation by means of genome-to-genome sequence comparison. Standards in Genomic Sciences 2, 117–134. doi:10.4056/sigs.531120

Ausubel, F.M., Brent, R., Kingston, R.E., Moore, D.D., Seidman, J.G., Smith, J.A., and Struhl, K. 1994. Current protocols in molecular biology. New York: John Wiley and Sons.

Aysan, Y., Sahin, S., Ulke, G., and Sahin, F. 2003. Bacterial rot of lettuce caused by Pseudomonas cichorii in Turkey. Plant Pathology 52:782-

Bartoli, C., Berge, O., Monteil, C.L., Guilbaud, C., Balestra, G.M., Varvaro, L., Jones, C., Dangl, J.L., Baltrus, D.A., Sands, D.C., Morris, C.E. 2014. The Pseudomonas viridiflava phylogroups in the P. syringae species complex are characterized by genetic variability and phenotypic plasticity of pathogenicity-related traits. Environ Microbiol 16, 2301–2315. doi:10.1111/1462-2920.12433

Berge, O., Monteil, C.L., Bartoli, C., Chandeysson, C., Guilbaud, C., Sands, D.C., and Morris, C.E. 2014. A user’s guide to a database of the diversity of Pseudomonas syringae and its application to classifying strains in this phylogenetic comples. PLos One. doi: 10.1371/journal.pone.0105547.

114

Bratcher, H.B., Corton, C., Jolley, K.A., Parkhill, J., and Maiden, M.C.J. 2014. A gene- by-gene population genomics platform: de novo assembly, annotation and genealogical analysis of 108 representative Neisseria meningitis genomes. BMC Genomics. 15:1138

Bouzar, H., Jones, J.B., Stall, R.E., Hodge, N.C., Minsavage, G.V., Benedict, A.A., and Alvarez, A.M. 1994. Physiological, chemical, serological, and pathogenic analyses of a worldwide collection of Xanthomonas campestris pv. vesicatoria strains. Phytopathology 84:663-671.

Bruen, T.C., Philippe, H., and Bryant, D. 2006. A simple and robust statistical test for detecting the presence of recombination. Genetics. 172:2665-2681.

Bull, C.T., and Koike, S.T. 2015. Practical Benefits of Knowing the Enemy: Modern Molecular Tools for Diagnosing the Etiology of Bacterial Diseases and Understanding the Taxonomy and Diversity of Plant-Pathogenic Bacteria. Annual Review of Phytopathology 53, 157–180. doi:10.1146/annurev-phyto-080614- 120122

Bull, C.T., Clarke, C.R., Cai, R., Vinatzer, B.A., Jardini, T.M., Koike, S.T. 2011. Multilocus Sequence Typing of Pseudomonas syringae Sensu Lato Confirms Previously Described Genomospecies and Permits Rapid Identification of P. syringae pv. coriandricola and P. syringae pv. apii Causing Bacterial Leaf Spot on Parsley. Phytopathology 101, 847–858. doi:10.1094/PHYTO-11-10-0318

Burkholder, W. H. 1954. Bacteria pathogenic on head lettuce in New-York state. Phytopathology, 44(10), 592–596.

Burnham, K.P, and Anderson, D.R. 2004. Multimodel inference: understanding AIC and BIC in model selection. Sociological Methods and Research 33:201-208.

Canteros, B.I, Minsavage, G.V., Jones, J.B., and Stall, R.E. 1995. Diversity of plasmids in Xanthomonas campestris pv. vesicatoria. Phytopathology 85:1482-1486.

Chan, J.Z.-M., Halachev, M.R., Loman, N.J., Constantinidou, C., Pallen, M.J. 2012. Defining bacterial species in the genomic era: insights from the genus Acinetobacter. BMC Microbiology 12, 302. doi:10.1186/1471-2180-12-302

Chase, A. R. 1986. Comparisons of three bacterial leaf spots of Hibiscus rosasinensis. Plant Disease, 70, 334-336.

Chen, W.P., and Kuo, T.T. 1993. A simple and rapid method for the preparation of gram-negative bacterial genomic DNA. Nucleic Acids Res 21, 2260.

Cho, J.-C., and Tiedje, J.M. 2001. Bacterial Species Determination from DNA-DNA Hybridization by Using Genome Fragments and DNA Microarrays. Appl. Environ. Microbiol. 67, 3677–3682.

115

Cock, P.J.A., Antao, T., Chang, J.T., Chapman, B.A., Cox, C.J., Dalke, A., Friedberg, I., Hamelryck, T., Kauff, F., Wilczynski, B., Hoon, M.J.L. de, 2009. Biopython: freely available Python tools for computational molecular biology and bioinformatics. Bioinformatics 25, 1422–1423. doi:10.1093/bioinformatics/btp163

Contreras-Moreira, B., Vinuesa, P., 2013. GET_HOMOLOGUES, a Versatile Software Package for Scalable and Robust Microbial Pangenome Analysis. Appl Environ Microbiol 79, 7696–7701. doi:10.1128/AEM.02411-13

Cottyn, B., Heylen, K., Heyrman, J., Vanhouteghem, K., Pauwelyn, E., Bleyaert, P., Baerenbergh, J.V., Höfte, M., De Vos, P., and Maes, M. 2009. Pseudomonas cichorii as the causal agent of midrib rot, an emerging disease of greenhouse- grown butterhead lettuce in Flanders. Systematic and applied microbiology 22: 211-225

Darriba, D., Taboada, G.L., Doallo, R. and Posada, D. 2012. jModelTest 2: more models, new heuristics and parallel computing. Nature methods 9(8):772. de Been, M., Pinholt, M., Top, J., Bletz, S., Mellmann, A., van Schaik, W., Brouwer, E., Rogers, M., Kraat, Y., Bonten, M., Corander, J., Westh, H., Harmsen, D., Willems, R.J.L. 2015. Core Genome Multilocus Sequence Typing Scheme for High- Resolution Typing of Enterococcus faecium. J. Clin. Microbiol. 53, 3788– 3797. doi:10.1128/JCM.01946-15

Deshmukh, A.M. 1997. Handbook of Media, Stains and Reagents in Microbiology. Pama Publication.

Didelot, X., Wilson, D.J., 2015. ClonalFrameML: Efficient Inference of Recombination in Whole Bacterial Genomes. PLOS Comput Biol 11, e1004041. doi:10.1371/journal.pcbi.1004041

Edgar, R.C. 2004. MUSCLE: multiple alignment with high accuracy and high throughput. Nucleic Acids. Res. 32(5): 1792-1797.

Forsythe, S.J., Dickins, B., Jolley, K.A., 2014. Cronobacter, the emergent bacterial pathogen Enterobacter sakazakii comes of age; MLST and whole genome sequence analysis. BMC Genomics 15, 1121. doi:10.1186/1471-2164-15-1121

Freeman, J.H., Dittmar, P.J., and Vallad, G.E. 2016. Commerical Vegetable Production in Florida. Vegetable Production Handbook of Florida 2016-17. (Available at: http://edis.ifas.ufl.edu/pdffiles/cv/cv29200.pdf) (Verified on: 23 Sept 2016)

Garrity, G. 2006. Bergey’s Manual® of Systematic Bacteriology: Volume Two: The Proteobacteria, Part A Introductory Essays. Springer Science & Business Media.

116

Gassman, W., Dahlbeck, D., Chesnokova, O., Minsavage, G.V., Jones, J.B., and Staskawicz, B.J. 2000. Molecular evolution of virulence in natural field strains of Xanthomonas campestris pv. vesicatoria. Journal of Bacteriology 182: 7053- 7059.

Gevers, D., Cohan, F.M., Lawrence, J.G., Spratt, B.G., Coenye, T., Feil, E.J., Stackebrandt, E., Van de Peer, Y., Vandamme, P., Thompson, F.L., and Swings, J. 2005. Re-evaluating prokaryotic species. Nature Reviews Microbiology 3, 733– 739.

Gomila, M., Peña, A., Mulet, M., Lalucat, J., and García-Valdés, E. 2015. Phylogenomics and systematics in Pseudomonas. Front Microbiol 6.

González, A.J., Cleenwerck, I., De Vos, P., and Fernández-Sanz, A.M. 2013. Pseudomonas asturiensis sp. nov., isolated from soybean and weeds. Systematic and Applied Microbiology 36, 320–324.

González, A.J., Rodicio, M.R., and Mendoza, M.C. 2003. Identification of an Emergent and Atypical Pseudomonas viridiflava Lineage Causing Bacteriosis in Plants of Agronomic Importance in a Spanish Region. Appl. Environ. Microbiol. 69, 2936– 2941.

Guindon, S., Dufayard, J.F., Lefort, V., Anisimova, M., Hordijk, W., and Gascuel, O. 2010. New algorithms and methods to estimate maximum-likelihood phylogenies: assessing the perforamce of PhyML 3.0. Systematic Biology. 59:307-321.

Gumtow, R. L., Khan, A. A., Bocsanczy, A. M., Yuen, J. M., Palmateer, A. J., and Norman, D. J. 2013. First Report of a Leaf Spot Disease of Golden Dewdrop (Duranta erecta) Caused by Pseudomonas cichorii and a Xanthomonas Species in Florida. Plant Disease, 97(6), 836–836.

Gururani, M.A., Venkatesh, J., Upadhyaya, C.P., Nookaraju, A., Pandey, S.K., and Park, S.W. 2012. Plant disease resistance genes: Current status and future directions. Physiological and Molecular Plant Pathology 78:51-65.

Hamza, A.A., Robene-Soustrade, I., Jouen, E., Lefeuvre, P., Chiroleu, F., Fisher-Le Saux, M., Gagnevin, L., Pruvost, O. 2012. MultiLocus Sequence Analysis- and Amplified Fragment Length Polymorphism-based characterization of xanthomonads associated with bacterial spot of tomato and pepper and their relatedness to Xanthomonas species. Systematic and Applied Microbiology 35, 183–190. doi:10.1016/j.syapm.2011.12.005

Hert, A. P., Roberts, P.D., Momol, M.T., Minsavage, G.V., Tudor-Nelson, S.M., and Jones, J.B. 2005. Relative Importance of Bacteriocin-Like Genes in Antagonism of Xanthomonas perforans Tomato Race 3 to Xanthomonas euvesicatoria Tomato Race 1 Strains. Appl. Environ. Microbiol. 71, 3581–3588

117

Horvath, D.M., Stall, R.E., Jones, J.B., Pauly, M.H., Vallad, G.E., Dahlbeck, D., Staskawicz, B.J., Scott, J.W. 2012. Transgenic Resistance Confers Effective Field Level Control of Bacterial Spot Disease in Tomato. PLOS ONE 7, e42036. doi:10.1371/journal.pone.0042036

Huang, C.-H., G.E. Vallad, S. Zhang, A. Wen, B. Balogh, J.F. Figueiredo, J.B. Jones, T. Momol, S. Olson. 2012. The effect of application frequency and reduced rates of acibenzolar-S-methyl on the field efficacy of induced resistance against bacterial spot of tomato. Plant Disease. 96:221-227.

Hudson, R.R., and Kaplan, N.L. 1985. Statistical properties of the number of recombination events in the history of a sample of DNA sequences. Genetics 111:147-164.

Huerta-Cepas, J., Serra, F., and Bork, P. 2016. ETE 3: Reconstruction, analysis, and visualization of phylogenomic data. Molecular Biology and Evolution. 33:1635- 1638.

Huson, D.H., and Bryant, D. 2006. Application of phylogenetic networks in evolutionary studies. Molecular Biology and Evolution. 23:254-267.

Hwang, M. S., Morgan, R. L., Sarkar, S. F., Wang, P. W., and Guttman, D. S. 2005. Phylogenetic characterization of virulence and resistance phenotypes of Pseudomonas syringae. Applied and Environmental Microbiology, 71(9), 5182– 5191.

Jibrin, M.O., Timilsina, S., Potnis, N., Minsavage, G.V., Shenge, K.C., Akpa, A.D., Alegbejo, M.D., Beed, F., Vallad, G.E., Jones, J.B. 2015. First Report of Atypical Xanthomonas euvesicatoria Strains Causing Bacterial Spot of Tomato in Nigeria. Plant Disease 99, 415–415. doi:10.1094/PDIS-09-14-0952-PDN

Jones, J. B., Engelhard, A. W., and Raju, B. C. 1983. Outbreak of a stem necrosis on chrysanthemum incited by Pseudomonas cichorii in Florida. Plant Disease, 67(4), 431–433.

Jones, J. B., Jones, J. P., McCarter, S. M., and Stall, R. E. 1984. Pseudomonas viridiflava: causal agent of bacterial leaf blight of tomato. Plant Disease, 68(4), 341–342.

Jones, J.B., Lacy, G.H., Bouzar, H., Stall, R.E., and Schaad, N.W. 2004. Reclassification of the xanthomonads associated with bacterial spot disease of tomato and pepper. Systematic and Applied Microbiology 27: 755-762.

Jones, J.D.G. and Dangl, J.L. 2006. The plant immune system. Nature 444:323-329.

Kearney, B., and Staskawicz, B.J. 1990. Widespread distribution and fitness contribution of Xanthomonas campestris avirulence gene avrBs2. Nature 346: 385-386.

118

Kearse, M., Moir, R., Wilson, A., Stones-Havas, S., Cheung, M., Sturrock, S., Buxton, S., Cooper, A., Markowitz, S., Duran, C., Thierer, T., Ashton, B., Meintjes, P., Drummond, A., 2012. Geneious Basic: An integrated and extendable desktop software platform for the organization and analysis of sequence data. Bioinformatics 28, 1647–1649. doi:10.1093/bioinformatics/bts199

Kebede, M., Timilsina, S., Ayalew, A., Admassu, B., Potnis, N., Minsavage, G. V., Goss, E.M., Hong, J.C., Strayer, A., Paret, M., Jones, J.B., and Vallad, G. E. 2014. Molecular characterization of Xanthomonas strains responsible for bacterial spot of tomato in Ethiopia. European Journal of Plant Pathology, 140(4), 677-688.

Kim, J.G., Li, X., Roden, J.A., taylor, K.W., Aakre, C.D., Su, B., Lalonde, S., Kirik, A;, Chen, Y., Baranage, G., McLane, H., Martin, G.B. and Mudgett, M.B. 2009. Xanthomonas T3S effector XopN suppresses PAMP-triggered immunity and interacts with a tomato atypical receptor-like kinase and TFT1. Plant Cell 21:1305-1323.

Kim, N.H, Choi, H.W., and Hwang, B.K. 2010. Xanthomonas campestris pv. vesicatoria effector avrBsT induces cell death in pepper, but suppresses defense responses in tomato. Molecular Plant-Microbe Interactions 23: 1069-1082.

Kim, N.H., and Hwang, B.K. 2015. Pepper heat shock protein 70a interacts with the type III effector AvrBsT and triggers plant cell death and immunity. Plant Physiology 167:307-322.

King, E.O., Ward, M.K., and Raney, D.E., 1954. Two simple media for the demonstration of pyocyanin and fluorescin. J. Lab. Clin. Med. 44, 301–307.

Kohl, T.A., Diel, R., Harmsen, D., Rothgänger, J., Walter, K.M., Merker, M., Weniger, T., and Niemann, S. 2014. Whole-Genome-Based Mycobacterium tuberculosis surveillance: a standardized, portable, and expandable approach. Journal of Clinicial Microbiology, 52(7): 2479-2486.

Konstantinidis, K.T., Ramette, A., Tiedje, J.M. 2006. The bacterial species definition in the genomic era. Philos Trans R Soc Lond B Biol Sci 361, 1929–1940. doi:10.1098/rstb.2006.1920

Kwong, J.C., Mccallum, N., Sintchenko, V., Howden, B.P., 2015. Whole genome sequencing in clinical and public health microbiology. Pathology 47, 199–210. doi:10.1097/PAT.0000000000000235

Lamichhane, J.R., Messéan, A., and Morris, C.E. 2015. Insights into epidemiology and control of diseases of annual plants caused by the Pseudomonas syringae species complex. J Gen Plant Pathol 81, 331–350.

Laue, H., Schenk, A., Li, H., Lambertsen, L., Neu, T.R., Molin, S., and Ullrich, M.S. 2006. Contribution of alginate and levan production to biofilm formation by Pseudomonas syringae. Microbiology 152, 2909–2918.

119

Leekitcharoenphon, P., Nielsen, E.M., Kaas, R.S., Lund, O., Aarestrup, F.M., 2014. Evaluation of Whole Genome Sequencing for Outbreak Detection of Salmonella enterica. PLoS One 9. doi:10.1371/journal.pone.0087991

Lelliott, R. A., Billing, E., and Hayward, A. C. 1966. A determinative scheme for the fluorescent plant pathogenic Pseudomonads. Journal of Applied Bacteriology, 29(3), 470–489.

Leopold, S.R., Goering, R.V., Witten, A., Harmsen, D., and Mellmann, A. 2014. Bacterial whole-genome sequencing revisited: portable, scalable, and standardized analysis for typing and detection of virulence and antibiotic resistance genes. Journal of Clinical Microbiology. 52:2365-2370.

Maiden, M.C.J., Harrison, O.B., 2016. The population and functional genomics of the Neisseria revealed with gene-by-gene approaches. J. Clin. Microbiol. JCM.00301-16. doi:10.1128/JCM.00301-16

Marcelletti, S., and Scortichini, M. 2014. Definition of Plant-Pathogenic Pseudomonas Genomospecies of the Pseudomonas syringae Complex Through Multiple Comparative Approaches. Phytopathology 104, 1274–1282.

Marchesi, J. R., Sato, T., Weightman, A. J., Martin, T. A., Fry, J. C., Hiom, S. J., and Wade, W. G. 1998. Design and evaluation of useful bacterium-specific PCR primers that amplify genes coding for bacterial 16S rRNA. Applied and Environmental Microbiology, 64(2), 795–799.

Marchi, G., Viti, C., Giovannetti, L., and Surico, G. 2005. Spread of levan-positive populations of Pseudomonas savastanoi pv. savastanoi, the causal agent of olive knot, in central Italy. European Journal of Plant Pathology, 112(2), 101–112.

Markowitz, V.M., Chen, I.-M.A., Palaniappan, K., Chu, K., Szeto, E., Grechkin, Y., Ratner, A., Jacob, B., Huang, J., Williams, P., Huntemann, M., Anderson, I., Mavromatis, K., Ivanova, N.N., Kyrpides, N.C., 2012. IMG: the integrated microbial genomes database and comparative analysis system. Nucleic Acids Res 40, D115–D122. doi:10.1093/nar/gkr1044

McCarter, S.M., Jones, J.B., Gitaitis, R.D., and Smitley, D.R.1983. Survival of Pseudomonas syringae pv. tomato in Association with Tomato Seed, Soil, Host Tissue, and Epiphytic Weed Hosts in Georgia. Phytopathology 73, 1393. doi:10.1094/Phyto-73-1393

McDonald, B.A., and Linde, C. 2002. Pathogen population genetics, evolutionary potential, and durable resistance. Annual Review of Phytopathology 40:349-379.

Meier-Kolthoff, J.P., Auch, A.F., Klenk, H.-P., and Göker, M. 2013. Genome sequence- based species delimitation with confidence intervals and improved distance functions. BMC Bioinformatics 14, 60.

120

Mhedbi-Hajri, N., Hajri, A., Boureau, T., Darrasse, A., Durand, K., Chrystelle, B., Saux, M.F., Manceau, C., poussier, S., Pruvost, O., Lemaire, C., Jacques, M.A. 2013. Evolutionary history of the plant pathogenic bacterium Xanthomonas axonopodis. PLOS one. doi: 10.1371/journal.pone.0058474.

Mirik, M., Aysan, Y., and Sahin, F. 2011. Characterization of Pseudomonas cichorii isolated from different hosts in Turkey. International Journal of Agriculture and Biology, 13, 203–209.

Momol, T., Jones, J.B., Olson, S., Obradovic, A., Balogh, B., and King, P. 2008. Integrated Management of Bacterial Spot on Tomato in Florida. University of Florida, IFAS Extension.

Morohoshi, T., Kato, T., Someya, N., and Ikeda, T. 2014. Complete genome sequence of N-Acylhomoserine Lactone-producing Pseudomonas sp. Strain StFLB209, isolated from potato phyllosphere. Genome Announcements. doi: 10.1128/genomeA.01037-14

Myung, I.-S., Choi, J.-K., Lee, J. Y., Yoon, M.J., Hwang, E. Y., and Shim, H. S. 2013. First Report of Bacterial Leaf Spot of Witloof, Caused by Pseudomonas cichorii in Korea. Plant Disease, 97(10), 1376–1376.

Newberry, E.A., Paret, M.L., Jones, J.B., and Bost, S.C. 2016. First report of leaf spot of pumpkin caused by Pseudomoans cichorii in Tennessee. Plant Disease. doi: 10.1094/PDIS-11-15-1387-PDN

Nimchuk, X., Eulgem, T., Holt III, B.F., Dangi, J.L. 2003. Recognition and response in the plant immune system. Annual Reviews of Genetics 37:579-609.

Palleroni, N.J. 1981. Introduction to the Family Pseudomonadaceae. In The Prokaryotes, M.P. Starr, H. Stolp, H.G. Trüper, A. Balows, and H.G. Schlegel, eds. Springer Berlin Heidelberg, pp. 655–665.

Palleroni, N.J. 2005. Pseudomonas. In S.P. borriello, P.R. Murray, and G. Funke eds., Topley and Wilson’s Microbiology and Microbial Infections. Bacteriology. Hodder Arnold, London. Pp. 1591-1605.

Palleroni, N. J. 2008. The road to the taxonomy of Pseudomonas. Pseudomonas: Genomics and Molecular Biology, 1–18.

Palleroni, N. J. 2010. The Pseudomonas story. Environmental Microbiology, 12(6), 1377–1383.

Parte, A.C. 2014. LPSN—list of prokaryotic names with standing in nomenclature. Nucleic Acids Res 42, D613–D616.

121

Paulsen, I.T., Press, C.M., Ravel, J., Kobayashi, D.Y., Myers, G.S.A., Mavrodi, D.V., DeBoy, R.T., Seshadri, R., Ren, Q., Madupu, R., et al. 2005. Complete genome sequence of the plant commensal Pseudomonas fluorescens Pf-5. Nat Biotech 23, 873–878.

Pohronezny, K., Sommerfeld, M. L., and Raid, R. N. 1994. Streptomycin resistance and copper tolerance among strains of Pseudomonas cichorii in celery seedbeds. Plant Disease, 78(2), 150–153.

Posada, D., 2008. jModelTest: Phylogenetic Model Averaging. Mol Biol Evol 25, 1253– 1256. doi:10.1093/molbev/msn083

Potnis, N., Krasileva, K., Chow, V., Almeida, N.F., Patil, P.B., Ryan, R.P., Sharlach, M., Behlau, F., Dow, J.M., Momol, M.T., White, F.F., Preston, J.F., Vinatzer, B.A., Koebnik, R., Setubal, J.C., Norman, D.J., Staskawicz, B.J., and Jones, J.B. 2011. Comparative genomics reveals diversity among xanthomonads infecting tomato and pepper. BMC Genomics 12:146

Potnis, N., Timilsina, S., Strayer, A., Shantharaj, D., Barak, J. D., Paret, M. L., Vallad, G.E., and Jones, J. B. 2015. Bacterial spot of tomato and pepper: diverse Xanthomonas species with a wide variety of virulence factors posing a worldwide challenge. Molecular plant pathology.

Ramkumar, G., Lee, S.W., Weon, H.-Y., Kim, B.-Y., and Lee, Y.H. 2014. First report on the whole genome sequence of Pseudomonas cichorii strain JBC1 and comparison with other Pseudomonas species. Plant Pathology 64:63-70.

Richter, M., and Rosselló-Móra, R. 2009. Shifting the genomic gold standard for the prokaryotic species definition. Proc Natl Acad Sci U S A 106, 19126–19131.

Robinson, D.F., and Foulds, L.R. 1981. Comparison of phylogenetic trees. Mathematical Biosciences. 53:131-147.

Rodriguez-R, L.M., and Konstantinidis, K.T. 2014. Bypassing Cultivation To Identify Bacterial Species: Culture-independent genomic approaches identify credibly distinct clusters, avoid cultivation bias, and provide true insights into microbial species. Microbe Magazine 9, 111–118.

Sarkar, S. F., and Guttman, D. S. 2004. Evolution of the core genome of Pseudomonas syringae, a highly clonal, endemic plant pathogen. Applied and Environmental Microbiology, 70(4), 1999–2012.

122

Schwartz, A.R., Potnis, N., Timilsina, S., Wilson, M., Patané, J., Martins, J., Minsavage, G.V., Dahlbeck, D., Akhunova, A., Almeida, N., Vallad, G.E., Barak, J.D., White, F.F., Miller, S.A., Ritchie, D., Goss, E., Bart, R.S., Setubal, J.C., Jones, J.B., Staskawicz, B.J. 2015. Phylogenomics of Xanthomonas field strains infecting pepper and tomato reveals diversity in effector repertoires and identifies determinants of host specificity. Xanthomonas 6, 535. doi:10.3389/fmicb.2015.00535

Scortichini, M., Marchesi, U., Rossi, M. P., and Di Prospero, P. 2002. Bacteria associated with hazelnut (Corylus avellana L.) decline are of two groups: Pseudomonas avellanae and strains resembling P. syringae pv. syringae. Applied and Environmental Microbiology, 68(2), 476–484.

Scott, J.W., Jones, J.B., and Somodi, G.C. 1989. Inheritance of resistance in tomato to race T3 of the bacterial spot pathogen. Journal of American Society of Horticulture Science 126: 436-441.

Sharlach, M., Dahlbeck, D., Liu, L., Chiu, J., Jimenez-Gomez, J.M., Kimura, S., Koenig, D., Maloof, J.N., Sinha, N., Minsavage, G.V., Jones, J.B., Stall, R.E., and Staskawicz, B.J. 2013. Fine genetic mapping of RXopJ4, a bacterial spot disease resistance locus from Solanum pennellii LA716. Theor. Appl. Genet. 126:601- 609. 10.1007/s00122-012-2004-6

Soltis, D.E., and Soltis, P.S. 2003. The role of phylogenetics in comparative genetics. Plant Physiology. 132:1790-1800.

Someya, N., Morohoshi, T., Okano, N., Otsu, E., Usuki, K., Sayama, M., Sekiguchi, H., Ikeda, T., and Ishida, S. 2009. Distribution of N-Acylhomoserine Lactone- producing fluorescent Pseudomonas in the phyllosphere and rhizosphere of Potato (Solanum tuberosum L.). Microbes and Environments. 24:305-314.

Staley, J.T. 2010. The phylogenomic species concept for Bacteria and Archaea. Issues.

Stall, R.E., Jones, J.B., Minsavage, G.V., 2009. Durability of Resistance in Tomato and Pepper to Xanthomonads Causing Bacterial Spot. Annual Review of Phytopathology 47, 265–284. doi:10.1146/annurev-phyto-080508-081752

Stead, D.E., Simpkins, S.A., Weller, S.A., Hennessy, J., Aspin, A., Stanford, H., Smith, N.C., and Elphinstone, J.G. 2013. Classification and identification of plant pathogenic Pseudomonas species by REP-PCR derived genetic fingerprints. In: Pseudomonas syringae and related pathogens. 411-420.

Strayer, A., Garcia-Maruniak, A., Sun, X., Schubert, T., and Sutton, B. 2012. First Report of Pseudomonas cichorii Causing Leaf Spot of Stevia Detected in Florida. Plant Disease, 96(11), 1690–1690.

123

Sun, X., M. C. Nielsen, and J. W. Miller. 2002. Bacterial spot of tomato and pepper. Plant Pathology Circular No. 129 (revised) [Online]. Florida Department of Agriculture & Conservation Services. (Available at: https://www.freshfromflorida.com/content/download/11136/143107/pp129rev.pdf) (verified 23 Sept 2016).

Swingle, D. B. 1925. Center rot of French endive or wilt of chicory (Cichorium intybus L.). Phytopathology, 15, 730.

Swords, K.M., Dahlbeck, D., Kearney, B., Roy, M., and Staskawicz, B.J. 1996. Spontaneous and induced mutations in a single open reading frame alter both virulence and avirulence in Xanthomonas campestris pv. vesicatoria avrBs2. Journal of Bacteriology 178: 4661-4669.

Tai, T.H., Dahlbeck, D., Clark, E.T., Gajiwala, P., Pasion, R., Whalen, M.C., Stall, R.E., and Staskawicz, B.J. 1999. Expression of the Bs2 pepper gene confers resistance to bacterial spot disease in tomato. Proceedings of the National Academy of Sciences of the United States of America 96:14153-14158.

Tamura, K., Stecher, G., Peterson, D., Filipski, A., and Kumar, S. 2013. MEGA6: Molecular evolutionary genetics analysis version 6.0. Molecular Biology and Evolution, 30(12), 2725–2729.

Teper, D., Salomon, D., Sunitha, S., Kim, J.G., Mudgett, M.B., and Sessa, G. 2014. Xanthomonas euvesicatoria type III effector XopQ interacts with tomato and pepper 14-3-3 isoforms to suppress effector-triggered immunity. The Plant Journal 77: 297-309.

Testen, A. L., Nahson, J., Mamiro, D. P., and Miller, S. A. 2015. First Report of Tomato Pith Necrosis Caused by Pseudomonas cichorii in Tanzania. Plant Disease, 99(7), 1035–1035.

Timilsina, S., Abrahamian, P., Potnis, N., Minsavage, G.V., White, F.F., Staskawicz, B.J., Jones, J.B., Vallad, G.E., Goss, E.M., 2016. Analysis of Sequenced Genomes of Xanthomonas perforans Identifies Candidate Targets for Resistance Breeding in Tomato. Phytopathology PHYTO-03-16-0119-FI. doi:10.1094/PHYTO-03-16-0119-FI

Timilsina, S., Jibrin, M.O., Potnis, N., Minsavage, G.V., Kebede, M., Schwartz, A., Bart, R., Staskawicz, B.J., Boyer, C., Vallad, G.E., Pruvost, O., Jones J.B., and Goss, E.M. 2015. Multilocus sequence analysis of xanthomonads causing bacterial spot of tomato and pepper plants reveals strains enerated by recombination among species and recent global spread of Xanthomonas gardneri. Appl. Environ. Microbiol. 81:1520-1529. 10.1128/AEM.03000-14

Trantas, E. A., Sarris, P. F., Mpalantinaki, E. E., Pentari, M. G., Ververidis, F. N., and Goumas, D. E. 2013. A new genomovar of Pseudomonas cichorii, a causal agent of tomato pith necrosis. European Journal of Plant Pathology, 137(3), 477–493.

124

Tudor-Nelson, S. M., Minsavage, G. V., Stall, R. E., and Jones, J. B. 2003. Bacteriocin- like substances from tomato race 3 strains of Xanthomonas campestris pv. vesicatoria. Phytopathology, 93(11), 1415-1421.

Ueda, K., Seki, T., Kudo, T., Yoshida, T., and Kataoka, M. 1999. Two distinct mechanisms cause heterogeneity of 16S rRNA. Journal of Bacteriology, 181(1), 78–82.

Vaidya, G., Lohman, D.J., Meier, R., 2011. SequenceMatrix: concatenation software for the fast assembly of multi-gene datasets with character set and codon information. Cladistics 27, 171–180. doi:10.1111/j.1096-0031.2010.00329.x

Vallad, G.E., S. Timilsina, H. Adkison, N. Potnis, G. Minsavage, J. Jones, and E. Goss. 2013. A recent survey of Xanthomonads causing bacterial spot of tomato in Florida provides insights into management strategies. Proceedings from the 2013 Florida Tomato Institute, Naples, FL.

VanSickle, J. and R. Weldon. 2009. The economic impact of bacterial leaf spot on the tomato industry. 2009 Proceedings of the Florida Tomato Institute. p. 30–31. von Neubeck, M., Huptas, C., Glück, C., Krewinkel, M., Stoeckel, M., Stressler, T., Fischer, L., Hinrichs, J., Scherer, S., and Wenning, M. 2016. Pseudomonas helleri sp. nov. and Pseudomonas weihenstephanensis sp. nov., isolated from raw cow’s milk. International Journal of Systematic and Evolutionary Microbiology 66, 1163–1173.

Wang, J.F. 1992. A new race of Xcv on tomato and screening tomato genotypes for resistance to this new race. Resistance to Xanthomonas campestris pv. vesicatoria in tomato. Ph.D. dissertation. University of Florida, Gainesville, FL.

White, F.F., Potnis, N., Jones, Jeffrey, B., and Koebnik, R. 2009. The type III effectors of Xanthomonas. Molecular Plant Pathology. 10:749-766.

Wilkie, J. P., and Dye, D. W. 1974. Pseudomonas cichorii causing tomato and celery diseases in New Zealand. New Zealand Journal of Agricultural Research, 17(2), 123–130.

Willems, S., Kampmeier, S., Bletz, S., Kossow, A., Köck, R., Kipp, F., Mellmann, A., 2016. Whole-Genome Sequencing Elucidates Epidemiology of Nosocomial Clusters of Acinetobacter baumannii. J. Clin. Microbiol. 54, 2391–2394. doi:10.1128/JCM.00721-16

Yamamoto, S., Kasai, H., Arnold, D.L., Jackson, R.W., Vivian, A., and Harayama, S. 2000. Phylogeny of the genus Pseudomonas: intrageneric structure reconstructed from the nucleotide sequence of gyrB and rpoD genes. Microbiology 146:2385-2394.

125

Young, J.M., Park, D.-C., Shearman, H.M., Fargier, E., 2008. A multilocus sequence analysis of the genus Xanthomonas. Systematic and Applied Microbiology 31, 366–377. doi:10.1016/j.syapm.2008.06.004

126

BIOGRAPHICAL SKETCH

Sujan Timilsina was born in Pokhara, Kaski district, Nepal. He attended Little

Step Higher Secondary School in Pokhara and later graduated from high school in 2006 from Nobel Academy, Kathmandu. He completed his undergraduate degree in

Agriculture (B.Sc.Ag) with a major in Plant Pathology from the Institute of Agriculture and Animal Sciences, Tribhuvan University, Nepal in 2010. He worked briefly as an agronomist for the non-governmental organization SAPPROS (Support Activities for

Poor Producers of Nepal) on food for work project funded by World Food Program

(WFP). In 2012, he joined the University of Florida (UF) for his Master of Science in plant pathology with Dr. Gary E. Vallad and Dr. Jeffrey B. Jones. His master’s research was titled “Phenotypic and genotypic variation in bacterial spot causing xanthomonads”.

Upon graduation in December 2013, Sujan continued to work with Dr. Gary E. Vallad and Dr. Jeffrey B. Jones and started his Ph.D. in spring 2014, in which he conducted research on Pseudomonads and Xanthomonads and received his Ph.D. from the

University of Florida in fall 2016.

127