University of Florida Thesis Or Dissertation Formatting

EVOLUTION OF PLANT PATHOGENICITY IN STREPTOMYCES

YUCHENG ZHANG

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 for their unconditional love and support

ACKNOWLEDGMENTS

I would like to express my gratitude to Dr. Rosemary Loria, my advisor, for her guidance in my research. She always encouraged me to be a passionate and independent researcher and gave me freedom to explore new research fields. During the last three years in Dr. Loria’s lab, I have learned microbiology and biotechnology skills, have developed the habits of logical thinking and proper questioning. From Dr. Loria, I also learned the importance of networking and collaboration for a successful career in academia.

I would like to thank Drs. Valerie de Crecy-Lagard, Jeffrey B. Jones, and G. Shad Ali for serving on my committee, and for providing their time and expertise. Their advice and recommendations are valuable resources for my projects, and for the improvement of my dissertation.

I am grateful to the past members of the Loria lab. Your friendship and support let me feel very happy in this international family. I would like to express my special thanks to Drs.

Isolde Francis and Jose C. Huguet Tapia for kindly teaching me Streptomyces techniques and bioinformatics.

I would like to thank Dr. Yousong Ding and his student Mr. Ran Zuo for helpful collaborations, particularly HPLC and LC-MS analysis.

Last but most importantly, I deeply thank my parents (Zhongle Zhang and Tongling Liu), my parents-in-law (Yingshun Qiu and Shuqin Zhao), my wife Sai Qiu, and my adorable daughter

Flora Zhang for their unconditional support and encouragement. They are the most important motivation for pursuing my academic and career goals.

TABLE OF CONTENTS

page

ACKNOWLEDGMENTS ...... 4

LIST OF TABLES ...... 8

LIST OF FIGURES ...... 9

ABSTRACT ...... 11

CHAPTER

1 GENERAL INTRODUCTION ...... 13

The Streptomyces Genus ...... 13 Streptomyces Genomes ...... 13 Plant Pathogenic Streptomyces Species ...... 14 Streptomyces scabiei (syn. S. scabies) ...... 15 Streptomyces acidiscabies ...... 15 Streptomyces turgidiscabies ...... 16 Streptomyces ipomoeae ...... 16 Streptomyces europaeiscabiei ...... 16 Streptomyces stelliscabiei ...... 16 Streptomyces reticuliscabiei ...... 17 Streptomyces caviscabies ...... 17 Streptomyces luridiscabiei, Streptomyces puniciscabiei, and Streptomyces niveiscabiei ...... 17 Pathogenicity and Virulence Factors ...... 17 Thaxtomins ...... 17 Nec1 ...... 19 TomA ...... 20 Fas Operon ...... 21 Coronafacic Acid ...... 21 Concanamycins ...... 22 FD-891 ...... 22 Borrelidin ...... 23 Other Unknown Pathogenicity/virulence Factors ...... 23 Integrative and Conjugative Elements ...... 23 Pathogenicity Islands of Streptomyces ...... 26 The PAI in S. turgidiscabies Car8 ...... 26 The Toxicogenic Region and Colonization Region in Other Pathogenic Species ...... 26 Goals and Objectives of This Study ...... 27

2 PHYLOGENY OF THE STREPTOMYCES GENUS ...... 29

Introduction ...... 29

Materials and methods ...... 31 Bacterial Strains, Culture Conditions and DNA Extraction ...... 31 Virulence Bioassays ...... 32 Analysis of Thaxtomin A Production ...... 32 Genome Sequencing and Assembly ...... 32 Ortholog Retrieval and Single-copy Cluster Search ...... 33 Gene Alignment and Filtering Criteria ...... 33 Maximum Likelihood (ML) Phylogeny Estimation ...... 33 Average Nucleotide Identity and in silico DNA-DNA Hybridization Estimation ...... 34 Results...... 34 Characterization of Pathogenic Streptomyces Isolates ...... 34 Two Copies of Selected Housekeeping Genes Were Detected in Some Streptomyces Genomes ...... 35 Phylogenomic Analysis of the Streptomyces Genus ...... 36 isDDH and ANI Comparison of Plant Pathogenic Streptomyces Isolates ...... 38 Discussion ...... 39

3 PROMISCUOUS PATHOGENICITY ISLANDS OF STREPTOMYCES SPP...... 51

Materials and Methods ...... 53 Phylogeny of the Thaxtomin Biosynthetic Cluster ...... 53 Insertion of Antibiotic Resistance Markers in the PAI ...... 53 Selection of Transconjugant Strains ...... 54 Virulence Bioassays ...... 55 Analysis of Thaxtomin A Production ...... 55 Results...... 56 Streptomyces Pathogens Contained Different Versions of PAIs ...... 56 ThxA-phylogeny Contradicts the Core-genome Phylogeny ...... 57 The S. scabiei PAI was Site-specifically Mobilized into the Chromosome of Non- pathogenic Species ...... 57 PAI Acquisition Conferred a Pathogenic Phenotype in S. diastatochromogenes Transconjugants ...... 59 Discussion ...... 60

4 SITE-SPECIFIC ACCRETION OF THE THAXTOMIN A BIOSYNTHETIC CLUSTER WITH AN INTEGRATIVE CONJUGATIVE ELEMENT LEADS TO CIS- MOBILIZATION AND CONJUGATIVE TRANSFER ...... 70

Introduction ...... 70 Materials and Methods ...... 73 Bacterial Strains and Culture Conditions ...... 73 Virulence Bioassays ...... 73 Analysis of Thaxtomin A Production ...... 73 Genome Sequencing and Assembly ...... 74 Ortholog Retrieval and Single-copy Cluster Search ...... 74 Gene Alignment and Filtering Criteria ...... 74 Maximum Likelihood Phylogeny Estimation ...... 75

Average Nucleotide Identity and in silico DNA-DNA Hybridization Estimation ...... 75 Selection of Transconjugant Strains ...... 75 Results...... 76 Four S. acidiscaibies Isolates Were Selected for Genome Sequencing ...... 76 Phylogenomic and Genome Similarity Analysis of Pathogenic and Related Non- pathogenic Species ...... 77 PAI Variation Exists within Pathogenic Streptomyces Species ...... 78 S. acidiscabies, S. niveiscabiei and S. sp. 96-12 Acquired TxtA Biosynthetic Genes Through a Recent HGT Event ...... 79 Site-specific Accretion Between TR1 and TR2 ...... 80 Conjugative Mobilization of S. acidiscabies TR1 in cis by S. scabiei TR2 ...... 81 TR1 is fixed in pathogenic strains lacking TR2 ...... 82 Discussion ...... 82

5 A RE-EVALUATION OF THE TAXONOMY OF PHYTOPATHOGENIC GENERA DICKEYA AND PECTOBACTERIUM USING WHOLE-GENOME SEQUENCING DATA ...... 100

Materials and Methods ...... 103 Genome Collection ...... 103 Ortholog Retrieval and Single-copy Ortholog Search ...... 104 Gene Alignment and Filtering Criteria ...... 104 Phylogenomic Analysis ...... 104 Average Nucleotide Identity (ANI) and in silico DNA-DNA Hybridization (isDDH) Analysis...... 105 Results...... 105 Phylogenomic Analysis of Dickeya and Pectobacterium ...... 106 Assessment of Genome Similarity Using isDDH and ANI ...... 107 Discussion ...... 108

6 CONCLUSIONS AND PERSPECTIVES ...... 122

LIST OF REFERENCES ...... 126

BIOGRAPHICAL SKETCH ...... 144

LIST OF TABLES

Table page

2-1 Genome summary statistics of plant pathogenic Streptomyces spp. and nonpathogenic Streptomyces diastatochromogenes ...... 43

2-2 Genomes of non-pathogenic Streptomyces species used in this study ...... 44

2-3 The list of Streptomyces strains that have two-copies of MLSA genes ...... 46

3-1 Bacterial strains, plasmids, and cosmids used in this study ...... 63

3-2 List of primers used in this study ...... 63

4-1 List of primers used in this study ...... 88

4-2 Genome summary statistics of plant pathogenic Streptomyces spp...... 89

5-1 Features of genomes used in this study...... 113

5-2 Pairwise isDDH and ANI values for different genomospecies of Pectobacterium carotovorum ...... 117

LIST OF FIGURES

Figure page

2-1 Virulence phenotypes of plant pathogenic Streptomyces isolates...... 47

2-2 Phylogenetic relationships among nine Streptomyces strains based on gyrB gene sequences ...... 48

2-3 Phylogenetic reconstruction of the genus Streptomyces based on the concatenation of all 433 single-copy gene clusters from 111 Streptomyces genomes with the maximum likelihood algorithm RAxML...... 49

2-4 ANI and isDDH values of plant pathogenic and closely related non-pathogenic Streptomyces isolates ...... 50

3-1 Schematic genetic organization of PAIs from different plant pathogenic Streptomyces species...... 64

3-2 Phylogenetic reconstruction of scab-causing Streptomyces strains based on the thaxtomin A biosynthetic cluster (txtA, txtB, txtC, txtD, txtE and txtR)...... 65

3-3 Comparison of phylogenetic relationships derived from the core-genome (left) and the thaxtomin A biosynthetic cluster (right)...... 66

3-4 Excision and site-specific integration of the S. scabiei toxicogenic region (TR) into the S. diastatochromogenes chromosome...... 67

3-5 Primers used to detect the integration of S. scabiei toxicogenic region (TR) into the chromosome of S. diastatochromogenes...... 69

4-1 Primers used to detect the integration of the toxicogenic region (TR) into the chromosome of S. acidiscabies and S. diastatochromogenes...... 91

4-2 Virulence assays of pathogenic Streptomyces isolates...... 92

4-3 Phylogeny of pathogenic isolates and their related non-pathogenic Streptomyces species based on the concatenation of 1116 single-copy gene clusters with the maximum likelihood algorithm RAxML ...... 93

4-4 ANI and isDDH values of pathogenic Streptomyces isolates ...... 94

4-5 Schematic genetic organization of PAIs from different pathogenic Streptomyces isolates...... 95

4-6 Phylogenetic reconstruction of scab-causing Streptomyces strains based on the thaxtomin A biosynthetic cluster (txtA, txtB, txtC, txtD, txtE, txtH, and txtR)...... 96

4-7 Comparison of evolutionary relationships derived from the core-genome tree (Figure 4-3) and the TxtA tree (Figure 4-6)...... 97

4-8 Schematic representation of the mating experiments for investigation of the site- specific accretion and cis-mobilization of TR1 and TR2 ...... 97

4-9 Cartoon illustrating the site-specific accretion and cis-mobilization of TR1 and TR2 .....98

4-10 Schematic representation of a hypothesis for the evolution of TR1 and TR2 by deletion, site-specific accretion, and cis-mobilization ...... 99

5-1 The pipeline for phylogenomic and systematic analysis of the genera Dickeya and Pectobacterium ...... 118

5-2 Phylogenetic reconstruction of the genera Dickeya and Pectobacterium based on the concatenation of nucleic acid sequences of all 895 single-copy orthologs using the maximum likelihood algorithm RAxML ...... 119

5-3 Pairwise ANI and isDDH values of Pectobacterium atrosepticum, Pectobacterium betavasculorum, and Pectobacterium wasabiae ...... 120

5-4 Pairwise ANI and isDDH values of the genus Dickeya ...... 121

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

EVOLUTION OF PLANT PATHOGENICITY IN STREPTOMYCES

Yucheng Zhang

August 2016

Chair: Rosemary Loria Major: Plant Pathology

At least 11 Streptomyces species cause disease on underground plant structures. The most economically important of these is potato scab, and the most studied of these pathogenic species is Streptomyces scabiei (syn. S. scabies). The main pathogenicity determinant of scab-causing

Streptomyces species is a nitrated diketopiperazine, known as thaxtomin A (ThxA). In the pathogenic species Streptomyces turgidiscabies, ThxA biosynthetic genes reside on a mobile pathogenicity island (PAI). However, the mobilization of PAIs in other Streptomyces species remains uncharacterized. Here, we investigated the mobilization of the toxicogenic region (TR) of S. scabiei 87-22. One internal and two flanking attachment sites separate TR into two sub- regions, designated TR1 and TR2. TR1 contains ThxA biosynthetic genes, and TR2 includes putative integrative and conjugative elements (ICEs). Based on whole genome sequences, we inferred the evolutionary relationships of pathogenic Streptomyces species, and discovered that

Streptomyces sp. 96-12, a novel pathogenic species isolated from potatoes in Egypt, was phylogenetically grouped with non-pathogenic species rather than with known pathogenic species. We also found that S. sp. 96-12 contains a TR region that is almost identical to the TR region in S. scabiei, despite significant differences in their genome sequences. This suggested direct or indirect in vivo mobilization of TR between S. scabiei and nonpathogenic Streptomyces

species. To test whether S. scabiei 87-22 TR could indeed be mobilized, S. scabiei 87-22 deletion mutants containing antibiotic resistance markers in TR were mated with Streptomyces diastatochromogenes, a nonpathogenic species. S. scabiei TR site-specifically inserted into the aviX1 gene of S. diastatochromogenes and conferred pathogenicity in radish seedling assays. Our results demonstrated that S. scabiei could be the source of TR responsible for the emergence of novel pathogenic species. We further demonstrated that TR1 would be a non-autonomous cis- mobilizable element (CIME), and it hijacks the recombination and conjugation machinery of

TR2 to excise, transfer and integrate, leading to the dissemination of pathogenicity genes. In addition, our study revealed some genomes deposited in GenBank from the Streptomyces genus as well as other bacterial genera, such as Pectobacetrium and Dickeya, were misnamed. This work highlights the potential of using whole genome sequences to re-evaluate current prokaryotic species definition for taxonomically challenging genera.

CHAPTER 1 GENERAL INTRODUCTION

The Streptomyces Genus

Gram-positive bacteria are divided into two phyla, the Firmicutes (low G+C) and the

Actinobacteria (high G+C). These two phyla diverged very early from their last common ancestor in the evolution history (Ventura et al. 2007), indicating Firmicutes and Actinobacteria are highly unlikely to share ancestral pathogenicity factors (Hogenhout and Loria 2008).

Actinomycetes are distinguished by high G+C content in their DNA (often over 70%), and also their ability to form branching hyphae at certain growth stages of their development (Hopwood

2006). Most gram-positive plant pathogenic bacteria are actinomycetes (Eichenlaub and

Gartemann 2011), such as Clavibacter and Streptomyces.

Streptomyces spp. are filamentous actinobacteria that produce spores; they also are famous for producing biologically active secondary metabolites, including antibiotics (Challis and Hopwood 2003). Streptomyces has a complex lifecycle, which begins when hypha emerge from spores and produce a branched vegetative mycelium; upon nutritional and possibly other environmental stresses, Streptomyces will develop aerial hyphae that undergo cell division and form unigenomic spores which serve as the primary reproductive structure (Seipke et al. 2012).

The Streptomyces genus is the largest genus of Actinobacteria, including over 700 species

(http://www.bacterio.net/streptomycesa.html).

Streptomyces Genomes

The genome of Streptomyces coelicolor A3(2) was the first sequenced in the genus

(Bentley et al. 2002), and was a milestone in understanding chromosomal organization of

Streptomyces. The Streptomyces chromosome is linear, with a central “core” region possessing essential genes and terminal “arms” regions that have variable composition and organization

(Hopwood 2006). Analysis of the genome of S. coelicolor suggested significant horizontal gene transfer. Doroghazi and Buckley (2010) demonstrated that widespread homologous recombination occurred within and between Streptomyces species, and suggested that horizontal gene transfer and recombination may have significant impact in shaping the Streptomyces evolution (Doroghazi and Buckley 2010).

The low cost of new genome sequencing technologies has enabled scientists to search microbial genomes for biosynthetic clusters encoding secondary metabolites. This has led to the rapid accumulation of Streptomyces genome sequences in public databases. At the time of this writing, the genomic sequences of over 600 Streptomyces isolates are publically available, and thousands more are expected in the next few years. The size of Streptomyces genomes is typically within the range of 7 Mb to 10 Mb (Studholme 2016). The Streptomyes scabiei 87-22 genome was the first sequenced genome of pathogenic Streptomyces species. Representatives of three other pathogenic Streptomyces species were also sequenced recently by our group, namely

Streptomyces acidsicabies 84-104 (Huguet-Tapia and Loria 2012), Streptomyces turgidiscabies

Car8 (Huguet-Tapia et al. 2011) and Streptomyces ipomoea 91-03 (Huguet-Tapia et al. 2016).

Plant Pathogenic Streptomyces Species

Although Streptomyces are viewed predominantly as free-living terrestrial soil-inhabiting bacteria, some species are symbionts with fungi, insects, plants and animals (Seipke et al. 2012).

Streptomycetes have also been isolated from marine soils to pursue novel secondary metabolites

(Fiedler et al. 2005). To date, at least 11 Streptomyces species are plant pathogens and cause diseases of underground crop structures such as roots and tubers (Dees and Wanner 2012). The most economically important disease caused by Streptomyces species is known as potato scab, characterized by raised, pitted, or superficial lesions formed on the tuber surface of Solanum tuberosum (Loria et al. 2006). Scab-causing pathogenic Streptomyces species do not display

tissue or host specificity; they can cause scab symptoms on underground structures of diverse crops, such as potato, carrot, radish, beet and other tap root crops (Loria et al. 2006). The best characterized scab-causing species are S. scabiei, S. acidiscabies, and S. turgidiscabies. Unlike scab-producing streptomycete pathogens, another pathogenic species Streptomyces ipomoeae tends to naturally cause disease to only family Convolvulaceae plant spp., including sweet potato

(Ipomoea batatas [L.] Lam.), but does not infect potato (Clark and Watson 1983). This disease is called soil rot, differing from potato scab by its ability to cause highly necrotic destruction of plant roots (Loria et al. 2003). A brief description of known plant pathogenic Streptomyces species follows.

Streptomyces scabiei (syn. S. scabies)

This species is the best studied and most widely distributed plant pathogenic

Streptomyces species. It can cause scab disease on beets, carrot, radish, and other tuber and tap root crops in agricultural systems. S. scabiei can also attack the pericarp of the peanut fruit and cause the pod wart disease of peanuts in Israel (Kritzman et al. 1996) and South Africa (De Klerk et al. 1997). S. scabiei produces thaxtomins, and thaxtomin A is the predominant form (King et al. 2002; King and Calhoun 2009).

Streptomyces acidiscabies

S. scabiei is generally controlled by maintenance of a soil pH below 5.2; however, S. acidiscabies can be found in soil between 4.5 and 5.2 (Lambert and Loria 1989) and cause potato scab in low pH soils. S. acidiscabies was first described in the northeastern United States

(Lambert 1991; Lambert and Loria 1989). Later this species was also reported from other regions, such as Korea (Song et al. 2004), Japan (Tóth et al. 2005), and China (Zhao et al. 2010).

This species produces thaxtomin A (Bukhalid et al. 1998).

Streptomyces turgidiscabies

This species was first reported in Japan as the cause of potato scab with a special erumpent lesion (Miyajima et al. 1998; Takeuchi et al. 1996), and later was reported in other region of the world, such as Europe (Kreuze et al. 1999; Thwaites et al. 2010), Korea (Kim et al.

1999), China (Zhao et al. 2008), and North America (Wanner 2009). It has been confirmed that this species produces thaxtomin A (Bukhalid et al. 1998).

Streptomyces ipomoeae

This species appears to cause disease only on plants in the family Convolvulaceae, most importantly on sweet potato (Ipomoea batatas [L.] Lam.) in agricultural systems. It is the causal agent of sweet potato soil rot disease (King et al. 1994), which can be distinguishable from the scab disease by its highly necrotic destruction of plant roots (Loria et al. 2003). Instead of thaxtomin A, S. ipomoea produces a less-modified, nonhydroxylated form of thaxtomin known as thaxtomin C (King et al. 1994).

Streptomyces europaeiscabiei

This species is closely related to S. scabiei and is the predominant scab-causing species in

Europe. S. europaeiscabiei has also been described in other regions including Korea (Song et al.

2004) and North America (Wanner 2009). This species also produce thaxtomins (Loria et al.

2006).

Streptomyces stelliscabiei

This species was first isolated from star-like lesions on potato tubers in France (Bouchek-

Mechiche et al. 2000). The16S ribosomal RNA (16S rRNA) sequences of S. stelliscabies indicate that it is closely related to the non-pathogenic species S. bottropensis (Bouchek-

Mechiche et al. 2000). This species has not been well studied, but it is able to produce thaxtomins (Loria et al. 2006).

Streptomyces reticuliscabiei

This is the main species that cause netted scab in France. As thaxtomin A does not cause netted scab lesions (Loria et al. 1997), S. reticuliscabiei might possess other virulence factors

(Bouchek-Mechiche et al. 2006). S. reticuliscabiei does not contain the thaxtomin A biosynthetic cluster (Aittamaa et al. 2010). DNA–DNA hybridization (DDH) and 16S rRNA gene sequences indicated S. turgidiscabies and S. reticuliscabiei are closely related (Bouchek-Mechiche et al.

2006).

Streptomyces caviscabies

This species was first isolated in Quebec, Canada. It can cause the deep-pitted lesions on potato tubers (Goyer et al. 1996a). This species was reported to produce thaxtomins in media made out of potato or sweet potato peels, but not in oatmeal media (Goyer et al. 1996b).

Streptomyces luridiscabiei, Streptomyces puniciscabiei, and Streptomyces niveiscabiei

These three species were first identified as causal agents of potato scab in Korea (Park et al. 2003a). Among these three species, thaxtomin A production was only confirmed in S. niveiscabiei (Park et al. 2003b).

Pathogenicity and Virulence Factors

Thaxtomins

The primary pathogenicity factor of plant pathogenic Streptomyces is the cyclic nitrated dipeptide phytotoxin thaxtomins (King and Calhoun 2009). Thaxtomins are cyclic dipeptides

(2,5-diketopiperazines) derived from L-phenylalanine and L-tryptophan and contain a 4- nitroindole moiety that is essential for the phytotoxicity of these compounds (King and Calhoun

2009). In S. acidiscabies, S. scabiei, and S. turgidiscabies, thaxtomin A (ThxA) is the most predominant form (Bukhalid et al. 1998). ThxA inhibits cellulose biosynthesis in expanding plant tissues, stimulates Ca2+ spiking, and causes cell death (Tegg et al. 2005). Disruption of

ThxA biosynthetic genes abolished ThxA production, and the mutants became avirulent on potato tubers (Healy et al. 2000). ThxA is synthesized via two nonribosomal peptide synthetases

(NRPS) encoding by txtA and txtB genes, a P450 monooxygenase (TxtC), a nitric oxide synthase

(TxtD), and a novel cytochrome P450 (TxtE) that site specifically nitrates tryptophan prior to cyclization (Bignell et al. 2014b). A 65 amino acid protein encoded by txtH is a predicted MbtH- like family protein and may be important in regulating NRPS activity (Herbst et al. 2013;

Stegmann et al. 2006). The expression of txtH was shown to be increased under thaxtomin- inducing conditions (Bignell et al. 2010a). However, its role in thaxtomin A production is the subject of ongoing research. S. ipomoeae produces a less-modified, nonhydroxylated member of this chemical family known as thaxtomin C (ThxC) (King et al. 1994), but this is also a required pathogenicity factor (Guan et al. 2012).

Previous studies have identified inducers that can trigger the ThxA production, including cereal xylans (Wach et al. 2007), potato suberins (Lauzier et al. 2008), and, more importantly, cellobiose and cellotriose (Johnson et al. 2007). Recent studies also demonstrated that the biosynthesis of ThxA is under strict transcriptional control by global and pathway-specific transcriptional regulators (Bignell et al. 2014a; Joshi et al. 2007a; Francis et al. 2015). The

AraC/XylS regulator TxtR is imbedded in the ThxA cluster, and the TxtR protein can bind cellobiose and activate expression of ThxA biosynthetic genes (Joshi et al. 2007a). The bld

(bald) gene family contains a group of pleiotropic regulators that are required for proper morphological development of Streptomyces (McCormick and Flardh 2012). TxtA production was reduced or abolished upon the deletion of bldA, bldC, bldD, bldG and bldH/adpA, indicating these five bld genes may act as global regulators of TxtA production (Bignell et al. 2014a).

Recently, the cellulose utilization repressor CebR was demonstrated to be the master regulator of

TxtA biosynthesis and virulence in pathogenic Streptomyces species (Francis et al. 2015). CebR binds to DNA motifs within txtB and upstream txtR and represses the expression of ThxA biosynthetic genes; binding of cellobiose to CebR can release CebR from the binding sites, leading to the increased expression of TxtA biosynthetic genes (Francis et al. 2015). ThxA inhibits plant cellulose synthesis in the ppm range (Fry and Loria 2002). Its superior potency has attracted industrial interests in developing ThxA as a natural and biodegradable herbicide to control weed growth (King et al. 2001). Several companies tried to market thaxtomin as a novel herbicide but the cost of producing it is too high

(http://reflexions.ulg.ac.be/cms/c_386474/en/thaxtomin-a-next-generation-weed- killer?portal=j_55&printView=true). Since deletion of cebR leads to constitutive thaxtomin A production, this will facilitate overproduction of this potential herbicide for commercial purposes

(Francis et al. 2015).

Nec1

Unlike many gram-negative pathogens, pathogenic actinobacteria lack the type III secretion system, which is used to inject virulence proteins directly into the cytoplasm of a eukaryotic host cell (Loria et al. 2006). A second virulence determinant identified in

Streptomyces is encoded by nec1, which lies on the pathogenicity island in S. turgidiscabies

(Kers et al. 2005). The nec1 gene encodes a necrogenic protein. Western blot analysis showed that Nec1 protein was detected in the supernatant of S. turgidiscabies culture, suggesting Nec1 is a secreted virulence protein (Joshi et al. 2007b). The nec1 gene confers a pathogenic phenotype on non-pathogenic Streptomyces lividans TK24 and Escherichia coli in an excised potato tuber tissue assay. Its low G+C content (54%) indicates it has moved into Streptomyces from another genus. However, no nec1 homologs are found in GenBank nuclecic acid or protein databases, and Nec1 protein also lacks characterized motifs. ThxA inhibits cellulose biosynthesis and can

help Streptomyces pathogens penetrate plant cell walls directly, which may allow proteins secreted by pathogens to move into host cells (Loria et al. 2006). SignalP 3.0 and TatP 1.0 programs indicate Nec1 can be secreted via a N-terminal signal sequence through the Sec or Tat secretion pathways (Joshi et al. 2007b). However, the agarase assay that tests the functionality of

Tat signal peptides, indicated that the Nec1 signal peptide was not targeted to the Tat pathway. It is therefore likely that Nec1 is secreted via Sec pathway (Mann 2009). Nec1 is expected to interact with plant proteins in the plant cell (Joshi et al. 2007b), but this has not been demonstrated and putative host targets have not been suggested. The nec1 gene is missing in some pathogenic Streptomyces isolates obtained from infected potatoes (Kreuze et al. 2007;

Flores González et al. 2008), indicating unlike ThxA, Nec1 is not essential for pathogenicity

(Lerat et al. 2009).

TomA

A putative tomatinase gene (tomA) is conserved on the pathogenicity islands of S. scabiei, S. acidiscabies, and S. turgidiscabies (Seipke and Loria 2008). Tomatinase a group of enzymes called saponinases; saponinases are able to hydrolyze and detoxify phytoanticipins, which are preformed antimicrobial compounds from plants (VanEtten et al. 1994). Tomatinase genes are typically found in fungal pathogens infecting tomato plants. However, a tomA gene was also found in Clavibacter michiganensis subspecies michiganensis, which is a plant pathogenic actinobacterium causing bacterial wilt and canker of tomato (Kaup et al. 2007). The tomA mutant of S. scabiei was not demonstrated to be compromised in virulence (Seipke and

Loria 2008), but the conservation of tomA in plant pathogenic Streptomyces species suggests this gene may play a role in host-pathogen interactions (Kers et al. 2005).

Fas Operon

In S. turgidiscabies, a six-gene plant fasciation (fas) operon lies adjacent to the thaxtomin biosynthetic cluster (Kers et al. 2005). This fas operon is absent from S. acidsicabies and S. scabiei, but is homologous to the fas operon in Rhodococcus fascians (Joshi and Loria 2007). R. fascians is an important gram-positive pathogen on ornamental plants, causing the leafy gall syndrome (Stes et al. 2013). The fas operon encodes a cytokinin biosynthesis pathway, which is responsible for a mixture of five cytokinins essential for virulence (Stes et al. 2013). Five of the six fas genes are collinear in S. turgidiscabies and R. fascians, indicating the fas genes in these two pathogens have a common ancestor (Joshi and Loria 2007). In an indistinguishable manner from R. fascians, S. turgidiscabies produces leafy galls on Arabidopsis and tobacco, as well as de novo meristem production on tobacco (Joshi and Loria 2007). Although the roles of cytokinins in the S. turgidiscabies-potato interaction are unknown, the large and erumpent lesions on potato produced by S. turgidiscabies may be due to cytokinins (Joshi and Loria 2007).

Coronafacic Acid

With the completion of genome sequencing of S. scabiei 87-22, the coronafacic acid

(CFA)-like gene cluster was identified (Bignell et al. 2010b). Unlike the thaxtomin A biosynthetic cluster and other virulence genes, this CFA-like gene cluster is not located on pathogenicity islands (Bignell et al. 2010b). This cluster was predicted to produce a molecule similar to the CFA component of the virulence-associated coronatine (COR) phytotoxin produced by the plant pathogen Pseudomonas syringae (Bender et al. 1999b; 1999a). As a key virulence factor of P. syringae, COR has several important roles in host infection, including facilitation of invasion of P. syringae by interfering with stomatal defenses, promotion of bacterial growth and persistence within the plant apoplast, and induction of disease susceptibility in uninfected plant regions (Xin and He 2013). A recent study demonstrated that the major

product of the S. scabiei CFA-like gene cluster was CFA-L-isoleucine (Fyans et al. 2015).

Similar to COR, CFA-L-isoleucine can elicit potato tissue hypertrophy and radish seeding stunting, though its bioactivity is not as toxic as that of COR (Fyans et al. 2015). In addition, knock-out of the cfa-like cluster reduced the virulence of S. scabiei on tobacco seedlings, indicating that the cfa-like cluster contributes to the full virulence of S. scabiei (Bignell et al.

2010b).

Concanamycins

Concanamycins are polyketide macrolides that are characterized by an 18-membered tetraenic macrolide ring with a ß-hydroxyhemiacetyl side chain and a methyl enol either (Bignell et al. 2014b). Concanamycins are inhibitors of vacuolar ATPase and are active against fungi but not against bacteria (Kinashi et al. 1984). Concanamycins A, B, and C are first isolated from the culture medium of S. diastatochromogenes (Kinashi et al. 1984), and later S. scabiei isolates from Japan were also found to produce concanamycin A and B that inhibited root growth in rice seeding assays (Natsume et al. 1996; 1998). Since other scab-causing Streptomyces species do not produce concanamycins, they may not be essential to the the scab-disease and their roles in disease development require further investigation (Natsume et al. 1998; 2001; 2005).

FD-891

FD-891 is a 16-membered polyketide macrolide, and has similar structure to concanamycins (Natsume et al. 2005). However, the action modes of the two types of phytotoxins seem to be different (Kataoka et al. 2000). It was isolated from Streptomyces cheloniumii (this name has not been officially accepted), which was reported to be responsible for the potato russet scab disease in Japan (Suzui et al. 1988). Similar to ThxA, FD-891 is a nonspecific phytotoxin, and can not only induce necrosis of potato tuber, but also cause stunting of rice and alfalfa seedlings (Natsume et al. 2005).

Borrelidin

Borrelidin is an 18-membered polyketide macrolide exhibiting anti-bacterial, anti-viral, anti-malarial and anti-angiogenic activity (Dickinson et al. 1965; Wakabayashi et al. 1997;

Otoguro et al. 2003). Furthermore, borrelidin purified from the culture of Streptomyces sp. GK18 was shown to induce deep pitted lesions on potato tubers, and cause severe stunting of potato plants grown in pots, indicating borrelidin also exhibits phytotoxic activity (Cao et al. 2012).

Other Unknown Pathogenicity/virulence Factors

Recently, Fyans et al. isolated a pathogenic Streptomyces strain 11-1-2 associated with common scab-infected potato in Newfoundland, Canada (Fyans et al. 2016). This strain did not produce thaxtomins, borrelidin or concanamycins, but it was shown to cause significant pitting of potato tuber tissue, and induce severe necrosis in Nicotiana benthamiana leaf infiltration assay

(Fyans et al. 2016). To better understand how S. sp. 11-1-2 is able to cause disease, further characterization of the phytotoxic substance it produces is needed.

Integrative and Conjugative Elements

Plant bacterial pathogens have evolved to possess diverse pathogenicity/virulence factors, such as specialized secretion systems, effectors, lipopolysaccharides (LPS), exopolysaccharides

(EPS), phytotoxins, and phytohormones, all of which may play important roles in preventing pathogen recognition, suppressing the host defense, promoting nutrient extraction, and facilitating pathogen reproduction and spread (Abramovitch and Martin 2004; Alfano and

Collmer 1996; Toth and Birch 2005). Many such pathogenicity/virulence determinants are associated with mobile genetic elements, including transposons, bacteriophage, and plasmids

(Canchaya et al. 2003; Dobrindt et al. 2004; Hacker et al. 2004). The horizontal gene transfer

(HGT) of these genetic elements allows the bacteria to acquire new traits and adapt to new and fluctuating environments rapidly (Burrus and Waldor 2004).

HGT of genetic materials between bacteria is mainly via three mechanisms: natural transformation, transduction and conjugation (Burrus and Waldor 2004). Some bacteria are able to take up exogenous DNA molecules released by other organisms into the cells via transformation and incorporate into their genomes (Dubnau 1999; Lorenz and Wackernagel

1994). In transduction, a bacteriophage packages a portion of the genome of an infected host into the phage head and transfers these genes to another bacterial host (Huddleston 2014).

Conjugation requires cell-to-cell contact, and DNA is often transferred via a mating pores that allows for the passage of DNA (Huddleston 2014). The transferred DNA is usually processed first in the donor cell to become single stranded DNA, and then converted back to double- stranded DNA in the recipient cell (Johnson and Grossman 2015). Often the identified conjugative systems located on plasmids. However, conjugation systems are frequently discovered in chromosome-borne mobile genetic elements, which are often referred to as integrative and conjugative elements (ICEs) (Wozniak and Waldor 2010).

An ICE encodes the machinery for recombination and conjugation as well as regulatory systems to mediate its excision from the chromosome and its conjugative mobilization (Wozniak and Waldor 2010). ICEs have two defining features: (1) they integrate in the host chromosome;

(2) they possess a conjugation machinery, a type IV secretion system, to mediate the transfer of

ICEs to other host cells (Johnson and Grossman 2015). Some prophages and integrative elements carry serine recombinases (integrases) (Smith and Thorpe 2002), however most ICEs possess tyrosine recombinases (integrases) (Johnson and Grossman 2015). Two recombinase families catalyze slightly different biochemical reactions, but both have the same functional consequence for the ICE. They can mediate the site-specific recombination between dsDNA molecules at the attachment (att) sites, which are short stretches of similar sequence (Grindley et al. 2006; Rajeev

et al. 2009). In addition to the recombinase, a recombination directionality factor (RDF), sometimes called Xis, is frequently discovered in ICEs and is also required for excision (Johnson and Grossman 2015). RDF influences the direction of recombinase-mediated recombination to favor excision (Hirano et al. 2011).

The ICE life cycle was summarized in a previous paper (Johnson and Grossman 2015): under normal conditions, ICEs are integrated into the att site of host chromosome and conjugation genes are not expressed; ICE gene expression can be induced stochastically or by specific cellular signal, leading to the ICE excision from the chromosome and formation of a circular double strand (dsDNA) structure, and leading several ICE genes to encoding proteins that assemble into a mating pore that allows for the passage of ICE DNA; the origin of transfer

(oriT) will be recognized, and one DNA strand of ICE will be nicked within the oriT to generate a linear single-stranded DNA (ssDNA)-protein complex; the linear ssDNA-protein complex will be pumped into the recipient cell via the mating pore; the ICE likely re-circularize to the circular dsDNA, and then integrate into the att site of the chromosome of new host mediated by an ICE- encoded integrase.

Many ICEs carry cargo genes that are not associated with the ICE life cycle and these genes are typically thought to provide host cells with certain benefits, including antibiotic resistance, heavy metal resistance, and the ability to metabolize a new carbon source (Johnson and Grossman 2015). In addition, pathogenicity/virulence genes are also frequently found to be located on ICEs, and these ICEs are called pathogenicity islands (PAIs) (Dobrindt et al. 2004).

Acquisition of a PAI may be sufficient to enable a non-pathogenic bacterial strain to become pathogenic (Burrus and Waldor 2004).

Pathogenicity Islands of Streptomyces

The PAI in S. turgidiscabies Car8

It has been thought that HGT contributes to the evolution of pathogenic streptomycetes

(Loria et al. 2006). S. turgidiscabies Car8, isolated from Japan, was demonstrated to contain a mobile 674 Kb PAI (Kers et al. 2005; Huguet-Tapia et al. 2011). This mobile large PAI carries the ThxA biosynthetic cluster as well as other virulence genes, including nec1, tomA, and the fas operon (Huguet-Tapia et al. 2011). During conjugation, the Car8 PAI can be transferred to non- pathogenic Streptomyces species as ~100 Kb or ~ 640 Kb modules, both of which can be site- specifically integrated into the att site (TTCATGAA) at the 3’ end of a bacitracin resistance gene

(bacA) (Kers et al. 2005). Acquisition of the 674 Kb PAI conferred a pathogenic phenotype on non-pathogenic species S. diastatochromogenes but not on S. coelicolor, indicating the genetic background of the recipient cell is important to the functionality of PAI (Kers et al. 2005).

The Car8 PAI is revealed to able to excise by recombination of two flanking and one internal att sites in three modules: (1) a 105 Kb circular module containing nec1 and tomA ; (2) a

569 Kb circular module carrying the ThxA biosynthetic cluster and the fas operon; and (3) a 674

Kb circular module of the whole Car8 PAI (Huguet-Tapia et al. 2014).

The Toxicogenic Region and Colonization Region in Other Pathogenic Species

Comparison of the genomes of pathogenic Streptomyces species revealed that the 105 Kb

PAI module of Car8 also existed in other pathogenic isolates, such as S. scabiei 87–22 and S. acidiscabies 84–104, but the 569 Kb PAI module was unique to S. turgidiscabies Car8 (Huguet-

Tapia et al. 2014). Interestingly, although other pathogenic Streptomyces species, namely S. scabiei and S. acidiscabies, lack the 569 Kb PAI, they all contain the ThxA cluster, which is located 4.9 Mb away from the virulence genes nec1 and tomA. These results suggest that, unlike

S. turgidiscabies, in S. scabiei and S. acidiscabies, the ThxA cluster, nec1 and tomA do not reside

on a single PAI. The region containing the ThxA cluster is called the toxicogenic region (TR), while the 105 Kb region carrying nec1 and tomA is called the colonization region (CR) (Lerat et al. 2009).

Recently, Chapleau et al. demonstrated experimentally the excision of the ThxA biosynthetic cluster from the chromosome of S. scabiei 87-22 (Chapleau et al. 2016). Two flanking and one att sites separate the S. scabiei TR region into two sub-regions, designated TR1 and TR2; the 20 Kb TR1 region possesses the ThxA biosynthetic cluster, and the 157 Kb TR2 region carries putative ICEs (Chapleau et al. 2016). The composite TR region was demonstrated to excise from the chromosome in three distinct circular modules: TR1 alone, TR2 alone and the whole composite TR region (Chapleau et al. 2016). Chapleau et al. also showed that most pathogenic Streptomyces strains lacked TR2. TR excision could not be detected in such strains, indicating that TR2 may be important for the excision of TR (Chapleau et al. 2016). However, excision alone does not mean the S. scabiei TR region has the ability to be transferred into other hosts. Further investigation into the HGT of S. scabiei TR will provide novel insights into the origin and evolution of pathogenicity of the genus Streptomyces.

Goals and Objectives of This Study

Emergence of new pathogens in agricultural systems is biologically interesting and economically important. Many reports of emergent plant pathogens exist in the largely saprophytic genus Streptomyces, making it an appropriate system for genomic and functional analysis of this phenomenon. The overall goal of this study is to investigate the evolution of plant pathogenicity in the genus Streptomyces, by 1) understanding the phylogenetic relationships of pathogenic species and their related non-pathogenic species in the genus

Streptomyces, and 2) characterizing the diversity and role of pathogenicity islands (PAIs) in the emergence of new plant pathogenic species in the Streptomyces genus.

Dissecting the origin and development of pathogenicity in Streptomyces requires an understanding of the phylogenic relationships among pathogenic species and related nonpathogenic species. In other bacterial genera, analysis of hundreds of single-copy genes as a single data set has provided a superior characterization of evolutionary relationships. This study aims to sequence the genomes of more than ten pathogenic Streptomyces isolates from different regions of the world, and use whole-genome sequences to infer the evolutionary relationships of pathogenic species and nonpathogenic species.

The HGT of PAIs is thought to be a driving force in the evolution of plant pathogenic streptomycetes. Whole genome sequences from different pathogenic species will enable us to characterize the genetic and structural diversity of PAIs and infer their evolution history. The mobilization of S. turgidiscabies Car8 has been confirmed. However, S. scabiei is the most widely distributed and best studied pathogenic Streptomyces species. The capacity for mobilization of the S. scabiei TR to saprophytic Streptomyces isolates, similar to what has been observed for the S. turgidiscabies PAI, has been suggested but not confirmed. This study aims to investigate the mobilization and conjugative transfer of the S. scabiei TR, which would have important implications for the evolution of pathogenicity.

Another objective of the research described here is to demonstrate the power of whole- genome sequencing in addressing complex taxonomic issues and accurately identifying novel species. We will develop a phylogenomic and systematic pipeline and apply it to the

Streptomyces genus as well as other important plant pathogenic genera, such as Dickeya and

Pectobacterium. We aim to show other researchers the potential of using whole genome sequences to re-evaluate current prokaryotic species definition and establish a unified prokaryotic species approach for taxonomically challenging genera.

CHAPTER 2 PHYLOGENY OF THE STREPTOMYCES GENUS

Introduction

Streptomyces spp. are Gram-positive, filamentous actinobacteria with high G+C content genomes. They are prolific producers of biologically-active secondary metabolites, including two-thirds of currently-used antibiotics (Bentley et al. 2002). Streptomyces are viewed predominantly as free-living, terrestrial, soil-inhabiting, and saprophytic bacteria, but a few species are plant pathogens (Loria et al. 2006). These pathogens do not display tissue or host specificity; they cause scab symptoms on underground structures of diverse crops such as potato, carrot, radish, beet and other tap root crops (Loria et al. 2006). Among these diseases, potato scab characteristically produces raised, pitted, or superficial lesions on the tuber surface and is the most economically-important of the Streptomyces diseases (Loria et al. 2006). The best characterized potato scab-causing species are Streptomyces scabiei, Streptomyces acidiscabies, and Streptomyces turgidiscabies (Loria et al. 2008). Unlike scab-causing streptomycetes,

Streptomyces ipomoeae has a limited host range and only causes the disease in members of the

Convolvulaceae family, including sweet potato (Ipomoea batatas [L.] Lam.) (Clark and Watson

1983).

Understanding the phylogenetic relationships of phytopathogenic Streptomyces species can provide insights into the evolution of virulence factors and the emergence of new pathogenic species. The analysis of 16S ribosomal RNA (rRNA) of pathogens has become the standard approach for determining evolutionary relationships among species (Kreuze et al. 1999), but highly similar 16S rRNA sequences among Streptomyces species limits the usefulness of this approach (Labeda et al. 2012). Genes encoding DNA topoisomerase, ribosomal protein S1, SecY and GTPase, among others were used for phylogenetic analysis of Streptomyces, but no single

gene yielded a resolved phylogeny (Alam et al. 2010). To address this issue, multilocus sequence analysis (MLSA) was developed (Guo et al. 2008). This approach compares and analyzes multiple internal fragments of housekeeping genes, namely ATP synthase F1 (ß-subunit; atpD),

DNA gyrase (ß-subunit; gyrB), recombinase A (recA), RNA polymerase (ß-subunit; rpoB), and tryptophan synthase (ß-subunit; trpB), to establish phylogenetic relationships (Guo et al. 2008).

Indeed, a higher resolution of closely related phytopathogenic Streptomyces species has been achieved using this approach (Labeda 2011). However, the MLSA approach assumes that the housekeeping genes chosen for analysis are present in all genomes of the taxonomic unit being studied, and not duplicated in any of those genomes. Previous MLSA studies of Streptomyces have not evaluated these assumptions, because large numbers of Streptomyces genomes have only recently become available.

However, concatenation of hundreds of selected single-copy genes from microbial genomes as a single data set is now possible, and comparative analyses of these data sets have been used to reconstruct highly resolved evolutionary relationships (Salichos and Rokas 2013).

The exponentially growing microbial genomic data further augments the application of this approach in precisely defining the taxa of new isolates. At the time writing, the genomic sequences of over 600 Streptomyces isolates are publically available, and thousands more are expected in the next few years. Although Streptomyces is now one of the most highly sequenced genera, many genomes are sequenced with the intent of mining secondary metabolism gene clusters, while leaving the genomes of phytopathogenic species unsequenced. Until recently, only four genomes of pathogenic Streptomyces isolates were available, namely S. scabiei 87-22

(Bignell et al. 2010b), S. acidsicabies 84-104 (Huguet-Tapia and Loria 2012), S. turgidiscabies

Car8 (Huguet-Tapia et al. 2011) and S. ipomoea 91-03 (Huguet-Tapia et al. 2016). Four genomes

provide valuable information for studying the pathogenicity and virulence factors of pathogenic

Streptomyces species, but have been insufficient in determining their evolutionary relationships.

Herein, we investigated the evolutionary history of scab-causing pathogenic streptomycetes. Specifically, we sequenced the genomes of ten phytopathogenic Streptomyces isolates of various geographic origins and one non-pathogenic species, Streptomyces diastatochromogenes. We performed a phylogenomic and systematic analysis of their genomes along with 100 Streptomyces genomes deposited in the GenBank database; fortunately among the 100 genomes nine were recently sequenced Streptomyces phytopathogens. Initially, we employed a MLSA approach to these genomes, based on five housekeeping genes (atpD, gyrB, recA, rpoB and trpB) typically used for phylogenetic analysis of Streptomyces strains. We found that some genomes contain multiple copies of these five genes, indicating the currently used

MLSA scheme is invalid for such an analysis. In contrast, whole-genome sequences, filtered to provide a data set consisting of conserved, single-copy genes, could be used to reconstruct evolutionary relationships of pathogenic species and related non-pathogenic species.

Additionally, we found that some pathogenic Streptomyces isolates were misidentified, and that

Streptomyces sp. 96-12 belongs to a novel pathogenic Streptomyces species.

Materials and methods

Bacterial Strains, Culture Conditions and DNA Extraction

Streptomyces strains were cultured at 28 °C on International Streptomyces Project medium 4 (ISP4) agar medium or in tryptic soy broth (TSB; BD Biosciences). All liquid cultures were shaken at 250 rpm. Genomic DNA was extracted from TSB-grown cultures using the

MasterPure Gram-positive DNA purification kit (Epicentre) based on the manufacturer’s instructions.

Virulence Bioassays

To assess the virulence phenotype of different strains, an in vitro radish seedling bioassay was performed as described previously (Bignell et al. 2010b) except that “German Giant” seeds

(Burpee) were used in place of the “Burpee White” variety. For each strain tested, mycelial suspensions were used to inoculate six newly germinated radish seedlings of similar size. The assay was performed in triplicate with each three different biological replicates.

Analysis of Thaxtomin A Production

Plant pathogenic Streptomyces strains were culture in TSB for 48 to 72 h. Mycelia were then pelleted by centrifugation, washed twice with sterile water, and re-suspended in sterile water to obtain an OD600 of 1.0. The mycelial suspension solution (400 µl) was then inoculated into 3 × 50 ml of OBB medium or OBB medium supplemented with 1% cellobiose (OBBc) in

250 ml flasks. After incubation for 6 days at 28 °C with shaking (250 rpm), thaxtomin A was purified from the culture supernatants and analyzed by HPLC as described previously (Johnson et al. 2007). All experiments were repeated at least three times using different biological replicates of the Streptomyces strains, with three technical repeats per strain. The pellets of the culture samples were dried and weighted as a measure of bacterial growth. The data for S. scabiei 87-22 was normalized to 100 %. The data for all pathogenic isolates were square-root transformed to improve homogeneity of variance. Transformed data were subjected to one-way

ANOVA using SAS 9.4 (SAS Institute, Cary, NC), and means were compared based on Tukey's

HSD test at P < 0.05.

Genome Sequencing and Assembly

We sequenced the genomes of Streptomyces isolates (Table 2-1) using the Illumina-

HiSeq 2500 paired-end technology. For each genome, the SPAdes version 3.5.0 assembly algorithm (Bankevich et al. 2012) was used to perform genome assemblies. Gene-finding and

annotation were achieved using the Rapid Annotation using Subsystems Technology (RAST) server (Aziz et al. 2008).

Ortholog Retrieval and Single-copy Cluster Search

To build the rooted phylogenetic tree, Streptacidiphilus albus JL83 and Kitasatospora setae NBRC 14216 were selected as outgroups. Genes from 111 Streptomyces genomes and the two outgroups were clustered into ortholog groups using OrthoMCL v1.4 (Li et al. 2003). We first conducted an all-against-all protein sequence comparison of 113 strains using BLASTP, and defined putative pairs of orthologs based on reciprocal BLASTP. Subsequently, the reciprocal

BLASTP values were converted to a normalized similarity matrix that was analyzed by a

Markov Cluster algorithm (MCL) (Enright et al. 2002), which yielded a set of clusters of homologs. OrthoMCL was run with a BLAST E-value cut-off of 1e-5, a minimum aligned sequence length coverage of 50% of the query sequence, and an inflation index of 1.5. Ortholog groups with a single copy per genome were selected as potential phylogenetic markers.

Gene Alignment and Filtering Criteria

To minimize inclusion of alignments of low quality in the analysis, we used various filtering criteria. Alignments for each of the single-copy orthologs were generated at the amino acid level using Probalign v1.1 (Roshan and Livesay 2006), then back-translated to DNA sequences, based on the original genomic sequence. Subsequently, alignment columns with a posterior probability < 0.6 were removed (Lefébure and Stanhope 2009). Alignments made after removing all ambiguous positions that were less than 50% of the original alignment length were excluded (Lefébure and Stanhope 2009).

Maximum Likelihood (ML) Phylogeny Estimation

The gene alignment was used for Maximum Likelihood (ML) inference on the RAxML webserver (https://www.phylo.org/portal2/) (Stamatakis et al. 2008) with the GTR+G substation

model, including an estimation of invariable sites. For inferring tree robustness, 1000 bootstrap replicates were computed.

Average Nucleotide Identity and in silico DNA-DNA Hybridization Estimation

The average nucleotide identity (ANI) analysis of whole-genome data was performed using the method proposed by Richter and Rosselló-Móra (Richter and Rosselló-Mora 2009).

Pair-wise ANI values were obtained from nucmer (Delcher et al. 2002) with the script calculate_ani.py

(https://github.com/widdowquinn/scripts/blob/master/bioinformatics/calculate_ani.py). The in silico DNA-DNA hybridization (isDDH) was estimated using the genome-to-genome distance calculator (GGDC) (Auch et al. 2010a; Meier-Kolthoff et al. 2013). The contig files were uploaded to the GGDC 2.0 Web server (http://ggdc.dsmz.de/distcalc2.php), where isDDH calculations were performed using Formula 2 alone. We used the point estimates plus the 95% model-based confidence intervals as the isDDH estimates.

Results

Characterization of Pathogenic Streptomyces Isolates

We sequenced the genomes of 10 isolates of plant pathogenic streptomycetes from different locations, and the genome of S. diastatochromogenes ATCC 12309, a nonpathogenic isolate (Table 2-1). To validate the pathogenicity of selected plant pathogens, we tested their ability to cause disease symptoms on radish seedlings. Scab-causing Streptomyces species cause tissue swelling and necrosis, and severe root and shoot stunting, most likely due to the production of thaxtomin A (ThxA) (Loria et al. 1997). All tested plant pathogenic isolates induced similar disease symptoms on radish seedlings following infection (Figure 2-1A).

ThxA inhibits plant cellulose synthesis in the ppm range (Fry and Loria 2002). Its superior potency has attracted industrial interests in developing ThxA as a natural and

biodegradable herbicide to control weed growth (King et al. 2001). Cellobiose, a strong elicitor of ThxA production (Johnson et al. 2007), can significantly induce ThxA production. However, marketing ThxA as a herbicide is not profitable, due to the high cost of cellobiose. Therefore, pathogenic isolates with the ability to produce a large amount of ThxA without cellobiose can be used as the industrial strain. We used HPLC analysis to quantify ThxA production levels of different pathogenic Streptomyces isolates cultured in oat bran broth (OBB), an undefined complex medium known to induce ThxA biosynthesis (Wach et al. 2007). Our assay also included three well-studied pathogenic isolates, namely S. scabiei 87-22, S. acidiscabies 84-104, and S. turgidiscabies Car8. The pathogenic Streptomyces isolates produced ThxA at different levels (Figure 2-1B). S. acidiscabies 84-104 produced the highest level of ThxA, followed by S. turgidiscabies Car8 and S. scabiei 96-08. All other tested isolates produced a significantly lower level of ThxA. ThxA was undetectable in the extracts of S. europaeiscabiei 96-14 and S. sp. 96-

12 cultured in OBB. This result was unexpected, but when OBB was supplemented with 1% cellobiose (OBBc), ThxA production levels of all strains were increased and ThxA production by

S. europaeiscabiei 96-14 and S. sp. 96-12 was also detected (data not shown), confirming their ability to produce ThxA.

Two Copies of Selected Housekeeping Genes Were Detected in Some Streptomyces Genomes

MLSA has been widely applied to unravel bacterial taxonomic relationships. The commonly used MLSA for Streptomyces utilizes partial sequences of housekeeping genes atpD, recA, gyrB, rpoB and trpB (Guo et al. 2008). The homologs of five housekeeping genes in 111

Streptomyces genomes (Tables 2-1, 2-2) were determined by OrthoMCL v1.4, which provides a scalable method for constructing homologous groups across multiple taxa (Li et al. 2003).

Interestingly, none of the five genes were single-copy in all 111 genomes; some genomes

contained two copies of these genes (Table 2-3). The genomes with two copies of these housekeeping genes were also confirmed manually via a BLASTN search. The atpD and rpoB were not present as single-copy only in one strain, namely Streptomyces regensis NRRL B-

11479 (Genbank accession # LFVR00000000.1). Indeed, none of the five genes were present as single-copy in this assembly. It may be that the sequenced DNA of S. regensis NRRL B-11479 was contaminated by DNA from other organisms. In fact, a large proportion of the predicted proteins in this assembly showed the greatest sequence similarity to Saccharomonospora species instead of Streptomyces species (Studholme 2016). However, DNA contamination cannot explain why there are two copies of gyrB in another nine assemblies, as two assemblies are complete genomes, namely S. bingchenggensis BCW-1 (Genbank accession #

GCA_000092385.1) and S. cattleya DSM 46488 (Genbank accession # GCA_000240165.1). The sequence identities of two copies of gyrB in each genome range from 76% to 82% at the nucleotide (nt) level (Table 2-3). The phylogenetic tree based on gyrB nt sequences showed that two copies of gyrB were separated into two major clades (Figure 2-2), indicating the extra copies of gyrB may come from horizontal gene transfer (HGT) events rather than duplication events.

These results suggest gyrB is not an ideal phylogenetic marker for streptomycete systematics. To avoid utilizing homologous genes acquired from duplication or HGT events in phylogenetic inference, we decided not to use these five commonly used housekeeping genes to infer the phylogenetic relationships of the Streptomyces genus.

Phylogenomic Analysis of the Streptomyces Genus

To accurately infer the phylogenetic relationships of pathogenic Streptomyces species in the genus, we performed a comprehensive phylogenomic analysis using 111 Streptomyces genomes (Figure 2-3). We first identified 542 single-copy ortholog groups that are shared among all 111 Streptomyces isolates as well as two outgroups (Streptacidiphilus albus JL83 and

Kitasatospora setae NBRC 14216) using OrthoMCL v1.4 (Li et al. 2003). The nucleotide sequence alignments of 542 ortholog groups were filtered by removing poorly aligned sites; 109 ortholog groups that were less than 50% of the original alignment length after filtering were excluded. A concatenated alignment was made from the remaining 433 genes and used to infer the phylogenetic relationships of the 111 Streptomyces isolates by maximum likelihood (ML).

The ML tree showed strong bootstraps in almost all branches with only a few nodes receiving less than 70% bootstrap support (Figure 2-3).

S. scabiei isolates from different geographical regions were clustered closely together in the phylogram. Interestingly, Streptomyces stelliscabiei was phylogenetically closer to the non- pathogenic species Streptomyces bottropensis than to other pathogenic species including S. scabiei. Our analysis demonstrated that S. scabiei and S. bottropensis were phylogenetically distant from S. diastatochromogenes (Figure 2-3), although they were previously thought as closely related species and assigned to the Diastatochromogenes cluster (Elesawy and Szabo

1979; Healy and Lambert 1991). S. ipomoeae, which produces a less-modified thaxtomin derivative (thaxtomin C) (Guan et al. 2012), formed a separate clade from the cluster containing

S. scabiei, S. europaeiscabiei, and S. stelliscabiei. S. sp. 96-12 was originally classified as S. scabiei. Interestingly, our phylogenomic analysis grouped it with the non-pathogenic strain

Streptomyces galbus KCCM 41354 rather than with other phytopathogenic species. This result suggested that S. sp. 96-12 may be a recently emerged, novel pathogenic species. S. acidiscabies

84-104 and S. turgidiscabies Car8 were also phylogenetically grouped with non-pathogenic isolates, substantiating the notions that these two pathogenic species are emergent pathogens

(Huguet-Tapia et al. 2011; Huguet-Tapia and Loria 2012).

isDDH and ANI Comparison of Plant Pathogenic Streptomyces Isolates

To further define the evolutionary relationships of pathogenic Streptomyces isolates, we used two different in silico approaches: in silico DNA-DNA hybridization (isDDH) and average nucleotide identity (ANI), to assess genome similarity. Wet-lab DDH was a cornerstone of bacterial species determination, but it is tedious and difficult to standardize among laboratories

(Achtman and Wagner 2008). In contrast, in silico DDH data are reproducible and straightforward when genome sequences are available (Goris et al. 2007). The ANI of all genes shared by isolates provides a good estimate of evolutionary distance (Richter and Rosselló-Mora

2009; Colston et al. 2014; Zhang and Qiu 2015; 2016). Generally, the 95–96% ANI value corresponds to the 70% DDH standard; both values have been used as cutoffs in species discrimination (Goris et al. 2007).

Our analysis demonstrated that the ANI and isDDH values of each strain pair (Figure 2-

4) correlated well with their phylogenetic relationships in the ML tree (Figure 2-3), and strains in the same phylogenetic clade also showed high ANI and isDDH values. For example, ANI and isDDH values (Figure 2-4), as well as the core genome-based phylogeny (Figure 2-3) all indicated S. sp. 96-12 is distant from S. bottropensis, S. scabiei and S. europaeiscabiei, although its 16S rRNA sequence is 100% identical to S. bottropensis, and 99% identical to the latter two strains (data not shown). In addition, S. europaeiscabiei NCPPB 4086 was named S. scabiei

NCPPB 4086 when deposited in GenBank, but our analysis unambiguously classified it as a S. europaeiscabiei isolate. The isDDH values between S. europaeiscabiei NCPPB 4086 and other

S. europaeiscabiei isolates were approximately 73%. Given that a isDDH value of 79% is usually used as the lower limit in classifying subspecies (Meier-Kolthoff et al. 2013), our work indicated that S. europaeiscabiei NCPPB 4086 and other S. europaeiscabiei isolates may belong to two subspecies. Furthermore, S. bottropensis and S. stelliscabiei shared high ANI (97%) and

isDDH (75%) values. This result agreed well with phylogenomic analysis (Figure 2-3) and suggested that S. stelliscabiei and S. bottropensis should be classified as subspecies of a single species. Further comparative analysis of closely related pathogenic and non-pathogenic strains may be useful in identifying virulence genes.

Discussion

The genus Streptomyces is comprised of more than 600 species, making it the largest among all bacterial genera. The phylogeny of Streptomyces has been studied using a wide array of molecular methods, such as 16S rRNA and MLSA (Guo et al. 2008; Labeda 2011; Labeda et al. 2012). These studies have provided valuable insights into the evolutionary relationships of the species within the genus Streptomyces. However, the high sequence similarity of 16S rRNA genes among Streptomyces resulted in poor phylogenetic resolution (Labeda et al. 2012). The current Streptomyces MLSA scheme is useful in delineating intraspecific diversity (Rong et al.

2010). However, its application in systematics has mainly been evaluated by comparing its performance with 16S rRNA or individual housekeeping genes (Guo et al. 2008). Here, our phylogenomic and in silico genome similarity analyses (ANI and isDDH) demonstrated that whole-genome sequences are valuable resources for accurate inferring evolutionary relationships of the genus Streptomyces. The present study revealed that some pathogenic Streptomyces isolates were misidentified and S. sp. 96-12 belonged to a novel pathogenic Streptomyces species.

A previous phylogenetic analysis of Streptomyces and its relatives based on whole- genome sequences demonstrated that single-gene phylogenies were poorly resolved or even misleading, and the genome-based phylogeny provided a notably approved resolution (Alam et al. 2010). S. scabiei was previously considered to be most closely related to the

Diastatochromogenes group (Elesawy and Szabo 1979; Healy and Lambert 1991). However, our

analysis of the core genome-based phylogeny and in silico genome similarity analyses strongly suggested that S. diastatochromogenes is distantly related to S. scabiei. These results highlight the advantages of genome sequence-based classification in the delineation of taxonomically challenging genera.

Inaccurate identification of Streptomyces may have underestimated the number of pathogenic species. During the last two decades, 16S rRNA gene sequencing has been widely used for bacterial identification. However, insufficient variations among the 16S rRNA gene sequences of Streptomyces result in limited taxonomic resolution (Labeda 2011; Labeda et al.

2012). Many scab-causing Streptomyces isolates were identified as S. scabiei based on 16S rRNA sequences; however, increased phylogenetic resolution indicates that some of them should be classified as other species. For example, S. europaeiscabiei NCPPB 4086 was classified as S. scabiei NCPPB 4086 when deposited in GenBank, but our study indicates it belongs to the S. europaeiscabiei species. Another example is S. sp. 96-12, which was also originally named S. scabiei 96-12 by our group because its 16S rRNA is 99% identical to S. scabiei 87-22. Taken together, our results suggest that more pathogenic species exist in the natural environment than previously thought. Recently developed genome-based tools (Richter and Rosselló-Mora 2009;

Auch et al. 2010b) will facilitate the accurate identification of Streptomyces species in the future.

Genome sequences enabled us to determine a detailed and robust phylogeny of the

Streptomyces genus. However, phylogeny-based approaches can only infer evolutionary relationships, but do not provide a threshold for microbial species delineation; a standard usually based on DDH values. The empirical DDH technique requires stringent control of experimental variables and is performed in only a few specialized labs internationally (Rosselló-Mora 2006).

The increasing availability of whole genome sequences has triggered the development of

genomic-derived values to replace traditional DDH (Auch et al. 2010b; Goris et al. 2007; Richter and Rosselló-Mora 2009; Auch et al. 2010a). ANI and isDDH are two of the most widely accepted in silico tools used to determine strain relationships These analyses have been successfully applied in several bacteria genera, such as Aeromonas (Colston et al. 2014), Erwnia

(Zhang and Qiu 2015), and Ralstonia (Zhang and Qiu 2016). We also demonstrated that ANI and isDDH analyses are powerful tools to re-evaluate the phylogenetic relationships of the

Streptomyces genus. For example, S. stelliscabiei, first isolated from star-like common scab lesions on potato in France (Bouchek-Mechiche et al. 2000), should be renamed S. bottropensis.

Although Streptomyces is one of the most highly sequenced genera now, most recent systematic studies are still dependent on the MLSA scheme proposed by Guo et al.(Guo et al.

2008). This MLSA scheme follows the general recommendations for genetic marker selection, which is to use genes coding for subunits of ubiquitous enzymes (Glaeser and Kämpfer 2015).

However, our study strongly suggests that gyrB is not a good genetic marker for Streptomyces systematic analysis, due to the presence of more than one copy of gyrB in almost 10% (9 out of

111) tested Streptomyces genome assemblies. An increasing number of studies support the view that different genes have different phylogenetic inference power. Some commonly used genetic markers do not reliably infer evolutionary relationships among species (Townsend 2007;

Aguileta et al. 2008; Capella-Gutierrez et al. 2014; Salichos and Rokas 2013). Although the current Streptomyces MLSA scheme has well-established protocols and databases for the convenience of lab-to-lab portability and reproducibility (Guo et al. 2008), a thorough assessment may be required to check whether these markers are single-copy and whether their resulting phylogenies can accurately reflect the evolutionary relationships of organisms. With the

increasing availability of genome sequences, genome-based tools will facilitate the accurate identification of species in the future.

Table 2-1. Genome summary statistics of plant pathogenic Streptomyces spp. and the nonpathogenic Streptomyces diastatochromogenes Species Strain name Genome Contig Contig Length Geographic Genbank assembly Source or reference status number N50 (Kb) (Mb) location accession # S. scabiei 87-22 Complete 10.1 USA: GCA_000091305.1 (Bignell et al. 2010b) Wisconsin S. scabiei 84-232 Draft 1464 12.6 9.8 USA: West GCA_001572115.1 This study Virginia S. scabiei 84-34 Draft 1650 10.9 10.0 USA: upstate GCA_001550245.1 This study New York S. scabiei 85-08 Draft 1383 13.1 10.0 USA: Maine GCA_001550215.1 This study S. scabiei 95-18 Draft 1798 10.6 10.5 South Africa GCA_001550225.1 This study S. scabiei 96-06 Draft 1173 9.9 9.0 Canada: GCA_001579685.1 This study Ontario S. scabiei 96-08 Draft 1374 10.6 11.2 Canada: GCA_001550235.1 This study Alberta S. scabiei 96-15 Draft 1330 12.5 9.2 Germany GCA_001550295.1 This study S. scabiei NCPPB 4066 Draft 440 40.1 10.0 USA: New GCA_000738715.1 University of Exeter York S. scabiei NRRL B- Draft 1091 15.1 9.8 USA: central GCA_001005405.1 USDA/ARS/NCAUR 16523 New York S. europaeiscabiei NCPPB 4086 Draft 295 97.2 10.5 Canada: GCA_000738695.1 University of Exeter Ontario S. europaeiscabiei 89-04 Draft 1080 17.7 10.5 USA: Alaska GCA_001550325.1 This study S. europaeiscabiei 96-14 Draft 931 21.2 10.5 Australia GCA_001550375.1 This study S. europaeiscabiei NRRL B- Draft 1056 16.4 10.0 France: Le GCA_000988945.1 USDA/ARS/NCAUR 24443 Rhue S. stelliscabiei NRRL B- Draft 991 19.0 10.0 France GCA_001008135.1 USDA/ARS/NCAUR 24447 S. turgidiscabies Car8 Draft 692 27.4 10.8 Japan GCA_000331005.1 (Huguet-Tapia et al. 2011) S, acidiscabies 84-104 Draft 244 113.4 11.0 USA: upstate GCA_000242715.2 (Huguet-Tapia and New York Loria 2012) S. ipomoeae 91-03 Draft 687 26.4 10.4 USA GCA_000317595.1 (Huguet-Tapia et al. 2016) S. sp. 96-12 Draft 1875 8.7 10.1 Egypt GCA_001550315.1 This study S. diastatochromogenes ATCC 12309 Draft 1280 18.6 12.3 GCA_001550305.1 This study aFor complete genomes, N50 is not provided.

Table 2-2. Genomes of non-pathogenic Streptomyces species used in this study Species Strain name GenBank assembly Genome Contigs Contig accession # size (Mb) number N50 (Kb) S. achromogenes NRRL B-2120 GCA_000720835.1 8.1 163 185.7 S. alboflavus NRRL B-2373 GCA_000716675.1 9.7 171 134.3 S. alboviridis NRRL B-1579 GCA_000716935.1 8.6 263 264.8 S. albus NBRC 13014 GCA_000813365.1 7.6 117 128.8 S. antibioticus NRRL B-2032 GCA_001270635.1 10.0 343 81.3 S. anulatus NRRL B-2873 GCA_000721175.1 8.4 167 178.4 S, atroolivaceus NRRL ISP-5137 GCA_000717025.1 8.2 106 332.1 S. aureocirculatus NRRL ISP-5386 GCA_000720475.1 9.3 223 156.1 S. aureus NRRL B-1941 GCA_000725495.1 9.9 135 314.8 S. avellaneus NRRL B-3447 GCA_000721255.1 7.4 270 129.2 S. avermitilis MA-4680 GCA_000009765.1 9.1 1 S. avicenniae NRRL B-24776 GCA_000719135.1 6.5 66 276.6 S. bikiniensis NRRL B-1049 GCA_000716465.1 7.4 164 118.9 S. bingchenggensis BCW-1 GCA_000092385.1 11.9 1 S. bottropensis ATCC 25435 GCA_000383595.1 9.0 4 4 ,831.2 S. brasiliensis NRRL B-1626 GCA_000717015.1 7.8 67 247.5 S. californicus NRRL B-2098 GCA_000717645.1 7.8 105 251.9 S. catenulae NRRL B-2342 GCA_000718015.1 7.0 91 174.2 S. cattleya DSM 46488 GCA_000240165.1 8.1 2 S. celluloflavus NRRL B-2493 GCA_000720995.1 8.7 193 182 .0 S. cellulosae NRRL B-2687 GCA_000721105.1 10.1 182 192.2 S. citreofluorescens (anulatus) NRRL B-3362 GCA_000717105.1 8.4 178 201.4 S. coelicolor A3(2) GCA_000203835.1 9.0 3 S. cyaneofuscatus NRRL B-2570 GCA_000718135.1 7.9 127 353.8 S. davawensis JCM 4913 GCA_000349325.1 9.6 4 5,836.1 S. decoyicus NRRL 2666 GCA_001270575.1 8.6 304 75.0 S. durhamensis NRRL B-3309 GCA_000725475.1 9.3 228 256.6 S. erythrochromogenes NRRL B-2112 GCA_000725555.1 7.5 134 235.7 S. flaveolus NRRL B-1589 GCA_000716535.1 10.3 232 194.9 S. flavidovirens DSM 40150 GCA_000429085.1 7.1 53 241.4 S. flavochromogenes NRRL B-2684 GCA_000717595.1 8.9 170 235.9 S. flavovariabilis NRRL B-16367 GCA_000725785.1 8.8 142 285.9 S .flavovirens NRRL B-2182 GCA_000720395.1 7.7 97 371.3 S. floridae NRRL2423 GCA_000717665.1 7.9 67 393.2 S. fulvissimus DSM 40593 GCA_000385945.1 7.9 1 S. fulvoviolaceus NRRL B-2870 GCA_000718165.1 11.6 263 163.8 S. fulvoviridis NRRL ISP-5210 GCA_000717065.1 7.4 67 518.3

Table 2-2. Continued Species Strain name GenBank assembly Genome Contig Contig accession # size (Mb) number N50 (Kb) S. galbus KCCM 41354 GCA_000772895.1 10.1 104 241.6 S. graminis NRRL B-1570 GCA_000716525.1 7.9 334 166.0 S. griseofuscus NRRL B-5429 GCA_000718315.1 8.9 244 135.2 S. griseolus NRRL B-2925 GCA_000721185.1 7.5 139 201.6 S. griseoluteus NRRL ISP-5360 GCA_000717735.1 8.1 190 112.0 S. griseus NBRC 13350 GCA_000010605.1 8.5 1 S. halstedii NRRL-ISP 5068 GCA_000720535.1 7.7 105 273.7 S. iakyrus NRRL ISP-5482 GCA_000717055.1 9.1 181 287.3 S. incarnatus NRRL8089 GCA_001027185.1 8.9 45 401.5 S. katrae NRRL B-16271 GCA_000717655.1 8.6 151 193.1 S. lavenduligriseus NRRL ISP-5487 GCA_000718625.1 8.7 217 193.2 S. leeuwenhoekii DSM 42122 GCA_001013905.1 8.1 3 S. lividans TK24 GCA_000739105.1 8.3 1 S. mediolani NRRL WC-3934 GCA_000721685.1 8.1 98 339.5 S. megasporus NRRL B-16372 GCA_000718985.1 5.9 195 156.0 S. natalensis ATCC 27448 GCA_000935125.1 8.7 63 375.9 S. niger NRRL B-3857 GCA_000718305.1 8.5 108 381.7 S. nodosus ATCC 14899 GCA_000819545.1 7.7 111 110.8 S. ochraceiscleroticus NRRL ISP-5594 GCA_000720485.1 8.4 147 212.0 S. olivaceus NRRL B-1125 GCA_000716815.1 7.4 65 518.3 S. peucetius NRRLWC-3868 GCA_000725565.1 9.5 129 220.2 S. pratensis ATCC 33331 GCA_000176115.2 7.7 3 S. prunicolor NBRC 13075 GCA_000367365.1 11.8 202 124.7 S. puniceus NRRL ISP-5083 GCA_000718695.1 8.2 83 276.6 S. purpeochromogenes NRRL B-3012 GCA_000717965.1 7.9 126 178.0 S. purpeofuscus NRRL B-1817 GCA_000718025.1 9.3 138 241.2 S. purpureus KA281 GCA_000384175.1 7.5 24 573.2 S. pyridomyceticus NRRL ISP-5024 GCA_000717725.1 6.3 139 107.1 S. regensis NRRL B-11479 GCA_001047335.1 15.1 960 71.4 S. resistomycificus NRRL 2290 GCA_001270505.1 9.2 1 S. roseoverticillatus NRRL B-3500 GCA_000719265.1 7.2 96 256.2 S. rubellomurinus ATCC 31215 GCA_000961885.1 8.6 411 74.3 S. ruber NRRL ISP-5378 GCA_000725575.1 9.7 125 184.7 S. scabrisporus DSM 41855 GCA_000372745.1 11.4 199 181.0 S. sclerotialus NRRL B-2317 GCA_000719105.1 7.0 92 262.6 S. sclerotialus NRRL ISP-5269 GCA_000720555.1 8.2 112 265.3 S. scopuliridis NRRL B-24574 GCA_000718095.1 8.2 117 275.1

Table 2-2. Continued Species Strain name GenBank assembly Genome Contig Contig accession # size (Mb) number N50 (Kb)a S. seoulensis NRRL B-24310 GCA_000725795.1 6.4 76 273.7 S. sulphureus DSM 40104 GCA_000381025.1 7.1 34 354.6 S. variegatus NRRL B-16380 GCA_000955965.1 8.8 233 220.7 S. varsoviensis NRRL ISP-5346 GCA_000718635.1 8.6 118 268.2 S. venezuelae ATCC 10712 GCA_000253235.1 8.2 114 112.6 S. vietnamensis GIMV4.0001 GCA_000830005.1 9.2 2 S. vinaceus NRRL ISP-5257 GCA_000717215.1 8.1 72 354.5 S. violaceoruber NRRL S-12 GCA_000717845.1 8.5 170 148.2 S. violaceorubidus NRRL B-16381 GCA_000717995.1 7.7 282 176.4 S. violens NRRL ISP-5597 GCA_000717745.1 7.9 64 355.8 S. virginiae NRRL ISP-5094 GCA_000720455.1 8.3 133 284.2 S. vitaminophilus DSM 41686 GCA_000380165.1 6.4 113 136.3 S. wedmorensis NRRL 3426 GCA_000716445.1 9.4 259 125.2 S. xanthophaeus NRRL B-5414 GCA_000725805.1 8.6 174 240.1 S. xylophagus NRRL B-12029 GCA_000716435.1 11.4 211 356.0 S. yeochonensis CN732 GCA_000745345.1 7.8 6 4,825.6 S. yerevanensis NRRL B-16943 GCA_000716805.1 10.9 323 150.5 aFor complete genomes, N50 is not provided.

Table 2-3. The list of Streptomyces strains that have two-copies of MLSA genes Gene Strain namea Nucleotide identities (% nt) gyrB S. albus NBRC 13014 81 S. bingchenggensis BCW-1 82 S. cattleya DSM 46488 82 S. avicenniae NRRL B-24776 80 S. graminis NRRL B-1570 81 S. peucetius NRRL WC-386 82 S. scabrisporus DSM 41855 80 S. vitaminophilus DSM 41686 81 S. xanthophaeus NRRL B-5414 76 recA S. fulvissimus DSM 40593 93 S. yerevanensis NRRL B-16943 84 trpB S. flavidovirens DSM 40150 92 aFive genes are not present as single-copy in Streptomyces regensis NRRL B-11479 (Genbank accession # LFVR00000000.1). Since this assembly may contain more than one organism, it was not included in the table.

Figure 2-1. Virulence phenotypes of plant pathogenic Streptomyces isolates. A) Surface- sterilized newly germinated radish seedlings (six per treatment) of similar size were inoculated with water (negative control) or with mycelia from tryptic soy broth- grown Streptomyces cultures. Plants were then incubated at room temperature under a 12-h photoperiod for 4 days. All seedlings infected with pathogenic isolates showed a stunted phenotype and representative plants for each treatment are shown. B) Production of thaxtomin A by plant pathogenic Streptomyces isolates. Isolates were grown in triplicate in OBB medium for 6 days at 28°C. The average thaxtomin production of S. scabiei 87-22 that is normalized using dried cell weights is set to 100%. The average % production for each strain relative to S. scabiei 87-22 and normalized using dried cell weights is shown. The errors bars represent the standard deviation from the mean. Letters represent results of a one-way ANOVA with Tukey’s HSD test; bars not sharing letters are significantly different at P < 0.05.

Figure 2-2. Phylogenetic relationships among nine Streptomyces strains based on gyrB gene sequences. These nine strains have two copies of gyrB genes, which are designated gyrB1 and gyrB2. The gyrB gene of S. coelicolor A3 (2) was selected as the reference. The tree was constructed using the ML method from 1132 base pairs of gyrB genes after trimming and gap removal. Numbers at nodes represent levels (%) of bootstrap support from 1000 resampled datasets.

Figure 2-3. Phylogenetic reconstruction of the genus Streptomyces based on the concatenation of all 433 single-copy gene clusters from 111 Streptomyces genomes with the maximum likelihood algorithm RAxML. 1000 bootstrap replicates were computed. Numbers above branches are bootstrap percentages. Streptacidiphilus albus JL83 and Kitasatospora setae NBRC 14216 were selected as outgroups (not shown on tree).

Figure 2-4. ANI and isDDH values of plant pathogenic and closely related non-pathogenic Streptomyces isolates. The lower triangle displays ANI values, and the upper triangle displays isDDH values. Displayed isDDH values are point estimates plus the 95% model-based confidence intervals. Boxes with ANI ≥ 99% or isDDH ≥ 79% are colored red, and boxes with ANI ≥ 96% and < 99% or isDDH ≥ 70% and < 79% are colored orange. Isolates belonging to the same species are marked with the same color.

CHAPTER 3 PROMISCUOUS PATHOGENICITY ISLANDS OF STREPTOMYCES SPP.

Streptomyces are viewed predominantly as free-living terrestrial soil-inhabiting bacteria, but some species are symbionts with fungi, insects, plants and animals (Seipke et al. 2012).

Furthermore, some strains have been isolated from marine soils (Fiedler et al. 2005). Eleven

Streptomyces species are plant pathogens and cause diseases of underground crop structure such as roots and tubers (Dees and Wanner 2012). Streptomyces scabiei, Streptomyces acidiscabies,

Streptomyces turgidiscabies and Streptomyces ipomoeae are well characterized (Loria et al.

2006). However, another seven pathogenic species have been the subject of relatively little research: namely Streptomyces europaeiscabiei, Streptomyces stelliscabiei, Streptomyces reticuliscabiei, Streptomyces caviscabies, Streptomyces luridiscabiei, Streptomyces puniciscabiei, and Streptomyces niveiscabiei (Loria et al. 2006; Goyer et al. 1996a). Potato scab, characterized by raised, pitted, or superficial lesions formed on the tuber surface is the most economically important disease caused by pathogenic Streptomyces species (Loria et al. 2006).

Mobilization of pathogenicity islands (PAIs) is thought to be a driving force in the evolution of plant pathogenic streptomycetes (Loria et al. 2006). Bioinformatics analyses of the genome sequences of S. scabiei 87-22 (Bignell et al. 2010b), S. acidiscabies 84-104 (Huguet-Tapia and

Loria 2012), S. turgidiscabies Car8 (Huguet-Tapia et al. 2014) and S. ipomoea 91-03 (Guan et al.

2012; Huguet-Tapia et al. 2016) revealed related but not identical PAIs.

The 674-Kb PAI discovered in S. turgidiscabies (Kers et al. 2005; Huguet-Tapia et al.

2014) was the first PAI described in any Gram-positive plant pathogen. This PAI carries genes encoding four pathogenicity factors: the biosynthetic pathway for thaxtomin A (ThxA), a functional tomatinase (TomA), a small, secreted protein (Nec1), and a cytokinin biosynthetic pathway (Fas) (Huguet-Tapia et al. 2014). ThxA is a nitrated 2,5-diketopiperazine synthesized

via two nonribosomal peptide synthetases (NRPS) encoding by txtA and txtB genes, a P450 monooxygenase (TxtC), a nitric oxide synthase (TxtD), and a novel cytochrome P450 (TxtE) that site specifically nitrates tryptophan prior to cyclization (Bignell et al. 2014b). A 65 amino acid protein encoded by txtH is a predicted MbtH-like family protein and may be important in regulating NRPS activity (Herbst et al. 2013; Stegmann et al. 2006). ThxA inhibits cellulose biosynthesis in plant cell walls, leading to abnormal cell expansion and cell death (Loria et al.

2008). TomA and Nec1 are not essential for pathogenicity but play roles in the infection process, possibly by suppressing plant defenses (Seipke and Loria 2008; Joshi et al. 2007b). These pathogenicity/virulence genes also exist in S. scabiei and S. acidiscabies; but they are separated in two remote chromosomal regions, designated the toxicogenic region (TR) and the colonization region (CR) (Lerat et al. 2009). The TR region contains the ThxA biosynthetic gene cluster, while the CR region includes nec1 and tomA. In addition to these genes, the PAI in S. turgidiscabies contains the fas operon with genes that are homologous to those in Rhodococcus fascians (Joshi and Loria 2007); Fas proteins mediate the biosynthesis and modification of a mix of cytokinins, which contribute to virulence phenotypes in both pathogens (Pertry et al. 2010).

In Chapter 2, we demonstrated that the pathogenic isolate Streptomyces sp. 96-12 from

Egypt belonged to a novel pathogenic species. In this study, we showed that S. sp. 96-12 acquired the TR region from S. scabiei directly or indirectly, likely in an agricultural environment. Remarkably, we further demonstrated that the S. scabiei TR region can transfer and site-specifically integrate into the chromosome of saprophytic species under laboratory conditions. Acquisition of TR conferred a pathogenic phenotype on saprophytic species. These results confirm the mobility of S. scabiei TR to non-pathogenic species both in vivo and in vitro,

and lay the groundwork to further understand and predict the pathogenic development of plant pathogens.

Materials and Methods

Phylogeny of the Thaxtomin Biosynthetic Cluster

Nucleic acid sequences of txtA, txtB, txtC, txtD, txtE, and txtR were aligned and manually corrected to remove ambiguous positions using MEGA 6. 06 (Tamura et al. 2013). Alignments of these six thaxtomin A biosynthetic genes were concatenated and submitted to the RAxML webserver (https://www.phylo.org/portal2/) (Stamatakis et al. 2008) for Maximum Likelihood

(ML) inference with the GTR+G substation model, including an estimation of invariable sites.

For inferring tree robustness, 1000 bootstrap replicates were computed.

Insertion of Antibiotic Resistance Markers in the PAI

Insertion of the antibiotic resistance markers into the TR1 or TR2 region in S. scabiei 87-

22 was accomplished by replacing txtH (SCAB_31771) with the apramycin resistance gene or replacing the lantibiotic biosynthesis genes lanA (SCAB_32021) and lanB (SCAB_32031) with the hygromycin B resistance gene using the Redirect PCR targeting system (Gust et al. 2003).

The cosmid 1989 (harboring the txtH gene) or the cosmid 2757 (harboring the lantibiotic biosynthesis genes) was introduced in Escherichia coli BW25113 strain harboring the arabinose- inducible λ red expression plasmid pIJ790 (aac(3)IV + oriT) or pIJ10700 (hyg + oriT) (Table 3-

1). The deletion cassette for txtH was PCR amplified using pIJ773 as the template and using primers DRB201 and DRB202 (Table 3-2), and the deletion cassette for lanA and lanB was PCR amplified using pIJ10700 as the template and using primers DRB431 and DRB432 (Table 3-2).

The gel-purified deletion cassettes were electroporated into the E. coli BW25113 cells that contained either the cosmid 1989 or cosmid 2757 and which had been induced with arabinose

(20 mM final concentration). The txtH and lanAB mutant cosmids were then isolated and

confirmed by PCR and sequencing. Following transformation of each cosmid into the non- methylating E. coli ET12567 strain (Table 3-1) (MacNeil et al. 1992), which contains the plasmid pUZ8002 as a driver for transfer (Bierman et al. 1992), the cosmids were conjugated into S. scabiei 87-22 on soy flour mannitol (SFM) agar (Kieser et al. 2000). Resulting exconjugants for ∆txtH were selected for resistance to apramycin (aprR) and sensitive to kanamycin, and exconjugants for ∆lanAB were selected for resistance to hygromycin B (hygR) and sensitive to kanamycin. The mutant strains were confirmed using PCR and were stored as spore stocks in 20% vol/vol glycerol at -80 C.

Selection of Transconjugant Strains

Conjugal mating was carried out using either S. scabiei 87-22 ∆txtH (aprR) or S. scabiei

87-22 ∆lanAB (hygR) as a PAI donor strain and naturally streptomycin-resistant (strR) S. diastatochromogenes as the recipient strain. Strains were co-cultured on SFM plates (Kieser et al. 2000) for 2 days at 28 °C. A soft nutrient agar overlay containing 20 µg/ml streptomycin and

50 µg/ml apramycin or 100 µg/ml hygromycin B (final concentration) was used to select for aprRstrR or hygRstrR transconjugants. Qualitative PCR assays were conducted to detect the integration of S. scabiei TR into the chromosome of S. diastatochromogenes (Table 3-2, Figure

3-5). Primers a + b flanking the att site located at the 3’ end of the S. diastatochromogenes aviX1 gene were used to detect the site-specific integration of TR in aviX1 (final PCR products for S. diastatochromogenes wild type, 553 bp). Primers e + f (final PCR products, 731 bp) were used to amplify the chromosomal junction formed by integration of TR or TR2 into S. diastatochromogenes aviX1. Primers c + d that are located within the integrase gene

(SCAB_31871) of TR2 were used to confirm TR2 integration (final PCR products, 656 bp).

Primers g + h flanking txtH (SCAB_31771) were used to confirm TR1 integration (final PCR products, 450 bp). All PCR products were sequenced.

Virulence Bioassays

To assess the virulence phenotype of strains, an in vitro radish seedling bioassay was performed as described previously (Bignell et al. 2010b) except that “German Giant” seeds

(Burpee) were used in place of the “Burpee White” variety. The assay was performed in triplicate with each three different biological replicates.

Analysis of Thaxtomin A Production

Plant pathogenic Streptomyces strains were cultured in TSB for 48 to 72 h. Mycelia were then pelleted by centrifugation, washed twice with sterile water, and re-suspended in sterile water to obtain an OD600 of 1.0. The mycelial suspension (400 µl) was then inoculated into 3 ×

50 ml of OBB medium or OBB medium supplemented with 1% cellobiose (OBBc) in 250 ml flasks. After incubation for 6 days at 28 °C with shaking (250 rpm), thaxtomin A was purified from the culture supernatants and analyzed by HPLC as described previously (Johnson et al.

2007). All experiments were repeated at least three times using different biological replicates of the Streptomyces strains, with three technical repeats per strain. The pellets of the culture samples were dried and weighted as a measure of bacterial growth. The data for S. scabiei 87-22 in OBB was normalized to 100 %. The data for all pathogenic isolates were square-root transformed to improve homogeneity of variance. Transformed data were subjected to one-way

ANOVA using SAS 9.4 (SAS Institute, Cary, NC), and means were compared based on Tukey's

HSD test at P < 0.05.

Results

Streptomyces Pathogens Contained Different Versions of PAIs

To understand the emergence and evolution of plant pathogenicity, we focused on comparative genomic analysis of PAIs in various Streptomyces species (Figure 3-1). S. turgidiscabies Car8 contains a uniquely large PAI of 674 Kb, which includes the ThxA biosynthesis cluster, the virulence genes nec1 and tomA, and a fas operon (Huguet-Tapia et al.

2011). In contrast, pathogenicity genes in the PAIs of other pathogenic species are separated into two distinct regions (TR and CR) and the fas operon is missing (Figure 3-1). One internal and two flanking attachment (att) sites of the S. scabiei TR region further separate its TR region into two sub-regions, designated TR1 and TR2 (Chapleau et al. 2016). The 20 Kb TR1 region contains the ThxA biosynthetic cluster, and the 157 Kb TR2 region includes putative integrative and conjugative elements (ICEs) (Chapleau et al. 2016). Our analysis demonstrated that all pathogenic isolates possess the TR1 region, indicating the importance of ThxA in disease development. The TR2 region is conserved only in S. scabiei isolates and S. sp. 96-12. Our genome based evolutionary analysis, reported in Chapter 2, revealed that S. sp. 96-12 did not cluster with S. scabiei (Figures. 2-3 and 2-4). Since TR1 and TR2 are almost identical in S. sp.

96-12 and S. scabiei, in vivo mobilization of these regions is a likely scenario. The TR of S. stelliscabiei NRRL B-24447 contains genes encoding the integrase and the recombination directionality factor (RDF) that are essential for integration and excision (Wozniak and Waldor

2010), but we failed to identify the TR2 region due to the draft genome's fragmentation (Figure

3-1). Furthermore, the S. stelliscabiei NRRL B-24447 integrase only had 83% sequence identity with the integrase of S. scabiei 87-22 at the amino acid level (data not shown). The boundaries and mobilization mechanism of S. stelliscabiei NRRL B-24447 PAI will require further research.

The CR region containing virulence genes nec1 and tomA is missing in S. europaeiscabiei NCPPB 4086 and S. sp. 96-12. Since the CR region is 4.9-Mb away from the TR region, these two regions may have been transferred into pathogenic Streptomyces genomes via different HGT events. In S. turgidiscabies Car8, the circular structure of the CR region from site- specific recombination at two flanking att sites has been detected (Huguet-Tapia et al. 2014).

However, the mobilization ability of the CR region itself requires further investigation. S. europaeiscabiei NCPPB 4086 and S. sp. 96-12 may have lost the CR region during evolution, but it is also possible that these two isolates have never acquired this region.

ThxA-phylogeny Contradicts the Core-genome Phylogeny

Almost all scab-causing Streptomyces species produce ThxA, and the transfer of the PAI containing the characteristic ThxA biosynthetic gene cluster is likely to be the genetic basis for emergence of new pathogenic Streptomyces species. Phylogenetic analysis of ThxA biosynthetic genes may thus infer the mechanism of PAI mobilization and its evolution. To test this, we constructed a phylogenetic tree of selected pathogens using their txtA, txtB, txtC, txtD, txtE and txtR genes (Figure 3-2). In this analysis, S. acidiscabies 84-104, all S. scabiei isolates, and S. sp.

96-12 were closely grouped, indicating the high similarity of their ThxA clusters (Figure 3-3). In contrast, S. acidiscabies 84-104, S. sp. 96-12 and S. scabiei were distantly related in the core- genome tree (Figure 3-3). The high similarity of the ThxA clusters in these three phylogenetically distant pathogenic taxa (S. acidiscabies, S. sp. 96-12 and S. scabiei) strongly suggested recent HGT.

The S. scabiei PAI was Site-specifically Mobilized into the Chromosome of Non-pathogenic Species

Our bioinformatics analyses suggested the direct or indirect in vivo mobilization of the

TR between S. sp. 96-12 and S. scabiei (Figure 3-3). In all S. scabiei strains and in S. sp. 96-12,

the whole TR region is integrated at the 3 end of an aviX1 gene (SCAB_33441, encoding a putative ATP/GTP binding protein) (Figure 3-1). The same integration occurred in S. acidsicabies and S. europaeiscabiei strains, although they only acquired the TR1 region (Figure

3-1). These results suggest that the 3’ end of the aviX1 gene of Streptomyces species is a common location for site-specific integration of the TR region.

We aimed to further understand TR mobilization and evolution by investigating the transfer of pathogenicity genes from S. scabiei to saprophytes in vitro. We first created two S. scabiei 87-22 deletion mutants, one with the apramycin resistance (aprR) marker inserted in TR1

[S. scabiei ΔtxtH (aprR)] and the other with the hygromycin B (hygR) resistance marker inserted in TR2 [S. scabiei ΔlanAB (hygR)]. In mating experiments, these mutants served as the PAI donors, while a naturally streptomycin resistant (strR) S. diastatochromogenes was used as the recipient strain (Figure 3-4A). Following selection with apramycin or hygromycin B, together with streptomycin, a number of S. diastatochromogenes mutants were isolated. We found that the aviX1 gene of S. diastatochromogenes shares 93% identity with S. scabiei aviX1

(SCAB_33441) at the nucleic acid level (data not shown) and an identical att site

(TTGAAGCGGAAC) (Figure 3-4B). We mapped the mobilized PAI in multi-antibiotic resistant

S. diastatochromogenes strains: the site-specific integration of the S. scabiei PAI into S. diastatochromogenes was examined using PCR primers flanking the att site at the 3’ end of the

S. diastatochromogenes aviX1 gene and other primers within TR1 or TR2 (Table 3-2). We detected the integration of different S. scabiei TR regions into the att site of the S. diastatochromogenes aviX1 gene (Figure 3-4C). Mating of S. scabiei ΔlanAB with S. diastatochromogenes resulted in the acquisition of TR2 alone or of the complete TR element by

S. diastatochromogenes; however, all transconjugants from the mating of S. scabiei ΔtxtH with

S. diastatochromogenes acquired the entire TR element but not TR1 alone. These observations suggested that the ICEs of the TR2 region are important to the mobilization and site-specific integration of TR2 or the entire TR region into a new host chromosome. In contrast, the TR1 region carrying the ThxA biosynthetic cluster is not mobile alone, and its transfer depends on the

ICEs in the TR2 region.

PAI Acquisition Conferred a Pathogenic Phenotype in S. diastatochromogenes Transconjugants

To determine the biological consequences of PAI transfer, we characterized S. diastatochromogenes transconjugants carrying different versions of TR by determining their production of ThxA in OBB and OBBc and evaluating their pathogenicity on radish seedlings. S. diastatochromogenes transconjugants with TR2 ΔlanAB alone produced an undetectable level of

ThxA (data not shown) and were not pathogenic on radish seedlings (Figure 3-4D). On the other hand, transconjugants that acquired the entire TRΔlanAB produced ThxA (Figure 3-4E) and were pathogenic (Figure 3-4D). Interestingly, the ThxA production level of these transconjugants was significantly higher than that of S. scabiei 87-22 in both OBB and OBBc. S. diastatochromogenes transconjugants carrying TRΔtxtH produced trace amounts of ThxA and two new metabolites with a distinctive yellow appearance (data not shown). HPLC and LC-MS analysis identified the new compounds as thaxtomin D and monohydroxylated thaxtomin D, presumably thaxtomin B (data not shown). The production of these two metabolites may be the result of insertion of the deletion cassette upstream of txtC, which encodes a P450 monooxygenase responsible for converting thaxtomin D into ThxA (Healy et al. 2002). As expected,

S. diastatochromogenes transconjugants carrying TRΔtxtH were not pathogenic in the radish seedlings assays (Figure 3-4D). Nonetheless, these results clearly demonstrated that the TR region of S. scabiei-type PAI conferred pathogenicity to a non-pathogenic Streptomyces species.

Discussion

The capacity for mobilization of the S. scabiei TR to saprophytic Streptomyces isolates, similar to what has been observed with the S. turgidiscabies PAI (Kers et al. 2005), has been suggested (Chapleau et al. 2016) but not confirmed. Here, we inserted two different antibiotic resistance markers into the S. scabiei TR1 and TR2 regions, and monitored their mobilization and site-specific integration into the chromosome of non-pathogenic species. We found that although TR1, TR2, and the entire TR region can be excised from the S. scabiei chromosome, only TR2 or the entire TR region was transferred into the recipient genome. Accordingly, recipients became pathogenic only after acquiring the entire TR region (Figure 3-4). These results suggested that TR2 is required for mobilization and TR1 is required for pathogenicity and, more importantly, they are both involved in the emergence of novel plant pathogenic

Streptomyces species in agricultural systems. Furthermore, the data presented here and in

Chapter 2 suggested that the TR region of S. sp. 96-12 may have been acquired, directly or indirectly, from S. scabiei through a recent HGT event.

Despite being only distantly related to S. scabiei, S. diastatochromogenes is still able to acquire the pathogenicity phenotype upon the acqusition of the S. scabiei PAI. However, very few new plant pathogenic Streptomyces species have been described since 1990 (Loria et al.

2008). The rarity of the pathogenicity phenotype within Streptomyces indicates that either PAI mobilization and stable integration is very infrequent in nature, or that only a small proportion of recipients can express the pathogenicity phenotype upon acquisition of the PAI. Alternatively, there may be other genetic elements responsible for viability and reproduction inside plant tissue, that have yet to be discovered, but are required for pathogenicity in nature. It is important to note that among characterized pathogens, only S. scabiei and S. sp. 96-12 harbor the TR2 region, which is prerequisite for the mobilization of the thaxtomin-containing TR1. The TR1 regions

may have become fixed in those pathogens lacking TR2. The ThxA biosynthetic clusters of S. stelliscabiei and S. europaescabiei are distantly related to those of S. scabiei, S. acidiscabies, and

S. sp. 96-12. We also observed significant genetic diversity among different S. europaescabiei isolates. It is possible that the ThxA biosynthetic clusters of S. europaescabiei and S. stelliscabiei have been fixed and genetic variation has accumulated after the TR fixation. However, the mobility of TR regions in S. europaescabiei and S. stelliscabiei requires further investigation.

Our comparative genomic analysis of PAIs revealed their diverse evolutionary paths and highlighted the central role of ThxA in plant pathogenesis. S. scabiei is thought to be the most ancient scab-causing Streptomyces species (Bukhalid et al. 2002). Furthermore, the presence of the TR2 region is rare among pathogens. It may be that TR1 and TR2 were co-transferred into the last common ancestor of S. scabiei, S. europaescabiei and S. stelliscabiei. Subsequently, the

TR2 region was lost in S. europaescabiei, leading to the fixation of the TR1 region in its genome. In contrast, the TR2 region retained in S. scabiei has been a driving force in the mobilization of TR1 to otherwise-non-pathogenic species, such as S. acidiscabies and S. sp. 96-

12. Another possible scenario is that the TR1 region itself was transferred to the last common ancestor of S. scabiei, S. europaescabiei and S. stelliscabiei via mechanisms independent of the

TR2 region. S. scabiei may have acquired the TR2 region subsequently and with it, the ability to transfer the TR1 region to other species, leading to the birth of novel pathogenic Streptomyces species. Due to the fragmentation of S. stelliscabiei’s draft assembly, we failed to identify the complete TR2 region. A high-quality genome assembly of S. stelliscabiei is required for reconstructing the evolution history of its TR region.

In summary, the results of this study revealed the horizontal transfer of the TR of

Streptomyces scabiei, and inferred its roles in the emergence of new pathogenic species. With the

increasing availability of genome sequences, it will be possible to characterize the diversity of

PAIs of plant pathogenic Streptomyces spp., and obtain a more comprehensive picture of the origin and evolution of pathogenicity in the Streptomyces genus.

Table 3-1. Bacterial strains, plasmids, and cosmids used in this study Strain or plasmid Description† Source or reference E. coli strains DH5α General cloning host Gibco-BRL BW25113 Host for the REDIRECT© PCR targeting system (Gust et al. 2003) ET12567 dam-, dcm-, hsdS-; non-methylating host for transfer of DNA (MacNeil et al. into Streptomyces spp. (cmlR, tetR) 1992) Plasmids or cosmids pIJ790 λ Red plasmid (tS, cmlR) (Gust et al. 2003) pUZ8002 Supplies transfer functions for mobilization of oriT- (Kieser et al. 2000) containing vectors from E. coli to Streptomyces (kanR) pIJ773 Template for the REDIRECT© PCR targeting system, (Gust et al. 2003) contains the [aac(3)IV+oriT] disruption cassette (ampR, aprR)

Cosmid 1989 SuperCos1 derivative containing the S. scabiei 87-22 txtH This study locus (kanR, ampR) Cosmid 2757 SuperCos1 derivative containing the S. scabiei 87-22 lanAB This study locus (kanR, ampR) Note: † aprR, apramycin resistance; cmlR, chloramphenicol resistance; tetR, tetracylcin resistance; ts, temperature sensitive; kanR, kanamycin resistance; ampR, ampicillin resistance; thioR, thiostrepton resistance

Table 3-2. List of primers used in this study Primer Sequence 5’- 3’ Use a TCCACCTCCTGACCACCAAG Detect the site-specific integration of TR in b AAGATCCCCGAACCGACCT S. diastatochromogenes aviX1 c GTAGCGAAGGCGAGAGTCTCACTG PCR amplification of the integrase gene d GAGCCGACGAACAAGTACTACCCG (SCAB_31871) located in the TR2 region e CGAAGATCGAGAACGTCAGGAAGG Detect the site-integration of TR2 or the f GACCGACGAGGACTTCAAGAACGA whole TR in S. diastatochromogenes aviX1 g TACGAGACCATCGGCAGGGA PCR amplification of txtH (SCAB_31771) h ACATCCTCACCGAGCCGGAA located in the TR1 region DRB201 TGCCGGGCCCTCTTTGCCGACTAGGAGAAA txtH (SCAB_31771) Redirect deletion TTCACCGTGTGTAGGCTGGAGCTGCTT cassette DRB202 GGCGACCCGTGGCCCCGCTCGATGTTATTG GCCGGGTCAATTCCGGGGATCCGTCGACC DRB431 ATGAAGAACTTCGAAGCCGCGACCACTCA lanA (SCAB_32021) and lanB GGTCGATGTGTGTAGGCTGGAGCTGCTTC (SCAB_32031) Redirect deletion cassette DRB432 TCACGGCGTCCTCCAGTGTTCGCGGGCGCT CTGGCGCAGATTCCGGGGATCCGTCGACC

Figure 3-1. Schematic genetic organization of PAIs from different plant pathogenic Streptomyces species. TR: the toxicogenic region, CR: the colonization region. Arrowed boxes represent the location and orientation of the open reading frames (ORFs) of ThxA biosynthetic genes (red), integrase and recombination directionality factor (RDF) (green), lantibiotic biosynthetic genes (orange), conjugative genes (blue), aviX1 (brown), bacA (light blue), nec1 (black), tomA (purple), sugar catabolism clusters (gray), and fas operon (yellow). The sequences of attachment sites at the junctions for TR1/TR2/TR are also shown.

Figure 3-2. Phylogenetic reconstruction of scab-causing Streptomyces strains based on the thaxtomin A biosynthetic cluster (txtA, txtB, txtC, txtD, txtE and txtR). 1000 bootstrap replicates were computed. Numbers above branches are bootstrap percentages.

Figure 3-3. Comparison of phylogenetic relationships derived from the core-genome (left) and the thaxtomin A biosynthetic cluster (right). The core-tree is from Figure 2-3 and the ThxA-tree is from Figure 3-2.

Figure 3-4. Excision and site-specific integration of the S. scabiei toxicogenic region (TR) into the S. diastatochromogenes chromosome. A) Schematic representation of the mating experiments between S. scabiei 87-22 deletion mutants and S. diastatochromogenes. B) The PAI integrates site-specifically into the attachment (att) site located at the 3’ end of the aviX1 gene. C) Cartoon illustrating the three excision forms and two site- specific integration forms of S. scabiei TR: TR1 alone, TR2 alone, and the whole TR can excise from the S. scabiei chromosome, but only TR2 alone and the whole TR can integrate into the S. diastatochromogenes chromosome (The entire TR integration can use either attP’ and attP’’, but our analysis only detected the integration using the attP’ site that resulted from the recombination between attL and attR). D) Surface- sterilized radish seedlings (six per treatment) were inoculated with mycelia from tryptic soy broth-grown cultures of S. diastatochromogenes transconjugants. Seedlings treated with S. scabiei 87-22, the non-pathogenic species S. diastatochromogenes or with water were used as controls. Plants were incubated at room temperature under a 12-h photoperiod for 4 days. Representative plants for each treatment are shown. E) Production of thaxtomin A by S. diastatochromogenes transconjugants that acquired the entire TRΔlanAB in OBB medium and OBB medium supplemented with 1% of cellobiose (OBBc). Isolates were grown in triplicate in OBB or OBBc medium for 6 days at 28 °C. The average thaxtomin production of S. scabiei 87-22 in OBB that is normalized using dried cell weights is set to 100%. The average % production for each strain relative to S. scabiei 87-22 in OBB and normalized using dried cell weights is shown, and the errors bars represent the standard deviation from the mean. Letters represent results of a one-way ANOVA with Tukey’s HSD test; bars not sharing letters are significantly different at P < 0.05.

Figure 3-4. Continued.

Figure 3-5. Primers used to detect the integration of S. scabiei toxicogenic region (TR) into the chromosome of S. diastatochromogenes.

CHAPTER 4 SITE-SPECIFIC ACCRETION OF THE THAXTOMIN A BIOSYNTHETIC CLUSTER WITH AN INTEGRATIVE CONJUGATIVE ELEMENT LEADS TO CIS-MOBILIZATION AND CONJUGATIVE TRANSFER

Introduction

Although Streptomyces are viewed predominantly as free-living terrestrial soil-inhabiting bacteria, some species are symbionts with fungi, insects, plants and animals (Seipke et al. 2012), and some strains have also been isolated from marine soils to explore novel secondary metabolites (Fiedler et al. 2005). A few Streptomyces species are plant pathogens and cause diseases of underground crop structures such as roots and tubers, and the most economically important disease caused by Streptomyces species is known as potato scab, characterized by raised, pitted, or superficial lesions formed on the tuber surface (Loria et al. 2006). The scab- causing species Streptomyces scabiei (syn. S. scabies) also infects the pericarp of the peanut fruit, causing peanut pod wart disease, however, this little known disease has only been described in Israel (Kritzman et al. 1996) and South Africa (De Klerk et al. 1997).

The best characterized scab-causing species are S. scabiei, Streptomyces acidiscabies, and Streptomyces turgidiscabies. Although plant pathogenic, Streptomyces ipomoeae only incites disease on plants in the family Convolvulaceae, including the economically important crop, sweet potato (Ipomoea batatas [L.] Lam.) (Clark and Watson 1983). Besides these four pathogenic species, seven other plant-pathogenic Streptomyces species have been described, namely Streptomyces europaeiscabiei, Streptomyces stelliscabiei, Streptomyces luridiscabiei,

Streptomyces puniciscabiei, Streptomyces niveiscabiei, Streptomyces reticuliscabiei, and S. caviscabies (Bouchek-Mechiche et al. 2000; Park et al. 2003a; Goyer et al. 1996a).

The primary pathogenicity determinant of scab-causing Streptomyces species is a family of dipeptide phytotoxins known as thaxtomins (Loria et al. 2006). A total of 11 family members

have been isolated; these thaxtomin analogues, are differentiated by the presence or absence of hydroxyl and/or N-methyl groups (Bignell et al. 2010a). In S. scabiei, S. acidiscabies, S. turgidiscabies, thaxtomin A (TxtA) is the predominant family member (Healy et al. 2000; King et al. 1989; Lawrence et al. 1990). However, the major thaxtomin analogue produced by S. ipomoeae is the nonhydroxylated analogue called thaxtomin C (King et al. 1994). S. europaeiscabiei, S. stelliscabiei, S. niveiscabiei, S. caviscabies also produce thaxtomins (Loria et al. 2006; Park et al. 2003a; Goyer et al. 1996a). Thaxtomin production by S. luridiscabiei and S. puniciscabiei, two species reported from Korea (Park et al. 2003a), has not been demonstrated

(Park et al. 2003b).

In S. turgidiscabies Car8, ThxA biosynthetic genes and other virulence genes, including nec1, a tomatinase gene (tomA), and a plant fasciation (fas) operon, reside on a mobile 674 Kb pathogenicity island (Kers et al. 2005; Huguet-Tapia et al. 2011). In contrast, virulence factors in

S. scabiei and S. acidiscabies reside on two remote chromosomal regions, namely the toxicogenic region (TR) and the colonization region (CR) (Lerat et al. 2009). In S. scabiei, TR is further separated into two sub-regions (TR1 and TR2) by an internal and two flanking attachment (att) sites (Chapleau et al. 2016). TR1 harbors the ThxA biosynthetic gene cluster while TR2 possesses the putative integrative and conjugative elements (ICEs), and CR contains nec1 and tomA (Chapleau et al. 2016) (Lerat et al. 2009). ICEs are self-transmissible elements that can encode the machinery for excision, conjugation, and integration, as well as regulatory systems to regulate their excision and conjugative transfer (Salyers et al. 1995; Osborn and

Boltner 2002; Burrus and Waldor 2004). Chapleau et al. demonstrated the excision of TR by three distinct forms, as TR1 alone, as TR2 alone, and as the entire TR region (Chapleau et al.

2016). In Chapter 3, we demonstrated the mobilization and site-specific integration of TR2 alone

or the entire TR, but not TR1 alone, from S. scabiei into the chromosome of the non-pathogenic

S. diastatochromogenes. These data suggest that TR1 is a cis-mobilizable element (CIME). A

CIME is a type of genomic island that lacks proteins involved in recombination and conjugation, but is flanked by att sites (Bellanger et al. 2014). We also observed that TR2, which is an ICE, only exists in S. scabiei and the novel species (S. sp. 96-12, GenBank accession #

GCA_001550315.1) but is absent in other pathogenic species, such as S. acidiscabies and S. europaeiscabiei (Chapter 3). These data suggest that S. scabiei is likely to be the source of the entire TR region and that its mobilization contributes to the emergence of novel pathogenic species.

We found that S. acidiscabies 84-104, isolated from a potato scab lesion in 1984

(Huguet-Tapia and Loria 2012), harbored a version of TR1 almost identical to that of S. scabiei, but lacked TR2 (Chapter 3). We hypothesized that S. acidiscabies, an emergent pathogenic species (Huguet-Tapia and Loria 2012), acquired the entire TR element but subsequently lost

TR2. However, the single available S. acidiscabies genome was insufficient to explore this hypothesis using a bioinformatics approach. Therefore, we sequenced genomes of four S. acidiscabies isolates from different locations. We were fortunate that although relatively few genome sequences were available for plant-pathogenic streptomycetes two years ago, genomes of both well-studied and less-well-studied pathogenic Streptomyces species have become available recently (Labeda 2016; Tomihama et al. 2016; Huguet-Tapia et al. 2016). These genomes allowed us to characterize the variation in the sequences of PAIs within and among pathogenic Streptomyces species, and investigate the origin and evolution of pathogenicity in the

Streptomyces genus. We also demonstrate that S. scabiei TR2 can transfer to a S. acidiscabies isolate harboring only TR1, and integrate into the att site at the 3’ end of the S. acidiscaibies

aviX1 gene, leading to simultaneous accretion of S. acidiscaibies TR1 and S. scabiei TR2.

Furthermore, the resulting composite island S. acidiscaibies TR1- S. scabiei TR2, can excise, and be transferred to a new host. These data demonstrated that TR1 is a CIME, and the site-specific accretion and cis-mobilization of TR1 together with the related ICE (TR2) have a key role in the dissemination of pathogenicity/virulence genes and emergence of novel pathogenic species in the

Streptomyces genus.

Materials and Methods

Bacterial Strains and Culture Conditions

MasterPure Gram-positive DNA purification kit (Epicentre) based on the manufacturer’s instructions.

Virulence Bioassays

To assess the virulence phenotype of different strains, an in vitro radish seedling bioassay was performed as described previously (Bignell et al. 2010b) except that “German Giant” seeds

Analysis of Thaxtomin A Production

shaking (250 rpm), thaxtomin A was purified from the culture supernatants and analyzed by

HPLC as described previously (Johnson et al. 2007). All experiments were repeated at least three times using different biological replicates of the Streptomyces strains, with three technical repeats per strain. The pellets of the culture samples were dried and weighted as a measure of bacterial growth. The data for S. scabiei 87-22 was normalized to 100 %. The data for all pathogenic isolates were square-root transformed to improve homogeneity of variance.

Transformed data were subjected to one-way analysis of variance (ANOVA) using SAS 9.4

(SAS Institute, Cary, NC), and means were compared based on Tukey's honestly significantly different (HSD) test at P < 0.05.

Genome Sequencing and Assembly

We sequenced the genomes of S. acidiscabies isolates using the Illumina-HiSeq 2500 paired-end technology. For each genome, the SPAdes version 3.5.0 assembly algorithm

(Bankevich et al. 2012) was used to perform genome assemblies. Gene-finding and annotation were achieved with Prokka version 1.12-beta (Seemann 2014).

Ortholog Retrieval and Single-copy Cluster Search

To build the rooted phylogenetic tree, Streptacidiphilus albus JL83 and Kitasatospora setae NBRC 14216 were selected as outgroups. Protein sequences from 24 Streptomyces genomes and two outgroups were clustered into ortholog groups using OrthoMCL v1.4 (Li et al.

2003). OrthoMCL uses BLASTP similarity for clustering, and was run using the default settings.

Ortholog groups with a single copy per genome were selected as potential phylogenetic markers.

Gene Alignment and Filtering Criteria

sequences, based on the original genomic sequence. Subsequently, alignment columns with a posterior probability < 0.6 were removed (Lefébure and Stanhope 2009). Alignments made after removing all ambiguous positions that were less than 50% of the original alignment length were excluded (Lefébure and Stanhope 2009).

Maximum Likelihood Phylogeny Estimation

The gene alignment was used for Maximum Likelihood (ML) inference on the RAxML webserver (https://www.phylo.org/portal2/) (Stamatakis et al. 2008) with the GTR+G substation model, including an estimation of invariable sites. For inferring tree robustness, 1000 bootstrap replicates were computed.

Average Nucleotide Identity and in silico DNA-DNA Hybridization Estimation

The Average Nucleotide Identity (ANI) analysis of whole-genome data was performed using the method proposed by Richter and Rosselló-Móra (Richter and Rosselló-Mora 2009).

Pair-wise ANI values were obtained from nucmer (Delcher et al. 2002) with the script calculate_ani.py

(https://github.com/widdowquinn/scripts/blob/master/bioinformatics/calculate_ani.py).

The isDDH was estimated using the genome-to-genome distance calculator (GGDC)

(Auch et al. 2010a; Meier-Kolthoff et al. 2013). The contig files were uploaded to the GGDC 2.0 webserver (http://ggdc.dsmz.de/distcalc2.php), where isDDH calculations were performed using

Formula 2 alone. We used the point estimates plus the 95% model-based confidence intervals as the isDDH estimates.

Selection of Transconjugant Strains

Streptomyces strains were first grown on ISP4 plates, and spores were harvested after 5-7 days. Conjugal mating was carried out by mixing spores of donor and recipient strains and co- culturing at 28 °C on SFM plates (Kieser et al. 2000). After 48 hours, 5 ml soft nutrient agar

containing antibiotics was spread onto the plates to select transconjugants. Final concentrations of antibiotics were 100 µg/ml hygromycin B, 50 µg/ml kanamycin, and/or 15 µg/ml thiostrepton.

Qualitative PCR assays were used to confirm the transfer and integration of TR elements (Figure

4-1, Table 4-1).

Results

Thaxtomin A Production of Four S. acidiscaibies Isolates

Two S. acidiscaibies isolates from the U.S. (85-06, FL01) and two isolates from Japan

(98-48, 98-49) were selected for genome sequencing (Table 4-2). Three of the four isolates caused the expected disease symptoms on radish seedlings (Figure 4-2A); noninoculated seedlings and seedlings inoculated with the nonpathogen S. diastatochromogenes grew normally.

In contrast, S. acidiscabies 98-48 had a phenotype intermediate to the healthy seedlings and those inoculated with the other three pathogenic strains.

We compared ThxA production by the four newly selected S. acidiscabies strains (85-06,

FL01, 98-48, and 98-49), three well-studied pathogenic isolates, namely S. acidiscabies 84-104,

S. scabiei 87-22 and S. turgidiscabies Car8 in OBB medium (Figure 4-2B), a complex medium used to induce ThxA biosynthesis (Wach et al. 2007). Four of the five S. acidiscabies isolates

(84-104, 85-06, 98-49, and FL01), S. turgidiscabies car8, and S. scabiei 87-22 produced thaxtomin as expected. However, S. acidiscabies 98-48 did not produce a detectable level of

TxtA in OBB. Even when OBB was supplemented with cellobiose, an excellent elicitor of ThxA biosynthesis (Johnson et al. 2007), ThxA production by S. acidiscabies 98-48 was still not detected (data not shown). We also found that all of the four ThxA-producing S. acidiscabies isolates (84-104, 85-06, 98-49, and FL01) produced a significant higher amount of ThxA than S. scabiei.

Phylogenomic and Genome Similarity Analysis of Pathogenic and Related Non-pathogenic Species

In Chapter 2, we demonstrated the power of core genome-based phylogenetic analysis to infer relationships among Streptomyces species, and the power of isDDH and ANI to identify novel species. However, Chapter 2 did not include several less-studied pathogenic Streptomyces species, namely S. luridiscabiei, S. puniciscabiei, and S. niveiscabiei, the genomes of which have been sequenced recently (Labeda 2016). To obtain a better understanding of evolutionary relationships of pathogenic Streptomyces species, we decided to update our phylogenomic and systematic analysis of pathogenic Streptomyces spp. A core genome-based phylogeny of using whole genome sequences of selected pathogenic isolates was constructed (Figure 4-3). We first identified 1192 single-copy ortholog groups that are shared among all 24 Streptomyces isolates as well as outgroups using OrthoMCL v1.4 (Li et al. 2003). The nucleotide sequence alignments of 1192 ortholog groups were filtered by removing poorly aligned sites, and 76 ortholog groups that were less than 50% of the original alignment length after filtering were excluded. The alignments of remaining 1116 genes were concatenated into a single alignment, and used to infer the phylogenetic relationships of the 24 Streptomyces isolates by maximum likelihood (ML).

The phylogenetic tree was distributed into three major clades: the first clade contains S. scabiei, S. stelliscabiei, S. bottropensis, S. europaeiscabiei, S. sp. 96-12, S. galbus KCCM

41354, S. ipomoeae and S. turgidiscabies; the second clade includes S. acidiscabies, S. niveiscabiei, and S. puniciscabiei; the third clade includes S. luridiscabiei alone.

ANI and isDDH are two of the most widely accepted in silico tools used to determine strain relationships, and 95-96% ANI and 70% isDDH values were applied as the threshold for species delineation (Goris et al. 2007; Auch et al. 2010b). In our analysis, we found that isolates in the same phylogenetic clade also showed high ANI and isDDH values (Figure 4-4). Although

S. niveiscabiei and S. acidiscabies are closely related in the phylogenetic tree, the low ANI

(~93%) and isDDH (52-54%) values unambiguously indicate that the current classification of these strains into two species is appropriate.

PAI Variation Exists within Pathogenic Streptomyces Species

Understanding the evolution of PAIs is of significance to yield novel insights into the emergence of pathogenic Streptomyces species in agricultural systems. With more genome sequences available recently, we can compare PAIs of different pathogenic Streptomyces isolates within and among species to explore the diversity of PAIs (Figure 4-5).

S. turgidiscabies Car8, isolated from Japan, contains a large mobile PAI of 674 Kb (Kers et al. 2005; Huguet-Tapia et al. 2011). Recently, another isolate from Japan, S. turgidiscabies

S58 (Tomihama et al. 2016), was selected for genome sequencing. Although the S. turgidiscabies S58 assembly is not complete, we identified several contigs of S. turgidiscabies

S58 that share synteny with the 674 Kb PAI of S. turgidiscabies Car8 . Two S. turgidiscabies isolates appear to contain the same 674 Kb PAI, but this requires confirmation by additional genome sequencing data to obtain the complete PAI of S. turgidiscabies S58.

In Chapter 3, we demonstrate that the TR2 region that contains the ICEs only exists in S. scabiei and S. sp. 96-12, but not in other pathogenic species. Since then, the genome sequence of

S. acidiscabies a10 was released (Tomihama et al. 2016); this strain also contains both TR1 and

TR2. However, S. acidiscabies 98-48 and the newly released sequence of S. acidiscabies NCPPB

4445 lack both TR1 and TR2. Therefore, within the S. acidiscabies species, three distinct compositions of the TR element exist: TR1 alone, the entire composite TR (TR1+TR2), and neither TR1 nor TR2. The genomes of S. niveiscabiei strains also vary in their PAI content. The

Korean isolate S. niveiscabiei NRRL B-24457 contains TR1 while S. niveiscabiei ST1015 and S.

niveiscabiei ST1020, two strains recently isolated from pitted potato scab lesions in Uruguay (M.

Julia-Pianzzola, personal communication), do not contain either TR1 or TR2.

Analysis of the genomes of two S. stelliscabiei isolates (NRRL B-24447 and P3825) suggest that both isolates contain TR1 and TR2. In these S. stelliscabiei assemblies, the TR1 regions were complete but the TR2 regions were separated into several contigs. For example, in

S. stelliscabiei P3825, TR1 and part of TR2 resides on a 111.4 Kb contig (Genbank accession #

JPPZ01000626.1) and the other part of TR2 resides on a 102.4 Kb contig (Genbank accession #

JPPZ01000624.1). Similar to S. scabei, the TR2 region of S. stelliscabiei P3825 was also inserted into the attachment (att) site at the 3’ end of avix1 gene. Chapleau et al. summarized the genes in TR2 that may be involved in conjugation and integration (Chapleau et al. 2016); S. stelliscabiei P3825 possesses all these genes except the gene encoding the ATPase (SCAB

_32961) (data not shown). Whether the lack of this gene affects the transfer of S. stelliscabiei TR requires further investigation. It is worth mentioning that the lantibiotic biosynthetic cluster was also missing in S. stelliscabiei TR2. Lantibiotics are small peptides that are active against other

Gram-positive bacteria and can therefore provide a fitness advantage to S. scabiei (Willey and van der Donk 2007; Nes and Tagg 1996). In addition, comparison of the TR regions in S. scabiei and S. stelliscabiei revealed significant genetic variation (data not shown). This indicates S. stelliscabiei has acquired TR for a long time.

S. acidiscabies, S. niveiscabiei and S. sp. 96-12 Acquired TxtA Biosynthetic Genes Through a Recent HGT Event

We demonstrated in Chapter 3 that although S. acidiscabies, S. sp. 96-12 and S. scabiei were phylogenetically distant pathogenic taxa, their TxtA clusters were almost identical, strongly suggesting a recent HGT of TxtA clusters. Since additional S. acidiscabies genomes and S. niveiscabiei genomes have become available, we have the opportunity to investigate whether we

can also detect the similar HGT event of ThxA clusters among these newly sequenced genomes.

To infer the evolution history of TxtA biosynthetic genes, we constructed a phylogenetic tree of pathogenic isolates using their txtA, txtB, txtC, txtD, txtE, txtH and txtR genes (Figure 4-6). In the

TxtA tree, five S. acidiscabies isolates (a10, 84-104, 85-06, 98-49, and FL01), two S. scabiei isolates (87.22 and NRRL B-16523), S. niveiscabiei NRRL B-2445 and S. sp. 96-12 were closely grouped, indicating the high similarity of their ThxA clusters. In contrast, S. acidiscabies, S. niveiscabiei, S. sp. 96-12 and S. scabiei were distantly related in the core-genome tree (Figure 4-

7). The high identity of the ThxA clusters in these four phylogenetically distant species (S. acidiscabies, S. niveiscabiei, S. sp. 96-12 and S. scabiei) support the notion that a recent PAI transfer among them.

Site-specific Accretion Between TR1 and TR2

We have demonstrated that TR2 presented a high integration specificity as it was always integrated into the att site at the 3’ end of aviX1 (Figure 3-4). Since TR1 is not mobile by itself, it seems to be a CIME. Recently, two Streptococcus CIMEs have been shown to become mobile after the site-specific accretion of the CIME with a related ICE (Bellanger et al. 2011; Puymege et al. 2013). It is possible that the cis-mobilization of Streptomyces TR1-TR2 is another example of site-specific accretion. To detected the site-specific accretion of TR1 and TR2, we carried the mating experiments using two strains from previous studies: (1) S. acidiscabies 84. 303. 1

(kanamycin resistance [kanR]) (Healy et al. 2000), containing only TR1 with a deletion mutation in txtA and txtB genes (hereinafter referred to as S. acidiscabies ∆txtAB); (2) and S. diastatochromogenes transconjugants (hygromycin B resistance [hygR]), containing only TR2 of

S. scabiei with a deletion mutation in lantibiotic biosynthesis genes lanA (SCAB_32021) and lanB (SCAB_32031) (hereinafter referred to as S. diastatochromogenes ∆lanAB) (This strain is from Chapter 3). S. diastatochromogenes ∆lanAB (hygR) served as the TR2 donor, while S.

acidiscabies ∆txtAB (kanR) acted as the recipient strain (Figures 4-8A, 4-9). Following selection with kanamycin together with hygromycin B, we obtained S. acidiscabies transconjugants containing both TR1 and TR2 (hereinafter referred to as S. acidiscabies ∆txtAB ∆lanAB). Using

PCR primers a+b, c+d, e+f, and g+h (Table 4-1), we found that TR2 was always inserted into the attachment site (attR) located at the 3’ end of the S. acidiscabies aviX1 gene rather than the attachment site (attL) on the other side of TR1 in all of the 15 tested S. acidiscabies ∆txtAB

∆lanAB transconguants (Figure 4-9), and the orientation and structure of their TR1 and TR2 is same with that of TR1 and TR2 in S. scabiei.

Conjugative Mobilization of S. acidiscabies TR1 in cis by S. scabiei TR2

The conjugative transfer of the composite element of S. acidiscabiei TR1-S. scabiei TR2 was then investigated by our second mating experiment (Figure 4-8B). S. acidiscabies ∆txtAB

∆lanAB served as the TR donor, while the S. diastatochromogenes transconjugant (thiostrepton resistance [thioR]) from a previous study (Kers et al. 2005), containing the 674 Kb PAI of S. turgidiscabies Car8 with a deletion mutation in nec1 (hereinafter referred to as S. diastatochromogenes ∆nec1) acted as the recipient strain. Following selection with hygromycin

B and thiostrepton, interestingly, we obtained three types of transconjugants: 1) S. diastatochromogenes strains containing TR2 and the S. turgidiscabies PAI, which are the result of mobilization of TR2 from S. acidiscabies to S. diastatochromogenes; 2) S. diastatochromogenes strains containing TR1, TR2 and S. turgidiscabies PAI, which are the result of mobilization of TR1 and TR2 from S. acidiscabies to S. diastatochromogenes; and 3) S. acidiscabies strains containing TR1, TR2 and S. turgidiscabies PAI, which are the result of mobilization of S. turgidiscabies PAI from S. diastatochromogenes to S. acidiscabies. These mating experiments showed that even though some pathogenic species lack TR2, introduction of external TR2 can lead to the site-specific accretion and cis-mobilization of TR1 and TR2. In

addition, these three types of transconjugants were also the evidence of the exchange of different

PAIs between species.

TR1 is fixed in pathogenic strains lacking TR2

We have demonstrated that S. scabiei TR1 is not mobile alone and its transfer depends on

TR2, however, comparative genomes showed that most pathogenic isolates only contained TR1 but not TR2 (Figure 4-5). It is possible that these pathogenic strains containing only TR1 use mechanisms rather than TR2 to disseminate their TR1 to non-pathogenic species. To test whether

TR1 of those pathogenic species lacking TR2 can be mobilized by other mechanisms, we carried the third mating experiment by mating S. acidiscabies ∆txtAB with S. diastatochromogenes

∆nec1 (Figure 4-8C). Following selection with thiostrepton and kanamycin, we only obtained S. acidiscabies transconjugants containing TR1 and S. turgidiscabies PAI, which are the result of mobilization of PAI of S. turgidiscabies from S. diastatochromogenes to S. acidiscabies. Failure to detect mobilization of S. acidiscabies TR1 to S. diastatochromogenes suggests without the

TR2 region, TR1 cannot be transferred and will get fixed.

Discussion

The present study would appear to be the most comprehensive investigation of PAI diversity and evolution of pathogenicity among the genus Streptomyces, including 24 pathogenic isolates from 10 species. The results, in combination with the findings of Chapter 3, demonstrated that the entire TR region could be a composite element probably stemmed from the accretion of a putative ICE (TR2), and a CIME (TR1). Their site-specific accretion leads to the cis-mobilization and conjugative transfer of ThxA biosynthetic genes and can be responsible for the emergence of novel Streptomyces pathogenic species in the agricultural system. Since the

Streptomyces genus is ubiquitous in soils, our study brings up the risk that new pathogenic species will continuously appear.

Previously, the cis-mobilization of two CIMEs by related ICEs has been reported: the transfer of CIMEL3catR3 by ICESt3 in Streptococcus thermophiles (Bellanger et al. 2011), and the transfer of CIME_Nem_tRNALys by ICE_515_tRNALys in Streptococcus agalactiae (Puymege et al. 2013). Our results demonstrated the cis-mobilization of TR1 by TR2 qualifies as another example of such mechanism. Additionally, CIMEs of S. thermophilus have been shown to derive from ICEs by deletion of genes involved in conjugation and integration (Pavlovic et al. 2004). It is possible that in Streptomyces TR1, the CIME harboring the TxtA biosynthetic cluster, has evolved by similar mechanism: an ICE integrated into the 3’ end of aviX1 would have lost its integration and conjugation genes while retaining two attachment sites (attL and attR), resulting in a CIME (TR1) (Figure 4-10, step 1); a related ICE (TR2) was subsequently acquired by conjugation (Figure 4-10, step 2) and site-specifically integrated between its attI site and the attR site of TR1 (Figure 4-10, step 3), resulting in the structure attL-CIME (TR1)-attPi-ICE2 (TR2)- attR-aviX1; the whole structure could be excised by recombination among three attachment sites in three different forms (Figure 4-10, step 4): as TR1, as TR2, or as the entire TR element; then,

TR2 or the entire TR can be integrated into the attachment site at the 3’end of the aviX1 gene of new hosts (Figure 4-10, step 5). Acquisition of the entire TR element can confers the pathogenic phenotype on new hosts, leading to the emergence of novel pathogenic species (Chapter 3).

S. luridiscabiei, S. puniciscabiei, and S. niveiscabiei were reported to cause potato common scab in Korea (Park et al. 2003a). Among these tree species, only S. niveiscabiei produced thaxtomins (Park et al. 2003a). Until recently, these three species have never been discovered in other regions of the world; interestingly, S. niveiscabiei strains were isolated from potato pitted scab lesions in Uruguay in 2010 (M. Julia-Pianzzola, personal communication).

However, these Uruguay S. niveiscabiei isolates do not produce thaxotmin but are pathogenic on

potato tuber disk assay (M. Julia-Pianzzola, personal communication). Comparative genomic analysis showed the isDDH values between Korean and Uruguay S. niveiscabiei isolates were only 87% (Figure 4-4), suggesting isolates from Uruguay and Korea possibly not originated from a single introduction. Indeed, S. niveiscabiei has been reported only in Korea and Uruguay does not mean that other potato growing regions are free of this species. Our core genome-based phylogenetic analysis revealed the close relationship between S. niveiscabiei and S. acidiscabies.

Owing to the poor characterization of S. niveiscabiei, S. niveiscabiei isolates will be easily misidentified as S. acidiscabies. S. niveiscabiei isolates from Uruguay lack the TxtA cluster, while isolates from Korea harbor the TxtA cluster, which is almost identical to the one in S. scabiei 87-22, despite significant differences in their whole genome sequences. These results suggest that similar to S. acidiscabies, S. niveiscabiei is also an emergent pathogenic species, and the Korean isolate acquired the TxtA cluster via a recent HGT event.

Recent international efforts have sequenced more than 30 genomes of pathogenic

Streptomyces isolates belonging to different species. The comparative genomic analysis presented here allows us to characterize the pathogenicity/virulence factors in different pathogenic isolates. A striking finding is that PAI variation exists not only between different species but also among strains from the same species. The seven S. acidiscabies isolates, for which genome sequences are available, carried three different PAI structures: TR1 alone (84-

104, 85-06, 98-49 and FL01), the entire TR element (a10), and no TR (98-48 and NCPPB 4445).

Interestingly, the PAI structures of the three Japanese S. acidiscabies isolates (98-48, 98-49, and a10) differed from each other. Similarly, three S. niveiscabiei carried two different PAI structures. The Korean isolate NRRL B-24457 carried the TR1 alone, but two isolates from

Uruguay (ST1015 and ST1020) do not contain the TR element. If we sequence additional S.

niveiscabiei genomes, we are likely to detect some isolates that carry the third TR version (the entire TR element). Together, these data show that the TR element is dynamic and that different forms of TR coexist in newly emerged pathogenic species, such as S. acidiscabies and S. niveiscabiei.

Excision of an ICE could cause the loss of this ICE from the host; if ICE excision happens before chromosome replication or cell division, this element probably will not be inherited to progeny (Wozniak and Waldor 2010). ICEs was thought to lack the autonomous replication ability; however, recent studies have demonstrated that many ICEs undergo autonomous replication and they share the conserved mechanisms (Lee et al. 2010; Sitkiewicz et al. 2011). Chapleau et al. reported that a TR2 locus (SCAB_32241) contained a replication- partition motif, indicating that TR may also possess the autonomous replication ability, which will be critical to the stability and propagation of TR (Chapleau et al. 2016). However, further functional study is required to confirm the autonomous replication ability of TR.

Some S. turgidiscabies isolates from the potato fields of Finland infected by both S. scabiei and S. turgidiscabies, were reported to contain a PAI with the similar structure to the PAI of S. scabiei 87-22 rather than to the PAI of S. turgidiscabies Car8, indicating the exchange of

PAI regions between two species (Aittamaa et al. 2010). Here, we detected the exchange of the

Car8-type PAI and TR-type PAI between S. acidiscabies and S. diastatochromogenes transconjugants. Such genetic exchange is critical to the evolution and pathogenicity of bacterial pathogens; the acquisition of new genes can enable pathogens to inhabit a new niche rapidly

(Holden et al. 2009).

The mobility of the 674 Kb PAI of S. turgidiscabies Car8 has been demonstrated (Kers et al. 2005), bringing up the possibility of the appearance of this large island in new pathogenic

species. However, all newly identified pathogenic species contain the TR-type PAI or its derivative but not the S. turgidiscabies Car8-type PAI. It may be that the 674 Kb Car8-type PAI cannot be stably transferred or maintained. This possibility can be supported by an earlier report that considerable variability existed in the composition of PAI genes of S. turgidiscabies isolates in Finland (Aittamaa et al. 2010): some isolates lack the fas operon, while other isolates lack all known virulence genes. However, in the present study, two Japanese S. turgidiscabies isolates

(Car8 and T45) contained identical PAIs. These two isolates probably originated from a single introduction. Further genome sequencing projects should include additional S. turgidiscabies strains to investigate the independent evolution of mosaic regions of the Car8-type PAI.

Until recently, all Streptomyces species causing the potato scab disease were thought to produce thaxtomins. Therefore, PCR primers that can amplify ThxA genes have been proposed for the detection of common scab pathogens (Flores González et al. 2008; Lerat et al. 2009).

However, a growing body of evidence suggests that other secondary metabolites may be primary virulence factors. A Streptomyces isolate was reported to cause pitted scab lesions by producing an 18-membered macrolide named borrelidin (Cao et al. 2012). Recently, Fyans et al. isolated a

Streptomyces strain from Newfoundland, Canada that has a severe pathogenic phenotype on several plant hosts; symptoms were associated with an unidentified phytotoxin (Fyans et al.

2016). S. reticuliscabiei, S. luridiscabiei, and S. puniciscabiei do not produce thaxtomins, but they are pathogenic species (Aittamaa et al. 2010; Park et al. 2003a). S. reticuliscabiei induces netted scab lesions on potato tubers (Aittamaa et al. 2010), while the latter two species cause the potato scab disease (Park et al. 2003a). Furthermore, two Uruguay S. niveiscabiei isolates

(ST1015 and ST1020) that lack thaxtomin production can cause deep pitted lesions on potato tubers (M. Julia-Pianzzola, personal communication). Taken together, these results indicate that

although ThxA is still the primary pathogenicity factor in most pathogenic Streptomyces species, plant pathogenicity is not exclusive to thaxtomin producing strains.

Table 4-1. List of primers used in this study Primer Sequence 5’- 3’ Use

a GAGAACGGCGAAGTCCCAG Primers flanking the attL site in S. acidiscabies ∆lanAB b GGTGATCTGCGTGGAGGACTG (Figures 4-1, 4-9) c GTACGTCGTCTCAGCCATGACC Primers flanking the attPi site in S. acidiscabies d GCCACAAGTCGATCAAGACCACG transconjugants TR1 + TR2 (Figures 4-1, 4-9) e GTAGCGAAGGCGAGAGTCTCACTG Primers within the integrase gene (SCAB_31871) f GAGCCGACGAACAAGTACTACCCG g CGAAGATCGAGAACGTCAGGAAGG Primers flanking the attR site in S. diastatochromogenes h GACCGACGAGGACTTCAAGAACGA with TR1 or TR1 + TR2 (Figures 4-1, 4-9) i TCCACCTCCTGACCACCAAG Detect the integration of TR into the attB site at the 3’ j AAGATCCCCGAACCGACCT end of S. diastatochromogenes aviX1 (Figures 4-1, 4-9) k GAAGAAGAAGCCCTCGAGTA Primers within the integrase (STRTUCAR8_02003) in l GGAAGTTCGCGTTGATGAG the PAI of S. turgidiscabies Car8 m GAGGAGGCGATCTTGCGTGAACT Primers within the putative methyltransferases gene n CGTAGTGCTTGTCGATGCTCCG (stPAI016) of the a plant fasciation (fas) operon in the PAI of S. turgidiscabies Car8 o TACTGGCCCTGCTCGATGTA Primers flanking the internal attachment site in the PAI p TGCAATCAGGCGCTAGGAAA of S. turgidiscabies Car8

Table 4-2. Genome summary statistics of plant pathogenic Streptomyces spp. Species Strain Genome Contig Contig Genome Geographic location GenBank assembly Source or reference name status number N50 (Kb) size (Mb) accession # S. acidiscabies 84-104 Draft 244 113.4 11.0 USA: upstate New GCA_000242715.2 (Huguet-Tapia and York Loria 2012) S. acidiscabies 85-06 Draft 1073 20.0 11.0 USA: West Virginia LYDU00000000 This study S. acidiscabies 98-48 Draft 983 20.3 10.6 Japan: Saga LYDS00000000 This study S. acidiscabies 98-49 Draft 1211 16.5 10.2 Japan: Saga LYDT00000000 This study S. acidiscabies FL01 Draft 993 19.3 10.6 USA: Florida LYDV00000000 This study S. acidiscabies a10 Draft 279 82.2 10.7 Japan: Kagoshima, GCA_001485105.1 (Tomihama et al. Nagashima 2016) S. acidiscabies NCPPB 4445 Draft 263 88.0 10.0 Unknown GCA_001189015.1 University of Exeter S. europaeiscabiei NCPPB 4064 Draft 206 81.4 8.6 Netherlands GCA_001189025.1 University of Exeter S. europaeiscabiei NCPPB 4086 Draft 295 97.2 10.5 Canada: Ontario GCA_000738695.1 University of Exeter S. europaeiscabiei 89-04 Draft 1080 17.7 10.5 USA: Alaska GCA_001550325.1 Chapter 3 S. europaeiscabiei 96-14 Draft 931 21.2 10.5 Australia GCA_001550375.1 Chapter 3 S. europaeiscabiei NRRL B- Draft 1056 16.4 10.0 France: Le Rhue GCA_000988945.1 (Labeda 2016) 24443 S. stelliscabiei NRRL B- Draft 991 19.0 10.0 France GCA_001008135.1 USDA/ARS/NCAU 24447 R S. stelliscabiei P3825 Draft 705 32.8 10.4 Unknown GCA_001189035.1 University of Exeter S. turgidiscabies Car8 Draft 692 27.4 10.8 Japan GCA_000331005.1 (Huguet-Tapia et al. 2011) S. turgidiscabies T45 Draft 101 264.4 10.6 Japan: Kagoshima, GCA_001485145.1 (Tomihama et al. Kushira 2016) S. ipomoeae 91-03 Draft 687 26.4 10.4 USA GCA_000317595.1 (Huguet-Tapia et al. 2016) S. scabiei 87-22 Complete 10.1 USA: Wisconsin GCA_000091305.1 (Bignell et al. 2010b) S. scabiei NRRL B- Draft 1091 15.1 9.8 USA: central New GCA_001005405.1 (Labeda 2016) 16523 York S. scabiei S58 Draft 149 193.2 10.0 Japan: Kagoshima, GCA_001485125.1 (Tomihama et al. Nagashima 2016) S. luridiscabiei NRRL B- Draft 821 18.0 7.9 South Korea: Jeju GCA_001418625.1 (Labeda 2016) 24455 S. puniciscabiei NRRL B- Draft 688 25.4 8.8 South Korea: Jeju GCA_001419685.1 (Labeda 2016) 24456 S. niveiscabiei NRRL B- Draft 1568 10.1 9.5 South Korea: Jeju GCA_001419795.1 (Labeda 2016) 24457

Table 4-2. Continued Species Strain Genome Contig Contig Genome Geographic location GenBank assembly Source or reference name status number N50 (Kb) size (Mb) accession # S. niveiscabiei ST1015 Draft 1126 14.6 9.8 Uruguay LXWC00000000 Universidad de la República S. niveiscabiei ST1020 Draft 1010 17.8 9.9 Uruguay LXWD00000000 Universidad de la República S. sp. 96-12 Draft 1875 8.7 10.1 Egypt GCA_001550315.1 Chapter 3

Figure 4-1. Primers used to detect the integration of the toxicogenic region (TR) into the chromosome of S. acidiscabies and S. diastatochromogenes.

Figure 4-2. Virulence assays of pathogenic Streptomyces isolates. A) Surface-sterilized radish seedlings (six per treatment) were inoculated with mycelia from tryptic soy broth-grown cultures of Streptomyces strains. Seedlings treated with S. scabiei 87-22, the non-pathogenic species S. diastatochromogenes or with water were used as controls. Plants were incubated at room temperature under a 12-h photoperiod for 4 days. Representative plants for each treatment are shown. B). Production of thaxtomin A by plant pathogenic Streptomyces isolates. Isolates were grown in triplicate in OBB medium for 6 days at 28°C. The average thaxtomin production of S. scabiei 87-22 that is normalized using dried cell weights is set to 100%. The average % production for each strain relative to S. scabiei 87-22 and normalized using dried cell weights is shown. The errors bars represent the standard deviation from the mean. Letters represent results of a one-way ANOVA with Tukey’s HSD test; bars not sharing letters are significantly different at P < 0.05.

Figure 4-3. Phylogeny of pathogenic isolates and their related non-pathogenic Streptomyces species based on the concatenation of 1116 single-copy gene clusters with the maximum likelihood algorithm RAxML.

Figure 4-4. ANI and isDDH values of pathogenic Streptomyces isolates. The lower triangle displays ANI values, and the upper triangle displays isDDH values. Displayed isDDH values are point estimates plus the 95% model-based confidence intervals. Boxes with ANI ≥ 99% or isDDH ≥ 70% are colored red, and boxes with ANI ≥ 96% and < 99% or isDDH ≥ 70% and < 79% are colored orange.

Figure 4-5. Schematic genetic organization of PAIs from different pathogenic Streptomyces isolates. TR: toxicogenic region; CR: colonization region. Arrowed boxes represent the location and orientation of the open reading frames (ORFs) of ThxA biosynthetic genes (red), integrase and recombination directionality factor (RDF) (green), lantibiotic biosynthetic genes (orange), conjugative genes (blue), aviX1 (brown), bacA (light blue), nec1 (black), tomA (purple), sugar catabolism clusters (gray), and fas operon (yellow). The sequences of attachment sites at the junctions for TR1/TR2/TR are also shown.

Figure 4-6. Phylogenetic reconstruction of scab-causing Streptomyces strains based on the thaxtomin A biosynthetic cluster (txtA, txtB, txtC, txtD, txtE, txtH, and txtR). 1000 bootstrap replicates were computed. Numbers above branches are bootstrap percentages.

Figure 4-7. Comparison of evolutionary relationships derived from the core-genome tree (Figure 4-3) and the TxtA tree (Figure 4-6).

Figure 4-8. Schematic representation of the mating experiments for investigation of the site- specific accretion and cis-mobilization of TR1 and TR2. A) the mating experiment between S. acidiscabies ∆txtAB and S. diastatochromogenes ∆lanAB for studying the of the site-specific accretion between TR1 and TR2. B) the mating experiment between S. acidiscabies ∆txtAB ∆lanAB and S. diastatochromogenes ∆nec1 for studying the cis-mobilization of TR1 and TR2. C) the mating experiment between diastatochromogenes ∆nec1 and S. acidiscabies ∆txtAB to confirm that for strains lacking TR2, their TR1 cannot be transferred. PAIst, PAI in S. turgidiscabies car8.

Figure 4-9. Cartoon illustrating the site-specific accretion and cis-mobilization of TR1 and TR2. S. scabiei TR2 was site-specifically into the attachment site at the 3’end of the S. acidiscabies aviX1 gene. The composite TR element can excise in three distinct forms: TR1, TR2, and the entire TR. TR2 alone or the whole TR can integrate into the S. diastatochromogenes chromosome. The cis-moblization of TR1 and TR2 can lead to the emergence of novel pathogenic species.

Figure 4-10. Schematic representation of a hypothesis for the evolution of TR1 and TR2 by deletion, site-specific accretion, and cis-mobilization. An ICE integrated into the 3’ end of aviX1 would have lost its integration and conjugation genes while retaining two attachment (att) sites (attL and attR), resulting in a CIME (TR1) (step 1). A related ICE (TR2) was subsequently acquired by conjugation (step 2) and site- specifically integrated between its attI site and the attR site of TR1 (step 3), resulting in the structure attL-CIME (TR1)-attPi-ICE2 (TR2)-attR-aviX1. The whole structure could be excised by recombination among three attachment sites in three different forms (step 4): as TR1, as TR2, or as the entire TR element. Then, TR2 or the entire TR can be integrated into the attachment site at the 3’end of the aviX1 gene of new hosts (step 5).

CHAPTER 5 A RE-EVALUATION OF THE TAXONOMY OF PHYTOPATHOGENIC GENERA DICKEYA AND PECTOBACTERIUM USING WHOLE-GENOME SEQUENCING DATA

The family Enterobacteriaceae includes some devastating human and animal pathogens, such as Escherichia coli and Salmonella enterica (Toth et al. 2006)*. This family also contains some important plant pathogens that are classified in the genera Dickeya, Pectobacterium,

Brenneria, Erwinia, and Pantoea (Samson et al. 2005). Five pathogens (Erwinia amylovora,

Pectobacterium carotovorum, Pectobacterium atrosepticum, Dickeya dadantii, and Dickeya solani) from above genera have been classified as the most important bacterial pathogens limiting crop yield and quality (Mansfield et al. 2012). The taxonomy of these five genera has been subjected to extensive taxonomic research and much revision over the years (Hauben et al.

2005; Samson et al. 2005; Denman et al. 2012).

Pectobacterium and Dickeya are formerly classified as rot pathogens of the Erwinia genus (Ma et al. 2007). The genus Pectobacterium was first established in 1945 (Waldee 1942), and now contains strains that attack a wide range of plants (Ma et al. 2007), causing soft rot, stem rot, wilt, and blackleg (Perombelon and Kelman 1980; Pérombelon 2002; del Pilar

Marquez-Villavicencio et al. 2011). However, Pectobacterium strains with a narrow-host-range were also isolated, namely P. atrosepticum and P. betavasculorum. P. atrosepticum is limited to potato (Solanum tuberosum L.), and P. betavasculorum is almost exclusively on sugar beet

(Gardan et al. 2003). P. atrosepticum and P. betavasculorum were formerly classified as the subspecies of P. carotovorum; however, these two species together with Pectobacterium wasabiae were elevated to the species level subsequently (Gardan et al. 2003). Currently, P.

*Reprinted with permission from Zhang, Y., Fan, Q., and Loria, R. 2016. Syst. Appl. Microbiol. 4: 252-259. Copyright © 2016, Elsevier GmbH. doi: 10.1016/j.syapm.2016.04.001. Specific contributions by YZ include design, and conduction of the comparative genomic and phylogenetic analysis, writing and submission of the manuscript.

100

carotovorum includes two subspecies with validly published names: carotovorum and odoriferum. P. carotovorum subsp. carotovorum causes soft rot in a wide spectrum of plant species (Barras and Van Gijsegem 1994), while P. carotovorum subsp. odoriferum causes slimy rot of witloof chicory (Gallois et al. 1992). Besides these two subspecies, a third subspecies, P. carotovorum subsp. brasiliense, was also proposed (Duarte et al. 2004) , albeit without valid acceptance yet. P. carotovorum subsp. brasiliense was first described as the causal agent of blackleg disease on potatoes in Brazil (Duarte et al. 2004), and was later reported to cause soft rot in Capsicum annum L., Ornithogalum spp., and Daucus carota subsp. sativus (Nabhan et al.

2012). Recently, Pectobacterium carotovorum subsp. actinidiae and Pectobacterium aroidearum were also proposed (Koh et al. 2012; Nabhan et al. 2013). P. carotovorum subsp. actinidiae causes canker-like symptoms in yellow kiwifruit (Actinidia chinensis) (Koh et al. 2012), and P. aroidearum contains soft rot pathogens with preference for monocotyledonous plants (Nabhan et al. 2013).

The genus Dickeya (originally classified as Erwinia chrysanthemi) was first established in 2005, as it was first reported as the cause of bacterial blight of Chrysanthemums (Burkholder et al. 1953). Later studies, however, revealed that it had a wide-host-range, including agronomically important crops such as potato and maize (Samson et al. 2005). The founding members of Dickeya contain six species, namely, Dickeya zeae, Dickeya dadantii, Dickeya chrysanthemi, Dickeya dieffenbachiae, Dickeya dianthicola, and Dickeya paradisiaca (Samson et al. 2005). Subsequently, D. dieffenbachiae was reclassified as a subspecies of D. dadantii

(Brady et al. 2012). Recently, some pectinolytic plant pathogenic isolates from European potatoes were classified as a novel species, Dickeya solani (van der Wolf et al. 2014). Some

101

newly isolated Dickeya strains were found not to fall within these current species, and may therefore constitute novel Dickeya species (Sławiak et al. 2009).

The genus Brenneria was first proposed in 1998 (Hauben et al. 1998a), and comprises five species with validly published names to date (Li et al. 2015b). Almost all species of this genus are plant pathogens; for example, Brenneria salicis is the causal agent of watermark disease in willows (Salix spp.) (Hauben et al. 1998b).

A variety of methods have been utilized to infer the taxonomic relationships of the genera

Dickeya, Pectobacterium, Brenneria, Erwinia, and Pantoea, including morphological and biological methods, serological methods, and molecular methods (Czajkowski et al. 2015). 16S ribosomal RNA (rRNA) (Hauben et al. 1998a), multilocus sequence analysis (MLSA) (Nabhan et al. 2012), and DDH (Gardan et al. 2003) are three of the most widely used molecular methods.

16S rRNA sequencing and MLSA are much easier to perform than DDH, but DDH is still the most widely accepted approach by the scientific community for microbial species discrimination.

Bacterial strains that show an approximate ≥ 70% DDH value will be considered to belong to the same species. However, DDH has been criticized for certain technique-inherent limitations: (i) the DDH experiment is cumbersome and currently only carried out in relatively few specialized molecular laboratories; (ii) DDH is error prone and distinct DDH methods may generate inconsistent results; (iii) DDH is unable to generate cumulative databases as the gene sequences

(Rosselló-Mora 2006) (Auch et al. 2010b).

Now, 20 years after the first bacterial genome was sequenced (Fleischmann et al. 1995), novel microbial genomic data is being generated on a daily basis. The increasing availability of genomic information has been triggering the development of genomic-derived values to replace wet-lab DDH (Auch et al. 2010b; Goris et al. 2007; Richter and Rosselló-Mora 2009; Auch et al.

102

2010a). The average nucleotide identity (ANI), and the genome-to-genome distance calculator, or in silico DDH (isDDH), are two of the most widely accepted bioinformatics tools that calculate whole-genome sequence similarities through comparing genomic data. A recent study in Aeromonas showed ANI values of ≥ 96% and isDDH values of ≥ 70% consistently grouped genomes originating from strains of the same species together (Colston et al. 2014). Our previous study also successfully applied whole-genome sequencing data to address the complex taxonomic status of the genera Erwinia and Pantoea (Zhang and Qiu 2015), e.g, the genome similarity (isDDH: 94.2%) between Pantoea stewartii subsp. stewartii and Pantoea stewartii subsp. indologenes is much greater than the 79% threshold for subspecies delineation (Meier-

Kolthoff et al. 2013), indicating it is not appropriate to classify these strains into two subspecies.

Whole-genome sequencing data has the potential to re-evaluate current prokaryotic species definition and establish a unified prokaryotic species definition frame for the taxonomically challenging genera Dickeya and Pectobacterium. Here, we evaluated ANI and isDDH in combination with phylogenetic studies of these genera. Our study included 83 publicly available genomes of the genera Dickeya and Pectobacterium as well as two Brenneria genomes.

This work provides new insights into the evolutionary relationships of different species and subspecies of these plant-pathogenic genera, reveals that a number of genomes deposited in

GenBank are misnamed and some strains represent novel species, and presents evidence that four subspecies of P. carotovorum should be elevated to the species level.

Materials and Methods

Genome Collection

All complete and draft genome sequences of Dickeya, Pectobacterium, and Brenneria were retrieved from the GenBank database. A total of two Brenneria genomes, 32 Dickeya genomes, and 51 Pectobacterium genomes were included in this study (Table 5-1). This study

103

also retrieved the genomes of Erwinia amylovora ATCC 49946 (GenBank assembly accession:

GCA_000027205.1) and Pantoea ananatis LMG 5342 (GenBank assembly accession:

GCA_000283875.1), which were used as the outgroups.

Ortholog Retrieval and Single-copy Ortholog Search

Orthologs were identified in the proteomes of all strains using the OrthoMCL v.1.4 (Li et al. 2003). OrthoMCL was run with a BLAST E-value cut-off of 1e-5, a minimum aligned sequence length coverage of 50% of the query sequence, and an inflation index of 1.5.

Orthologous groups present as single-copy in all genomes including the outgroups E. amylovora

ATCC 49946 and P. ananatis LMG 5342 were chosen for rooted phylogenetic tree reconstruction.

Gene Alignment and Filtering Criteria

To minimize the use of alignments of low quality, several filtering criteria were used.

Alignments for each of the one-to-one rooted core genes were generated at the amino acid level using Probalign (v1.1) (Roshan and Livesay 2006). These sequences were then back-translated to nucleotide sequences based on the genomic DNA sequences, and alignment columns with a posterior probability < 0.6 were removed (Suzuki et al. 2012; Lefébure and Stanhope 2009).

Alignments that were less than 50% of the original alignment length after removing all ambiguous positions were excluded (Lefébure and Stanhope 2009). Nucleotide sequences of filtered alignments were then translated to corresponding amino acid sequences to obtain the filtered alignments of amino acids.

Phylogenomic Analysis

Gene alignments of both nucleic acid and amino acid sequences were concatenated and submitted to the RAxML webserver (https://www.phylo.org/portal2/) for Maximum Likelihood

(ML) inference (Stamatakis et al. 2008). The nucleic acid phylogeny was inferred using the GTR

104

+ G substitution model, including an estimation of invariable sites. The amino acid phylogeny was inferred using the GAMMA AUTO model, which enables RAxML to automatically determine the best-scoring protein substitution model (Stamatakis 2014). For both phylogenies,

1000 bootstrap replicates were computed to estimate reliability.

Average Nucleotide Identity (ANI) and in silico DNA-DNA Hybridization (isDDH) Analysis

Pair-wise ANI values were obtained from nucmer (Delcher et al. 2002) with the script calculate_ani.py

(https://github.com/widdowquinn/scripts/blob/master/bioinformatics/calculate_ani.py), while isDDH was estimated using the Genome-to-Genome Distance Calculator (GGDC) (Auch et al.

2010b; Meier-Kolthoff et al. 2013). The genomic DNA sequences were uploaded to the GGDC

2.0 Web server (http://ggdc.dsmz.de/distcalc2.php). We used the point estimates plus the 95% model-based confidence intervals from Formula 2 alone as the isDDH estimates (Colston et al.

2014).

Results

In brief, our pipeline used for re-evaluating the taxonomy of the genera Dickeya and

Pectobacterium based on whole genome sequences proceeds as follows (Figure 5-1, see more details in the ‘Materials and methods’ section): (i) 87 genomes were analyzed by OrthoMCL v1.4 to determine single-copy orthologous groups; (ii) each single-copy orthologous group was aligned at the amino acid level by Probalign and ambiguously aligned regions were removed;

(iii) amino acid sequences of filtered alignments were concatenated and submitted to RAxML to reconstruct the whole-genome-based amino acid phylogeny; (iv) nucleic acid sequences of filtered alignments were concatenated and submitted to RAxML to reconstruct the whole- genome-based nucleic acid phylogeny; (v) genome similarities were assessed using ANI; (vi)

105

genome similarities were assessed using isDDH; (vii) taxonomy was evaluated based on the combination of phylogeny, ANI and isDDH.

Phylogenomic Analysis of Dickeya and Pectobacterium

Using OrthoMCL v1.4, we identified 898 single-copy orthologous groups. After alignments and removing poorly aligned sites, 895 filtered alignments passed the alignment quality criteria as described in the Material and methods. Nucleic acid sequences and amino acid sequences of 895 alignments were concatenated, and submitted to the RAxML webserver to build the phylogeny using the maximum likelihood algorithm. The whole-genome-derived nucleic acid phylogeny (Figure 5-2) and amino acid phylogeny (data not shown) shared the same topology, except for several closely related strains.

Our data revealed that three genera formed two major well-solved clades (Figure 5-2).

One clade consisted of species belonging to the genera Brenneria and Pectobacterium, while the genus Dickeya formed the other clade. Interestingly, strains of P. carotovorum were clustered into five distinct clades: clade 1, for P. carotovorum subsp. brasiliense strains and another five

P. carotovorum subsp. carotovorum strains (YC D60, PCC21, BC S2, BC D6, and YC D46); clade 2, for two P. carotovorum subsp. carotovorum strains (NCPPB 3395 and YC T1); clade 3, for most P. carotovorum subsp. carotovorum strains; clade 4, for P. carotovorum subsp. odoriferum strains; and clade 5, for P. carotovorum subsp. actinidiae KKH3T. Clade 1 was split further as follows: one subclade contained the type strain P. carotovorum subsp. brasiliense

LMG 21371T as well as another four P. carotovorum subsp. brasiliense strains, and the other subclade harbored ten P. carotovorum subsp. brasiliense strains and five P. carotovorum subsp. carotovorum strains. This indicates that significant genetic diversity exists among different P. carotovorum subsp. brasiliense strains, and five P. carotovorum subsp. carotovorum strains (YC

D60, PCC21, BC S2, BC D6, and YC D46) appear to be misnamed. Furthermore, two P.

106

carotovorum subsp. carotovorum strains (UGC32 and PC1) could not be grouped into clade 3, which contained the type strain P. carotovorum subsp. carotovorum ICMP 5702T. These two strains could not be clustered together with any other strain either; they therefore may represent two novel species. Similarly, several Dickeya strains were not clustered with type strains of their species, namely, D. solani M005, D. chrysanthemi M074, and D. dadantii Ech586.

Assessment of Genome Similarity Using isDDH and ANI

Here, we used two different bioinformatics methods, isDDH and ANI, to test whether such in silico approaches can be used as alternatives of wet-lab DDH to determine species delineation of these plant-pathogenic strains. The isDDH and ANI values (Figures 5-3, 5-4, and

Table 5-2) were consistent with their phylogenetic relationships (Figure 5-2): strains grouped in the same phylogenetic clade also showed high ANI and isDDH values.

The ANI and isDDH values among different subspecies of P. carotovorum were very low

(Table 5-2), below the suggested 96% cutoff for the ANI value (Richter and Rosselló-Mora

2009) and 70% cutoff for the isDDH value (Auch et al. 2010a) for species delineation. Based on

ANI and isDDH values, strains of P. carotovorum was grouped into five different genomospecies (Table 5-2), corresponding to the five clades (GI-GV) in the phylogenetic tree

(Figure 5-2). In other words, P. carotovorum’s four subspecies (actinidiae, odoriferum, carotovorum, and brasiliense) can be evaluated to the species level, and P. carotovorum subsp. carotovorum NCPPB 3395 and P. carotovorum subsp. carotovorum YC T1 can represent a novel species. Furthermore, another two strains (P. carotovorum subsp. carotovorum UGC32 and P. carotovorum subsp. carotovorum PC1) are not the same genomospecies with any other strain

(data not shown), indicating these two strains were misidentified. We also found strains of P. wasabiae can also be classified into two species (Figure 5-3): one species includes the type strain

P. wasabiae NCPPB 3701T, P. wasabiae NCPPB3702, and P. wasabiae CFBP 3304; the other

107

species consists of P. wasabiae RNS08.42.1A, P. wasabiae CFIA1002, and P. wasabiae

WPP163.

It is clear that several Dickeya strains were also misnamed (Figures 5-2, 5-4), namely, D. solani M005, D. chrysanthemi M074, and D. dadantii Ech586. D. solani M005 and D. chrysanthemi M074 are closely related, and may represent a novel species, while D. dadantii

Ech586 and strains of D. zeae distribute into related but distinctive genomospecies. The ANI and isDDH values between the two subspecies of D. dadantii (dadantii and dieffenbachiae) are about

97% and 75%, respectively. As the cutoff 79% DDH value is usually used for subspecies delineation (Meier-Kolthoff et al. 2013), the current classification of D. dadantii into two subspecies is appropriate.

Discussion

Here, by combining the whole genome-based phylogeny with ANI and isDDH estimations of 85 genomes, we showed the potential of using whole-genome sequences to evaluate the taxonomy of plant-pathogenic genera Dickeya and Pectobacterium. These two genera contain important plant pathogens; however, the taxonomic positions of some species and subspecies of these genera have been controversial. Reliable identification and accurate classification of pathogens are important for legislating against and establishing quarantine for specific pathogens (Toth et al. 2015).

Concatenation, compilation and analysis of hundreds of genes as a single data set has became the standard approach for reconstructing major branches of the tree of life (Salichos and

Rokas 2013). In this study, we reconstruct the phylogeny using 895 single-copy orthologous genes. Single-copy genes have been demonstrated to be resistant to horizontal gene transfer

(HGT), even in ancient bacterial groups subject to high rates of HGT (Lerat et al. 2003).

However, phylogeny-based approaches have a limitation: it infers the evolutionary relationships

108

but does not provide the common threshold that can be used for substituting DDH for taxonomy definition. This also can explain why traditional DDH has never been abandoned, even it has been 20 years since the first bacterial genome was sequenced (Fleischmann et al. 1995). The low cost of genome sequencing has enabled plant pathologist to make use of this technique to sequence the pathogen genomes on a daily basis. However, most of these genomes were just sequenced in the search for pathogenicity and virulence factors, and we never take full advantage of these valuable information. With so many complete genome sequences now on hand, a reliable, reproducible and informative taxonomic scheme will be established.

Previously reported by studies in other genera, namely, Aeromonas, Erwinia, Pantoea, and Ralstonia, most of the intraspecies ANI values were ≥ 96% and isDDH values were ≥ 70%

(Colston et al. 2014; Zhang and Qiu 2015; 2016). In our study, phylogenetic relationships, ANI and isDDH values were by and large consistent with the current understanding of the taxonomic relationships of these phytopathogenic species, which had been usually based on wet-lab DDH values, biochemical tests, 16S rRNA, and MLSA (Brady et al. 2012; Nabhan et al. 2012).

However, the ANI and isDDH from this study and the laboratory-based DDH values from previous studies gave conflicting indications as to whether P. carotovorum’s four subspecies

(actinidiae, odoriferum, carotovorum, and brasiliense) should be at the subspecies level or elevated to the species level. Sequences of 16S rRNA gene and intergenic spacer regions, and biochemical reaction analysis (Duarte et al. 2004), MLSA (Ma et al. 2007), and a pan-genomic approach (Glasner et al. 2008) all suggested that P. carotovorum subsp. brasiliense should represent a separates species. The isDDH values between P. carotovorum subsp. brasiliense and other P. carotovorum subspecies obtained in the present study were only 41-58% (data not shown), consistently substantiating its differentiation from other P. carotovorum strains.

109

However, the laboratory-based DDH values between P. carotovorum subsp. brasiliense and the

P. carotovorum type strain was 73-77% (Nabhan et al. 2012). Similarly, isDDH values of P. carotovorum subsp. actinidiae and P. carotovorum subsp. odoriferum with the P. carotovorum type strain were also lower than laboratory-based DDH values (Koh et al. 2012). Such conflict was also reported in a recent study of whether D. solani represented a novel species: the laboratory-based DDH value exceeded 70%, but the ANI was below 96%, and all the evidence from the phenotypic or molecular characterizations supported D. solani to be a separate species

(van der Wolf et al. 2014). Due to the irreproducibility of laboratory-based DDH, these DDH experiments were carried out only in single studies, but not validated by other labs. Whether we should continue to use the tedious traditional DDH or shift to the modern genome-based in silico methods has been the subject of much debate (Sentausa and Fournier 2013). However, in silico methods without doubt will speed up the process of cataloging prokaryotes. Whether P. carotovorum’s four subspecies ought to be considered as separate species requires further study.

At present, a reliable classification can only be achieved by the exploration of two sources of information: genomic information and phenotype (Rosselló-Mora and Amann 2001). A valid prokaryotic species designation requires the species to be characterized for at least one discriminative trait, and one isolate to be selected as its type strain (Rosselló-Mora and Amann

2001).

Another important finding of our study was that, out of the 85 Dickeya, Pectobacterium and Brenneria genomes that we analyzed, at least ten (11.8%) were misnamed, namely, P. carotovorum subsp. carotovorum YC D60, P. carotovorum subsp. carotovorum PCC21, P. carotovorum subsp. carotovorum BC S2, P. carotovorum subsp. carotovorum BC D6, P. carotovorum subsp. carotovorum YC D46, P. carotovorum subsp. carotovorum UGC32, P.

110

carotovorum subsp. carotovorum PC1, D. dadantii Ech586, D. solani M005, and D. chrysanthemi M074. Because wet-lab DDH is tedious and performed in only few specialized labs internationally, genome project submitters do not have easy and reliable tools to classify their strains to the correct taxonomic positions, leading to so many misnamed genome sequencing projects in GenBank. More importantly, these misnamed genomes deposited in

GenBank will sometimes cause the further misidentification of other genomes. At present, as only the original genome sequence submitters are allowed to revise the strain name

(http://www.ncbi.nlm.nih.gov/books/NBK51157/), the submitters should take advantage of the whole-genome sequences and in silico bioinformatics tools, such as ANI and isDDH, to assign strains the right name.

Taxonomy of plant pathogens is critically important for growers, plant pathologist, and policy-makers. However, the traditional classification methods, such as DDH, could not help researchers to classify pathogens to the correct taxonomic positions. Thanks to the low cost of next-generation sequencing technologies, a growing number of plant pathogen genomes have been sequenced and deposited in GenBank. It is a pity that, to date, taxonomists have barely taken advantage of the wealth of information contained in the whole-genome sequencing data. In this study, we demonstrated the potential of using whole genome sequences to resolve complex taxonomic questions of the genera Dickeya and Pectobacterium. We confirmed that a number of genomes deposited in GenBank were misnamed, and some strains represented new species. This study also obtained evidence to propose the elevation of the four subspecies of P. carotovorum to the species level. The phylogenomic and systematic strategy presented here will allow researchers to use genome sequences to identify and classify their isolates. Eventually, the in

111

silico methods based on whole genome sequences will substitute traditional DDH as the new gold standard for bacterial species delineation.

112

Table 5-1. Features of genomes used in this study Genus Species Subspecies Strain Genome GenBank Isolation source Reference/genome submitter name length (Mb) Accession# Brenneria sp. EniD312 4.9 CM001230.1 NA DOE Joint Genome Institute Brenneria goodwinii OBR1 5.4 CGIG00000000.1 NA University of Liverpool, UK Dickeya chrysanthemi NCPPB 516 4.6 CM001904.1 Parthenium (Pritchard et al. 2013b) argentatum Dickeya chrysanthemi NCPPB 402T 4.7 CM001974.1 Chrysanthemum (Pritchard et al. 2013b) Dickeya chrysanthemi NCPPB 3533 4.7 CM001981.1 Potato (Pritchard et al. 2013b) Dickeya chrysanthemi M074 5 JRWY00000000.1 Waterfall University of Malaya, Malaya Dickeya chrysanthemi L11 4.8 JSYH00000000.1 Lake Water University of Malaya, Malaya Dickeya dadantii dadantii 3937 4.9 CP002038.1 Pear (Glasner et al. 2011) Dickeya dadantii Ech586 4.8 CP001836.1 Philodendron DOE Joint Genome Institute Dickeya dadantii dadantii NCPPB 898T 4.9 CM001976.1 Pelargonium (Pritchard et al. 2013b) Dickeya dadantii dieffenbachiae NCPPB 2976T 4.8 CM001978.1 Dieffenbachia (Pritchard et al. 2013b) Dickeya dadantii dadantii NCPPB 3537 4.8 CM001982.1 Potato (Pritchard et al. 2013b) Dickeya dianthicola IPO 980 4.8 CM002023.1 Potato (Pritchard et al. 2013a) Dickeya dianthicola RNS04.9 4.7 APVF00000000.1 Potato (Raoul des Essarts et al. 2015) Dickeya dianthicola NCPPB 453T 4.7 CM001841.1 Dianthus (Pritchard et al. 2013a) Dickeya dianthicola GBBC 2039 4.8 CM001838.1 Potato (Pritchard et al. 2013a) Dickeya dianthicola NCPPB 3534 4.9 CM001840.1 Potato (Pritchard et al. 2013a) Dickeya paradisiaca NCPPB 2511T 4.6 CM001857.1 Plantain (Pritchard et al. 2013b) Dickeya solani IFB 0099 5.1 JXRS00000000.1 Potato (Golanowska et al. 2015) Dickeya solani IPO 2222T 4.9 CM001859.1 Potato (Pritchard et al. 2013a) Dickeya solani MK10 4.9 CM001839.1 Potato (Pritchard et al. 2013a) Dickeya solani MK16 4.9 CM001842.1 River water (Pritchard et al. 2013a) Dickeya solani GBBC 2040 4.9 CM001860.1 Potato (Pritchard et al. 2013a) Dickeya solani RNS 08.23.3.1.A 4.9 AMYI00000000.1 Potato (Khayi et al. 2014) Dickeya solani D s0432-1 4.9 AMWE00000000.1 Potato (Garlant et al. 2013) Dickeya solani M005 5.1 JSXD00000000.1 Waterfall University of Malaya, Malaya Dickeya zeae MK19 4.7 CM001985.1 River water (Pritchard et al. 2013b) Dickeya zeae MS1 4.7 APWM00000000.1 Banana (Zhang et al. 2013) Dickeya zeae DZ2Q 4.6 APMV00000000.1 Rice (Bertani et al. 2013)

113

Table 5-1. Continued Genus Species Subspecies Strain Genome GenBank Isolation source Reference/genome submitter name length (Mb) Accession# Dickeya zeae EC1 4.5 CP006929.1 Rice (Zhou et al. 2015) Dickeya zeae NCPPB 3532 4.6 CM001858.1 Potato (Pritchard et al. 2013b) Dickeya zeae CSL RW192 4.7 CM001972.1 River water (Pritchard et al. 2013b) Dickeya zeae ZJU1202 4.6 AJVN00000000.1 Rice (Li et al. 2012)9

Dickeya zeae NCPPB 3531 4.6 CM001980.1 Potato (Pritchard et al. 2013b) Pectobacterium carotovorum actinidiae KKH3T 4.9 JRMH00000000.1 Kiwi fruit National Academy of Agricultural Science, South Korea Pectobacterium carotovorum brasiliense CFIA1009 4.8 JPSN00000000.1 Potato (Li et al. 2015a) Pectobacterium carotovorum brasiliense CFIA1033 4.7 JPSO00000000.1 Potato (Li et al. 2015a) Pectobacterium carotovorum brasiliense LMG 21371T 4.8 JQOE00000000.1 Potato Food and Environment Research Agency, UK Pectobacterium carotovorum brasiliense LMG 21372 4.9 JQOD00000000.1 Potato Food and Environment Research Agency, UK Pectobacterium carotovorum brasiliense YC D49 4.7 JUJB00000000.1 Chinese cabbage Beijing Academy of Agriculture and Forestry Sciences, China Pectobacterium carotovorum brasiliense YC D62 4.9 JUJD00000000.1 Chinese cabbage Beijing Academy of Agriculture and Forestry Sciences, China Pectobacterium carotovorum brasiliense YC D64 4.8 JUJE00000000.1 Chinese cabbage Beijing Academy of Agriculture and Forestry Sciences, China Pectobacterium carotovorum brasiliense YC D21 4.8 JUJJ00000000.1 Chinese cabbage Beijing Academy of Agriculture and Forestry Sciences, China Pectobacterium carotovorum brasiliense YC D29 4.8 JUJK00000000.1 Chinese cabbage Beijing Academy of Agriculture and Forestry Sciences, China Pectobacterium carotovorum brasiliense YC T31 4.8 JUJL00000000.1 Chinese cabbage Beijing Academy of Agriculture and Forestry Sciences, China Pectobacterium carotovorum brasiliense YC D65 4.7 JUJN00000000.1 Chinese cabbage Beijing Academy of Agriculture and Forestry Sciences, China Pectobacterium carotovorum brasiliense YC D52 4.8 JUJC00000000.1 Chinese cabbage Beijing Academy of Agriculture and Forestry Sciences, China Pectobacterium carotovorum brasiliense BD255 4.8 LGRF00000000.1 Potato University of South Africa, South Africa Pectobacterium carotovorum brasiliense PBR1692 4.9 ABVX00000000.1 NA (Glasner et al. 2008) Pectobacterium carotovorum brasiliensie ICMP 19477 5 ALIU00000000.1 Potato (Panda et al. 2015a) Pectobacterium carotovorum carotovorum PC1 4.9 CP001657.1 Potato DOE Joint Genome Institute Pectobacterium carotovorum carotovorum PCC21 4.8 CP003776.1 Chinese Cabbage (Park et al. 2012) Pectobacterium carotovorum carotovorum WPP14 4.8 ABVY00000000.1 NA (Glasner et al. 2008) Pectobacterium carotovorum carotovorum NCPPB 312 4.8 JQHJ00000000.1 Potato Food and Environment Research Agency, UK

114

Table 5-1. Continued Genus Species Subspecies Strain Genome GenBank Isolation source Reference/genome submitter name length (Mb) Accession# Pectobacterium carotovorum carotovorum NCPPB 3395 4.6 JQHN00000000.1 Potato Food and Environment Research Agency, UK Pectobacterium carotovorum carotovorum YC D57 4.8 JUJG00000000.1 Chinese cabbage Beijing Academy of Agriculture and Forestry Sciences, China Pectobacterium carotovorum carotovorum YC D46 4.8 JUJH00000000.1 Chinese cabbage Beijing Academy of Agriculture and Forestry Sciences, China Pectobacterium carotovorum carotovorum BC D6 4.7 JUJT00000000.1 Chinese cabbage Beijing Academy of Agriculture and Forestry Sciences, China Pectobacterium carotovorum carotovorum BC S2 4.8 JUJF00000000.1 Chinese cabbage Beijing Academy of Agriculture and Forestry Sciences, China Pectobacterium carotovorum carotovorum YC T1 3.9 JUJI00000000.1 Chinese cabbage Beijing Academy of Agriculture and Forestry Sciences, China Pectobacterium carotovorum carotovorum YC D16 4.7 JUJO00000000.1 Chinese cabbage Beijing Academy of Agriculture and Forestry Sciences, China Pectobacterium carotovorum carotovorum YC D60 4.8 JUJP00000000.1 Chinese cabbage Beijing Academy of Agriculture and Forestry Sciences, China Pectobacterium carotovorum carotovorum YC T39 5 JUJQ00000000.1 Chinese cabbage Beijing Academy of Agriculture and Forestry Sciences, China Pectobacterium carotovorum carotovorum BC T2 4.8 JUJR00000000.1 Chinese cabbage Beijing Academy of Agriculture and Forestry Sciences, China Pectobacterium carotovorum carotovorum BC T5 4.8 JUJS00000000.1 Chinese cabbage Beijing Academy of Agriculture and Forestry Sciences, China Pectobacterium carotovorum carotovorum ICMP 5702T 4.8 AODT00000000.1 Potato (Panda et al. 2015b) Pectobacterium carotovorum carotovorum UGC32 4.8 AODU00000000.1 Potato (Panda et al. 2015a) Pectobacterium carotovorum odoriferum NCPPB3841 5.5 JQOF00000000.1 Chicory Food and Environment Research Agency, UK Pectobacterium carotovorum odoriferum NCPPB 3839T 5.1 JQOG00000000.1 Chicory Food and Environment Research Agency, UK Pectobacterium carotovorum odoriferum BC S7 4.9 CP009678.1 Chinese cabbage Beijing Academy of Agriculture and Forestry Sciences, China Pectobacterium atrosepticum NCPPB 3404 4.9 JQHO00000000.1 Potato Food and Environment Research Agency, UK Pectobacterium atrosepticum SCRI1043 5.1 BX950851.1 Potato (Bell et al. 2004) Pectobacterium atrosepticum JG10-08 5 CP007744.1 Potato Agricultural University of Hebei, China Pectobacterium atrosepticum 21A 5 CP009125.1, Potato (Nikolaichik et al. 2014) CP009126.1 Pectobacterium atrosepticum CFBP 6276 4.9 ASAB00000000.1 Potato (Kwasiborski et al. 2013) Pectobacterium atrosepticum NCPPB 549T 5.1 JQHK00000000.1 Potato Food and Environment Research Agency, UK Pectobacterium atrosepticum ICMP 1526 4.9 ALIV00000000.1 Potato (Panda et al. 2015a)

115

Table 5-1. Continued Genus Species Subspecies Strain Genome Accession# Isolation source Reference/genome submitter length (Mb) Pectobacterium betavasculor NCPPB 2795T 4.7 JQHM00000000.1 Sugar beet Food and Environment Research um Agency, UK Pectobacterium betavasculor NCPPB 2793 4.7 JQHL00000000.1 Sugar beet Food and Environment Research um Agency, UK Pectobacterium wasabiae RNS08.42.1A 5 JMDL00000000.1 Potato Institut des Sciences du Vegetal, France Pectobacterium wasabiae CFBP 3304 5.1 AKVS00000000.1 Wasabi (Nykyri et al. 2012) Pectobacterium wasabiae CFIA1002 5 JENG00000000.1 Potato (Yuan et al. 2014) Pectobacterium wasabiae NCPPB 3701T 5 JQHP00000000.1 Wasabi Food and Environment Research Agency, UK Pectobacterium wasabiae NCPPB3702 5 JQOH00000000.1 Wasabi Food and Environment Research Agency, UK Erwinia amylovora ATCC 49946 3.9 GCA_000027205.1 Apple (Sebaihia et al. 2010)

Pantoea ananatis LMG 5342 4.9 GCA_000283875.1 Human wound (De Maayer et al. 2012)

116

Table 5-2. Pairwise isDDH and ANI values for different genomospecies of Pectobacterium carotovorum I II III IV V

I 96 - 100 (67 - 100) 56 - 58 54 - 58 51 - 52 47 - 48

II 94 99 (98) 56 - 57 51 - 54 47 - 49

III 93 - 94 94 96 - 100 (73 - 100) 61- 66 52 - 54

IV 92 - 93 93 95 99 (93 - 95) 53 - 54

V 92 92 93 93 - Note: Data are shown as ranges of pairwise percent values of isDDH and ANI for different strains. Numbers in bold represent the pairwise isDDH percent values, and the rest numbers represent the pairwise ANI percent values. Strains belonging to each genomospecies are listed in Figure 5-2.

117

Figure 5-1. The pipeline for phylogenomic and systematic analysis of the genera Dickeya and Pectobacterium. i) 87 genomes were analyzed by OrthoMCL v1.4 to determine single-copy orthologous groups. ii) each single-copy orthologous group was aligned at the amino acid level by Probalign and ambiguously aligned regions were removed. iii) amino acid sequences of filtered alignments were concatenated and submitted to RAxML to reconstruct the whole-genome-based amino acid phylogeny. iv) nucleic acid sequences of filtered alignments were concatenated and submitted to RAxML to reconstruct the whole-genome-based nucleic acid phylogeny; v) genome similarities were assessed using ANI. vi) genome similarities were assessed using isDDH, vii) taxonomy was evaluated based on the combination of phylogeny, ANI and isDDH.

118

Figure 5-2. Phylogenetic reconstruction of the genera Dickeya and Pectobacterium based on the concatenation of nucleic acid sequences of all 895 single-copy orthologs using the maximum likelihood algorithm RAxML. 1000 bootstrap replicates were computed. Numbers above branches are bootstrap percentages. Erwinia amylovora ATCC 49946 and Pantoea ananatis LMG 5342 were used as the outgroups (not shown on the tree). Strains belonging to the same species are marked with the same color. GI, GII, GIII, GIV and GV represent the five clades of Pectobacterium carotovorum. Based on the pairwise ANI and isDDH values (Table 5-2), these five clades also correspond to the five genomospecies of Pectobacterium carotovorum. Strains marked with a superscript “T” represent the type strains of the species or subspecies.

119

Figure 5-3. Pairwise ANI and isDDH values of Pectobacterium atrosepticum, Pectobacterium betavasculorum, and Pectobacterium wasabiae. The lower triangle displays ANI values (%), and the upper triangle displays isDDH values (%). Displayed isDDH values are point estimates plus the 95% model-based confidence intervals. Boxes with ANI ≥ 96% or isDDH ≥ 69.5% are colored red. Strains belonging to the same species are marked with the same color, and strains marked with a superscript “T” represent the type strains of the species or subspecies.

120

Figure 5-4. Pairwise ANI and isDDH values of the genus Dickeya. The lower triangle displays ANI values (%), and the upper triangle displays isDDH values (%). Displayed isDDH values are point estimates plus the 95% model-based confidence intervals. Boxes of ANI ≥ 96% or isDDH ≥ 69.5% are colored red. Strains belonging to the same species are marked with the same color, and strains marked with a superscript “T” represent the type strains of the species or subspecies.

121

CHAPTER 6 CONCLUSIONS AND PERSPECTIVES

To date, at least 11 Streptomyces species have been identified to be the casual agent of the globally important potato scab disease. However, only four species, namely Streptomyces scabiei, Streptomyces acidiscabies, Streptomyces turgidiscabies, and Streptomyces ipomoea have been well studied (Loria et al. 2008). Although Streptomyces is one of the most highly sequenced genera now, many genomes are sequenced with the intent of mining secondary metabolism gene clusters, with considerably less attention being paid to the genomes of plant pathogenic

Streptomyces species. When we initiated this study, only four genomes (S. scabiei 87-22, S. acidiscabies 84-104, S. turgidiscabies Car8, and S. ipomoea 91-03) of pathogenic Streptomyces species were available. These four genomes have provided us important insights into genetic basis of pathogenesis of the genus Streptomyces. The large and mobile pathogenicity island (PAI) of S. turgidiscabies Car8 was identified in 2005 (Kers et al. 2005), however, comparative genomic analysis of four pathogenic genomes revealed related but not identical PAIs. These four genomes have been insufficient in determining how PAIs evolved and how the pathogenic species emerged in the nature environment.

To obtain a more comprehensive picture of the origin and evolution of pathogenicity in the Streptomyces genus, we sequenced another 14 plant pathogenic Streptomyces isolates of different origins. Luckily, other Streptomyces researchers also sequenced some pathogenic

Streptomyces genomes recently (Labeda 2016; Tomihama et al. 2016). With these pathogenic

Streptomyces genomes on hand, we were able to accurately infer the evolutionary relationships of pathogenic species and related nonpathogenic species, which is essential for our understanding of the origin and evolution of pathogenesis of the Streptomyces genus. In silico genome similarity analyses (ANI and isDDH) demonstrated that several pathogenic

122

Streptomyces isolates were misidentified. Applying the phylogenomic and in silico genome similarity analyses to other important plant pathogenic bacterial genera, namely Pectobacterium and Dickeya, revealed misidentification and misclassification of isolates is a common problem.

The low cost of genome sequencing will enable us to use whole genome sequences to re-evaluate current prokaryotic species definition and establish a unified prokaryotic species definition frame for taxonomically challenging genera.

Genomes of pathogenic Streptomyces isolates also provide us opportunities to characterize the PAI variation within and among pathogenic Streptomyces species using more than 30 pathogenic Streptomyces genomes. The PAI of S. turgidiscabis Car8 is the first identified

PAI in the Streptomyces genus and its mobility has been demonstrated (Kers et al. 2005), bringing up the possibility of the appearance of this huge island in new pathogenic species.

However, the toxicogenic region (TR) identified in S. scabiei 87-22 is the major PAI version contained not only by S. scabiei, but also by other pathogenic species, including S. europaeiscabiei, S. stelliscabiei, S. acidiscabies, and S. niveiscabiei. Pathogenic isolates containing the S. turgidiscabies-type PAI has not been discovered to date. This probably because the 674 Kb Car8-type of PAI cannot be stably transferred or maintained. Considerable variability in the composition of PAI genes has been reported for S. turgidiscabies isolates in Finland, suggesting that the mosaic regions of Car8-type PAI may undergo independent evolution

(Aittamaa et al. 2010). Further genome sequencing projects including S. turgidiscabies isolates with different compositions of PAI genes can contribute to our understanding of independent evolution of this large mosaic PAI.

Interestingly, our results revealed high similarity of the thaxtomin biosynthetic gene

(ThxA) clusters in four phylogenetically distant pathogenic taxa (S. acidiscabies, S. niveiscabiei,

123

S. sp. 96-12 and S. scabiei), strongly suggesting a recent HGT of PAI among them. S. sp. 96-12, a pathogenic isolate from potatoes in Egypt, is phylogentically related to a non-pathogenic species rather than with other known pathogenic species, but S. sp. 96-12 has an almost identical

TR region with S. scabiei. In these two taxa, two flanking and one internal attachment (att) sites further separated the TR region into two sub-regions, designated TR1 and TR2. The 20 Kb TR1 region contains the ThxA biosynthetic cluster, and the 157 Kb TR2 region includes a putative integrative and conjugative elements (ICE). Although the recombination among three att sites resulted in three distinct circular structures, as TR1, as TR2, and as the entire TR region, only

TR2 or the entire TR region was transferred into the recipient genome. Recipients gained pathogenic only after receiving the entire TR region. It seems that TR1 could be a cis- mobilizable element (CIME), which is a type of genomic islands that does not encode proteins involved in in recombination and conjugation and is flanked by att sites (Bellanger et al. 2014).

TR region could be a composite element probably stem from the accretion of a putative ICE

(TR2), and a CIME (TR1). Their site-specific accretion leads to the cis-mobilization and conjugative transfer of the ThxA cluster and can be responsible for the emergence of novel

Streptomyces pathogenic species in the agricultural system. Since the Streptomyces genus is ubiquitous in soils, our study brings up the risk that new pathogenic species will continuously appear.

In the last two years, the study of pathogenicity of Streptomyces has been focused on thaxtomins. However, a growing body of evidence has shown that other pathogenicity/virulence factors exist in Streptomyces pathogens. FD-891, a 16-membered polyketide macrolide isolated from Streptomyces cheloniumii, was reported to be responsible for the potato russet scab disease in Japan (Suzui et al. 1988). Borrelidin, an 18-membered polyketide macrolide purified from the

124

culture of Steptomyces sp. GK18, was shown to induce deep pitted lesions on potato tubers.

Recently, a phytotoxic substance isolated from a pathogenic Streptomyces strain 11-1-2 was shown to cause significant pitting of potato tuber tissue, and induce severe necrosis of the leaf tissue in Nicotiana benthamiana leaf infiltration assay (Fyans et al. 2016). All the pathogenic isolates mentioned above do not produce thaxtomins. Comparative genomics analysis presented here also revealed several pathogenic species lack the ThxA cluster, namely S. puniciscabiei and

S. luridiscabiei. In addition, although S. niveiscabiei NRRL B-24457 from Korea contains the

ThxA cluster, two S. niveiscabiei isolates (ST1015 and ST1020) from Uruguay causing pitted scab lesions on potatoes (M. Julia-Pianzzola, personal communication) do not contain the ThxA cluster. These results suggest that although ThxA is still the main pathogenicity factor of pathogenic Streptomyces species, more work is required to characterize the pathogenic

Streptomyces isolates carrying novel pathogenicity/virulence factors, which are important for our understanding of their disease mechanisms and the development of strategies for disease management.

125

LIST OF REFERENCES

Abramovitch, R. B., and Martin, G. B. 2004. Strategies used by bacterial pathogens to suppress plant defenses. Curr Opin Plant Biol. 7:356–364

Achtman, M., and Wagner, M. 2008. Microbial diversity and the genetic nature of microbial species. Nat Rev Microbiol. 6:431–440

Aguileta, G., Marthey, S., Chiapello, H., Lebrun, M. H., Rodolphe, F., Fournier, E., Gendrault- Jacquemard, A., and Giraud, T. 2008. Assessing the performance of single-copy genes for recovering robust phylogenies. Systematic Biol. 57:613–627

Aittamaa, M., Somervuo, P., Laakso, I., Auvinen, P., and Valkonen, J. P. T. 2010. Microarray- based comparison of genetic differences between strains of Streptomyces turgidiscabies with focus on the pathogenicity island. Mol. Plant Pathol. 11:733–746

Alam, M. T., Merlo, M. E., Takano, E., and Breitling, R. 2010. Genome-based phylogenetic analysis of Streptomyces and its relatives. Mol. Phylogenet. Evol. 54:763–772

Alfano, J. R., and Collmer, A. 1996. Bacterial pathogens in plants: Life up against the wall. Plant Cell. 8:1683–1698

Auch, A. F., Jan, von, M., Klenk, H.-P., and Göker, M. 2010a. Digital DNA-DNA hybridization for microbial species delineation by means of genome-to-genome sequence comparison. Stand Genomic Sci. 2:117–134

Auch, A. F., Klenk, H.-P., and Goeker, M. 2010b. Standard operating procedure for calculating genome-to-genome distances based on high-scoring segment pairs. Stand Genomic Sci. 2:142–148

Aziz, R. K., Bartels, D., Best, A. A., DeJongh, M., Disz, T., Edwards, R. A., Formsma, K., Gerdes, S., Glass, E. M., Kubal, M., Meyer, F., Olsen, G. J., Olson, R., Osterman, A. L., Overbeek, R. A., McNeil, L. K., Paarmann, D., Paczian, T., Parrello, B., Pusch, G. D., Reich, C., Stevens, R., Vassieva, O., Vonstein, V., Wilke, A., and Zagnitko, O. 2008. The RAST Server: rapid annotations using subsystems technology. BMC Genomics. 9:75

Bankevich, A., Nurk, S., Antipov, D., Gurevich, A. A., Dvorkin, M., Kulikov, A. S., Lesin, V. M., Nikolenko, S. I., Pham, S., Prjibelski, A. D., Pyshkin, A. V., Sirotkin, A. V., Vyahhi, N., Tesler, G., Alekseyev, M. A., and Pevzner, P. A. 2012. SPAdes: a new genome assembly algorithm and its applications to single-cell sequencing. J. Comput. Biol. 19:455–477

Barras, F., and Van Gijsegem, F. 1994. Extracellular enzymes and pathogenesis of soft-rot Erwinia. Annu Rev Phytopathol. 32:201–234

126

Bell, K. S., Sebaihia, M., Pritchard, L., Holden, M. T. G., Hyman, L. J., Holeva, M. C., Thomson, N. R., Bentley, S. D., Churcher, L. J. C., Mungall, K., Atkin, R., Bason, N., Brooks, K., Chillingworth, T., Clark, K., Doggett, J., Fraser, A., Hance, Z., Hauser, H., Jagels, K., Moule, S., Norbertczak, H., Ormond, D., Price, C., Quail, M. A., Sanders, M., Walker, D., Whitehead, S., Salmond, G. P. C., Birch, P. R. J., Parkhill, J., and Toth, I. K. 2004. Genome sequence of the enterobacterial phytopathogen Erwinia carotovora subsp. atroseptica and characterization of virulence factors. Proc. Natl. Acad. Sci. U.S.A. 101:11105–11110

Bellanger, X., Morel, C., Gonot, F., Puymege, A., Decaris, B., and Guedon, G. 2011. Site- specific accretion of an integrative conjugative element together with a related genomic island leads to cis mobilization and gene capture. Mol. Microbiol. 81:912–925

Bellanger, X., Payot, S., Leblond-Bourget, N., and Guedon, G. 2014. Conjugative and mobilizable genomic islands in bacteria: evolution and diversity. FEMS Microbiol. Rev. 38:720–760

Bender, C. L., Alarcon-Chaidez, F., and Gross, D. C. 1999a. Pseudomonas syringae phytotoxins: Mode of action, regulation, and biosynthesis by peptide and polyketide synthetases. Microbiol Mol Biol Rev. 63:266–292

Bender, C. L., Rangaswamy, V., and Loper, J. 1999b. Polyketide production by plant-associated pseudomonads. Annu Rev Phytopathol. 37:175–196

Bentley, S. D., Chater, K. F., Cerdeño-Tárraga, A.-M., Challis, G. L., Thomson, N. R., James, K. D., Harris, D. E., Quail, M. A., Kieser, H., Harper, D., Bateman, A., Brown, S., Chandra, G., Chen, C. W., Collins, M., Cronin, A., Fraser, A., Goble, A., Hidalgo, J., Hornsby, T., Howarth, S., Huang, C. H., Kieser, T., Larke, L., Murphy, L., Oliver, K., O'Neil, S., Rabbinowitsch, E., Rajandream, M.-A., Rutherford, K., Rutter, S., Seeger, K., Saunders, D., Sharp, S., Squares, R., Squares, S., Taylor, K., Warren, T., Wietzorrek, A., Woodward, J., Barrell, B. G., Parkhill, J., and Hopwood, D. A. 2002. Complete genome sequence of the model actinomycete Streptomyces coelicolor A3(2). Nature. 417:141–147

Bertani, I., da Silva, D. P., Abbruscato, P., Pietro Piffanelli, and Venturi, V. 2013. Draft Genome Sequence of the Plant Pathogen Dickeya zeae DZ2Q, Isolated from Rice in Italy. Genome Announc. 1:e00905–13

Bierman, M., Logan, R., O'Brien, K., Seno, E. T., Rao, R. N., and Schoner, B. E. 1992. Plasmid cloning vectors for the conjugal transfer of DNA from Escherichia coli to Streptomyces spp. Gene. 116:43–49

Bignell, D. R. D., Francis, I. M., Fyans, J. K., and Loria, R. 2014a. Thaxtomin A Production and Virulence Are Controlled by Several bld Gene Global Regulators in Streptomyces scabies. Mol Plant Microbe Interact. 27:875–885

Bignell, D. R. D., Fyans, J. K., and Cheng, Z. 2014b. Phytotoxins produced by plant pathogenic Streptomyces species. J. Appl. Microbiol. 116:223–235

127

Bignell, D. R. D., Huguet-Tapia, J. C., Joshi, M. V., Pettis, G. S., and Loria, R. 2010a. What does it take to be a plant pathogen: genomic insights from Streptomyces species. 98:179–194

Bignell, D. R. D., Seipke, R. F., Huguet-Tapia, J. C., Chambers, A. H., Parry, R. J., and Loria, R. 2010b. Streptomyces scabies 87-22 Contains a Coronafacic Acid-Like Biosynthetic Cluster That Contributes to Plant–Microbe Interactions. Mol Plant Microbe Interact. 23:161–175

Bouchek-Mechiche, K., Gardan, L., Andrivon, D., and Normand, P. 2006. Streptomyces turgidiscabies and Streptomyces reticuliscabiei: one genomic species, two pathogenic groups. Int. J. Syst. Evol. Microbiol. 56:2771–2776

Bouchek-Mechiche, K., Gardan, L., Normand, P., and Jouan, B. 2000. DNA relatedness among strains of Streptomyces pathogenic to potato in France: description of three new species, S. europaeiscabiei sp. nov. and S. stelliscabiei sp. nov. associated with common scab, and S. reticuliscabiei sp. nov. associated with netted scab. Int. J. Syst. Evol. Microbiol. 50:91–99

Brady, C. L., Cleenwerck, I., Denman, S., Venter, S. N., Rodríguez-Palenzuela, P., Coutinho, T. A., and De Vos, P. 2012. Proposal to reclassify Brenneria quercina (Hildebrand and Schroth 1967) Hauben et al. 1999 into a new genus, Lonsdalea gen. nov., as Lonsdalea quercina comb. nov., descriptions of Lonsdalea quercina subsp. quercina comb. nov., Lonsdalea quercina subsp. iberica subsp. nov. and Lonsdalea quercina subsp. britannica subsp. nov., emendation of the description of the genus Brenneria, reclassification of Dickeya dieffenbachiae as Dickeya dadantii subsp. dieffenbachiae comb. nov., and emendation of the description of Dickeya dadantii. Int. J. Syst. Evol. Microbiol. 62:1592–1602

Bukhalid, R. A., Chung, S. Y., and Loria, R. 1998. nec1, a Gene Conferring a Necrogenic Phenotype, Is Conserved in Plant-Pathogenic Streptomyces spp. and Linked to a Transposase Pseudogene. Mol Plant Microbe Interact. 11:960–967

Bukhalid, R. A., Takeuchi, T., Labeda, D., and Loria, R. 2002. Horizontal Transfer of the Plant Virulence Gene, nec1, and Flanking Sequences among Genetically Distinct Streptomyces Strains in the Diastatochromogenes Cluster. Appl. Environ. Microbiol. 68:738–744

Burkholder, W. R., McFadden, L. A., and Dimock, E. W. 1953. A bacterial blight of Chrysanthemums. Phytopathology. 43:522–526

Burrus, V., and Waldor, M. K. 2004. Shaping bacterial genomes with integrative and conjugative elements. Res Microbiol. 155:376–386

Canchaya, C., Proux, C., Fournous, G., Bruttin, A., and Brussow, H. 2003. Prophage genomics. Microbiol Mol Biol Rev. 67:238–276

Cao, Z., Khodakaramian, G., Arakawa, K., and Kinashi, H. 2012. Isolation of borrelidin as a phytotoxic compound from a potato pathogenic Streptomyces strain. Biosci. Biotechnol. Biochem. 76:353–357

Capella-Gutierrez, S., Kauff, F., and Gabaldón, T. 2014. A phylogenomics approach for selecting robust sets of phylogenetic markers. Nucleic Acids Res. 42:e54

128

Challis, G. L., and Hopwood, D. A. 2003. Synergy and contingency as driving forces for the evolution of multiple secondary metabolite production by Streptomyces species. Proc. Natl. Acad. Sci. U.S.A. 100:14555–14561

Chapleau, M., Guertin, J. F., Lerat, S., Burrus, V., and Beaulieu, C. 2016. Identification of genetic and environmental factors stimulating excision from Streptomyces scabiei chromosome of the toxicogenic region responsible for pathogenicity. Mol. Plant Pathol. 17:501–509

Clark, C. A., and Watson, B. 1983. Susceptibility of Weed Species of Convolvulaceae to Root- Infecting Pathogens of Sweet-Potato. Plant Dis. 67:907–909

Colston, S. M., Fullmer, M. S., Beka, L., Lamy, B., Gogarten, J. P., and Graf, J. 2014. Bioinformatic genome comparisons for taxonomic and phylogenetic assignments using Aeromonas as a test case. MBio. 5:e02136

Czajkowski, R., Pérombelon, M. C. M., Jafra, S., Lojkowska, E., Potrykus, M., Van Der Wolf, J. M., and Sledz, W. 2015. Detection, identification and differentiation of Pectobacterium and Dickeya species causing potato blackleg and tuber soft rot: a review. Ann Appl Biol. 166:18–38

De Klerk, A., McLeod, A., Faurie, R., and Van Wyk, P. S. 1997. Net Blotch and Necrotic Warts Caused by Streptomyces scabies on Pods of Peanut (Arachis hypogaea). Plant Dis. 81:958– 958

De Maayer, P., Chan, W. Y., Rezzonico, F., Bühlmann, A., Venter, S. N., Blom, J., Goesmann, A., Frey, J. E., Smits, T. H. M., Duffy, B., and Coutinho, T. A. 2012. Complete genome sequence of clinical isolate Pantoea ananatis LMG 5342. J Bacteriol. 194:1615–1616

Dees, M. W., and Wanner, L. A. 2012. In Search of Better Management of Potato Common Scab. Potato Res. 55:249–268 del Pilar Marquez-Villavicencio, M., Groves, R. L., and Charkowski, A. O. 2011. Soft Rot Disease Severity Is Affected by Potato Physiology and Pectobacterium taxa. Plant Dis. 95:232–241

Delcher, A. L., Phillippy, A., Carlton, J., and Salzberg, S. L. 2002. Fast algorithms for large- scale genome alignment and comparison. Nucleic Acids Res. 30:2478–2483

Denman, S., Brady, C., Kirk, S., Cleenwerck, I., Venter, S., Coutinho, T., and De Vos, P. 2012. Brenneria goodwinii sp. nov., associated with acute oak decline in the UK. Int. J. Syst. Evol. Microbiol. 62:2451–2456

Dobrindt, U., Hochhut, B., Hentschel, U., and Hacker, J. 2004. Genomic islands in pathogenic and environmental microorganisms. Nat Rev Microbiol. 2:414–424

Doroghazi, J. R., and Buckley, D. H. 2010. Widespread homologous recombination within and between Streptomyces species. ISME J. 4:1136–1143

129

Duarte, V., De Boer, S. H., Ward, L. J., and Oliveira, A. M. R. 2004. Characterization of atypical Erwinia carotovora strains causing blackleg of potato in Brazil. J. Appl. Microbiol. 96:535– 545

Dubnau, D. 1999. DNA uptake in bacteria. Annu. Rev. Microbiol. 53:217–244

Eichenlaub, R., and Gartemann, K.-H. 2011. The Clavibacter michiganensis Subspecies: Molecular Investigation of Gram-Positive Bacterial Plant Pathogens. Annu Rev Phytopathol. 49:445–464

Elesawy, A. A., and Szabo, M. 1979. Isolation and characterization of Streptomyces scabies strains from scab lesions of potato tubers. Designation of the neotype strain of Streptomyces scabies. Acta Microbiol Acad Sci Hung. 26:311–320

Enright, A. J., Van Dongen, S., and Ouzounis, C. A. 2002. An efficient algorithm for large-scale detection of protein families. Nucleic Acids Res. 30:1575–1584

Fiedler, H.-P., Bruntner, C., Bull, A. T., Ward, A. C., Goodfellow, M., Potterat, O., Puder, C., and Mihm, G. 2005. Marine actinomycetes as a source of novel secondary metabolites. Antonie Van Leeuwenhoek. 87:37–42

Fleischmann, R. D., Adams, M. D., White, O., Clayton, R. A., Kirkness, E. F., Kerlavage, A. R., Bult, C. J., Tomb, J. F., Dougherty, B. A., and Merrick, J. M. 1995. Whole-genome random sequencing and assembly of Haemophilus influenzae Rd. Science. 269:496–512

Flores González, R., Velasco, I., and Montes, F. 2008. Detection and characterization of Streptomyces causing potato common scab in Western Europe. 57:162–169

Francis, I. M., Jourdan, S., Fanara, S., Loria, R., and Rigali, S. 2015. The cellobiose sensor CebR is the gatekeeper of Streptomyces scabies pathogenicity. MBio. 6:e02018

Fry, B. A., and Loria, R. 2002. Thaxtomin A: evidence for a plant cell wall target. Physiol Mol Plant Path. 60:1–8

Fyans, J. K., Altowairish, M. S., Li, Y., and Bignell, D. R. D. 2015. Characterization of the Coronatine-Like Phytotoxins Produced by the Common Scab Pathogen Streptomyces scabies. Mol Plant Microbe Interact. 28:443–454

Fyans, J. K., Bown, L., and Bignell, D. R. D. 2016. Isolation and Characterization of Plant- Pathogenic Streptomyces Species Associated with Common Scab-Infected Potato Tubers in Newfoundland. Phytopathology. 106:123–131

Gallois, A., Samson, R., Ageron, E., and Grimont, P. A. D. 1992. Erwinia carotovora subsp. odorifera subsp. nov., Associated with Odorous Soft Rot of Chicory (Cichorium intybus L.). Int. J. Syst. Evol. Microbiol. 42:582–588

130

Gardan, L., Gouy, C., Christen, R., and Samson, R. 2003. Elevation of three subspecies of Pectobacterium carotovorum to species level: Pectobacterium atrosepticum sp. nov., Pectobacterium betavasculorum sp. nov. and Pectobacterium wasabiae sp. nov. Int. J. Syst. Evol. Microbiol. 53:381–391

Garlant, L., Koskinen, P., Rouhiainen, L., Laine, P., Paulin, L., Auvinen, P., Holm, L., and Pirhonen, M. 2013. Genome Sequence of Dickeya solani, a New soft Rot Pathogen of Potato, Suggests its Emergence May Be Related to a Novel Combination of Non-Ribosomal Peptide/Polyketide Synthetase Clusters. Diversity. 5:824–842

Glaeser, S. P., and Kämpfer, P. 2015. Multilocus sequence analysis (MLSA) in prokaryotic taxonomy. Syst. Appl. Microbiol. 38:237–245

Glasner, J. D., Marquez-Villavicencio, M., Kim, H. S., Jahn, C. E., Ma, B., Biehl, B. S., Rissman, A. I., Mole, B., Yi, X., Yang, C. H., Dangl, J. L., Grant, S. R., Perna, N. T., and Charkowski, A. O. 2008. Niche-Specificity and the Variable Fraction of the Pectobacterium Pan-Genome. Mol Plant Microbe Interact. 21:1549–1560

Glasner, J. D., Reverchon, S., Hugouvieux-Cotte-Pattat, N., Condemine, G., Bohin, J.-P., Van Gijsegem, F., Yang, S., Franza, T., Expert, D., Plunkett, G., III, Francisco, M. J. S., Charkowski, A. O., Py, B., Bell, K., Rauscher, L., Rodríguez-Palenzuela, P., Toussaint, A., Holeva, M. C., He, S. Y., Douet, V., Boccara, M., Blanco, C., Toth, I., Anderson, B. D., Biehl, B. S., Mau, B., Flynn, S. M., Barras, F., Lindeberg, M., Birch, P., Tsuyumu, S., Shi, X., Hibbing, M., Yap, M.-N., Carpentier, M., Dassa, E., Umehara, M., Kim, J. F., Rusch, M., Soni, P., Mayhew, G. F., Fouts, D. E., Gill, S. R., Blattner, F. R., Keen, N. T., and Perna, N. T. 2011. Genome sequence of the plant pathogenic bacterium Dickeya dadantii 3937. J Bacteriol. 193:2076–2077

Golanowska, M., Galardini, M., Bazzicalupo, M., Hugouvieux-Cotte-Pattat, N., Mengoni, A., Potrykus, M., Slawiak, M., and Lojkowska, E. 2015. Draft Genome Sequence of a Highly Virulent Strain of the Plant Pathogen Dickeya solani, IFB0099. Genome Announc. 3:e00109–15

Goris, J., Konstantinidis, K. T., Klappenbach, J. A., Coenye, T., Vandamme, P., and Tiedje, J. M. 2007. DNA-DNA hybridization values and their relationship to whole-genome sequence similarities. Int. J. Syst. Evol. Microbiol. 57:81–91

Goyer, C., Faucher, E., and Beaulieu, C. 1996a. Streptomyces caviscabies sp. nov., from Deep- Pitted Lesions in Potatoes in Québec, Canada. Int. J. Syst. Evol. Microbiol. 46:635–639

Goyer, C., Otrysko, B., and Beaulieu, C. 1996b. Taxonomic studies on Streptomycetes causing potato common scab: A review. Can J Plant Pathol. 18:107–113

Grindley, N. D. F., Whiteson, K. L., and Rice, P. A. 2006. Mechanisms of Site-Specific Recombination. Annu Rev Biophys Biomol Struct. 75:567–605

131

Guan, D., Grau, B. L., Clark, C. A., Taylor, C. M., Loria, R., and Pettis, G. S. 2012. Evidence That Thaxtomin C Is a Pathogenicity Determinant of Streptomyces ipomoeae, the Causative Agent of Streptomyces Soil Rot Disease of Sweet Potato. Mol Plant Microbe Interact. 25:393–401

Guo, Y., Zheng, W., Rong, X., and Huang, Y. 2008. A multilocus phylogeny of the Streptomyces griseus 16S rRNA gene clade: use of multilocus sequence analysis for streptomycete systematics. Int. J. Syst. Evol. Microbiol. 58:149–159

Gust, B., Challis, G. L., Fowler, K., Kieser, T., and Chater, K. F. 2003. PCR-targeted Streptomyces gene replacement identifies a protein domain needed for biosynthesis of the sesquiterpene soil odor geosmin. Proc. Natl. Acad. Sci. U.S.A. 100:1541–1546

Hacker, J., Hochhut, B., Middendorf, B., Schneider, G., Buchrieser, C., Gottschalk, G., and Dobrindt, U. 2004. Pathogenomics of mobile genetic elements of toxigenic bacteria. Int J Med Microbiol. 293:453–461

Hauben, L., Moore, E. R. B., Vauterin, L., Steenackers, M., Mergaert, J., Verdonck, L., and Swings, J. 1998a. Phylogenetic Position of Phytopathogens within the Enterobacteriaceae. Syst. Appl. Microbiol. 21:384–397

Hauben, L., Steenackers, M., and Swings, J. 1998b. PCR-Based detection of the causal agent of watermark disease in willows (Salix spp.). Appl. Environ. Microbiol. 64:3966–3971

Hauben, L., Van Gijsegem, F., and Swings, J. 2005. Genus XXIV. Pectobacterium. Bergey's manual of systematic bacteriology.

Healy, F. G., and Lambert, D. H. 1991. Relationships among Streptomyces spp. Causing Potato Scab†. Int. J. Syst. Evol. Microbiol. 41:479–482

Healy, F. G., Krasnoff, S. B., Wach, M., Gibson, D. M., and Loria, R. 2002. Involvement of a cytochrome P450 monooxygenase in thaxtomin A biosynthesis by Streptomyces acidiscabies. J Bacteriol. 184:2019–2029

Healy, F. G., Wach, M., Krasnoff, S. B., Gibson, D. M., and Loria, R. 2000. The txtAB genes of the plant pathogen Streptomyces acidiscabies encode a peptide synthetase required for phytotoxin thaxtomin A production and pathogenicity. Mol. Microbiol. 38:794–804

Herbst, D. A., Boll, B., Zocher, G., Stehle, T., and Heide, L. 2013. Structural basis of the interaction of MbtH-like proteins, putative regulators of nonribosomal peptide biosynthesis, with adenylating enzymes. J. Biol. Chem. 288:1991–2003

Hirano, N., Muroi, T., Takahashi, H., and Haruki, M. 2011. Site-specific recombinases as tools for heterologous gene integration. Appl Microbiol Biotechnol. 92:227–239

Hogenhout, S. A., and Loria, R. 2008. Virulence mechanisms of Gram-positive plant pathogenic bacteria. Curr Opin Plant Biol. 11:449–456

132

Holden, M. T. G., Heather, Z., Paillot, R., Steward, K. F., Webb, K., Ainslie, F., Jourdan, T., Bason, N. C., Holroyd, N. E., Mungall, K., Quail, M. A., Sanders, M., Simmonds, M., Willey, D., Brooks, K., Aanensen, D. M., Spratt, B. G., Jolley, K. A., Maiden, M. C. J., Kehoe, M., Chanter, N., Bentley, S. D., Robinson, C., Maskell, D. J., Parkhill, J., and Waller, A. S. 2009. Genomic evidence for the evolution of Streptococcus equi: host restriction, increased virulence, and genetic exchange with human pathogens. M.R. Wessels, ed. PLoS Pathog. 5:e1000346

Hopwood, D. A. 2006. Soil to genomics: the Streptomyces chromosome. Annu. Rev. Genet. 40:1–23

Huddleston, J. R. 2014. Horizontal gene transfer in the human gastrointestinal tract: Potential spread of antibiotic resistance genes. Infect Drug Resist. :167–176

Huguet-Tapia, J. C., and Loria, R. 2012. Draft genome sequence of Streptomyces acidiscabies 84-104, an emergent plant pathogen. J Bacteriol. 194:1847–1847

Huguet-Tapia, J. C., Badger, J. H., Loria, R., and Pettis, G. S. 2011. Streptomyces turgidiscabies Car8 contains a modular pathogenicity island that shares virulence genes with other actinobacterial plant pathogens. Plasmid. 65:118–124

Huguet-Tapia, J. C., Bignell, D. R. D., and Loria, R. 2014. Characterization of the integration and modular excision of the integrative conjugative element PAISt in Streptomyces turgidiscabies Car8. PLoS ONE. 9:e99345

Huguet-Tapia, J. C., Lefébure, T., Badger, J. H., Guan, D., Pettis, G. S., Stanhope, M. J., and Loria, R. 2016. Genome content and phylogenomics reveal both ancestral and lateral evolutionary pathways in plant pathogenic Streptomyces species. Appl. Environ. Microbiol. 82:2146-2155

Johnson, C. M., and Grossman, A. D. 2015. Integrative and Conjugative Elements (ICEs): What They Do and How They Work. Annu Rev Genet. 49:577–601

Johnson, E. G., Joshi, M. V., Gibson, D. M., and Loria, R. 2007. Cello-oligosaccharides released from host plants induce pathogenicity in scab-causing Streptomyces species. Physiol Mol Plant Path. 71:18–25

Joshi, M. V., and Loria, R. 2007. Streptomyces turgidiscabies possesses a functional cytokinin biosynthetic pathway and produces leafy galls. Mol. Plant Microbe Interact. 20:751–758

Joshi, M. V., Bignell, D. R. D., Johnson, E. G., Sparks, J. P., Gibson, D. M., and Loria, R. 2007a. The AraC/XylS regulator TxtR modulates thaxtomin biosynthesis and virulence in Streptomyces scabies. Mol. Microbiol. 66:633–642

Joshi, M., Rong, X., Moll, S., Kers, J., Franco, C., and Loria, R. 2007b. Streptomyces turgidiscabies secretes a novel virulence protein, Nec1, which facilitates infection. Mol. Plant Microbe Interact. 20:599–608

133

Kataoka, T., Yamada, A., Bando, M., Honma, T., Mizoue, K., and Nagai, K. 2000. FD-891, a structural analogue of concanamycin A that does not affect vacuolar acidification or perforin activity, yet potently prevents cytotoxic T lymphocyte-mediated cytotoxicity through the blockage of conjugate formation. Immunology. 100:170–177

Kaup, O., Gräfen, I., Zellermann, E.-M., Eichenlaub, R., and Gartemann, K.-H. 2007. Identification of a Tomatinase in the Tomato-Pathogenic Actinomycete Clavibacter michiganensis subsp. michiganensis NCPPB382. Mol Plant Microbe Interact. 18:1090–1098

Kers, J. A., Cameron, K. D., Joshi, M. V., Bukhalid, R. A., Morello, J. E., Wach, M. J., Gibson, D. M., and Loria, R. 2005. A large, mobile pathogenicity island confers plant pathogenicity on Streptomyces species. Mol. Microbiol. 55:1025–1033

Khayi, S., Mondy, S., Beury-Cirou, A., Moumni, M., Hélias, V., and Faure, D. 2014. Genome Sequence of the Emerging Plant Pathogen Dickeya solani Strain RNS 08.23.3.1A. Genome Announc. 2:e01270–13

Kieser, T., Bibb, M. J., Buttner, M. J., Chater, K. F., and Hopwood, D. A. 2000. Practical Streptomyces Genetics. Norwich, United Kindom.

Kim, Y.-S., Cho, J. M., Park, D. H., Lee, H.-G., Kim, J.-S., Seo, S.-T., Shin, K.-Y., Hur, J. H., and Lim, C. K. 1999. Production of Thaxtomin A by Korean Isolates of Streptomyces turgidiscabies and Their Involvement in Pathogenicity. Plant Pathol J. 15:168–171

Kinashi, H., Someno, K., and Sakaguchi, K. 1984. Isolation and characterization of concanamycins A, B and C. J. Antibiot. 37:1333–1343

King, R. R., and Calhoun, L. A. 2009. The thaxtomin phytotoxins: Sources, synthesis, biosynthesis, biotransformation and biological activity. Phytochemistry. 70:833–841

King, R. R., Lawrence, C. H., and Calhoun, L. A. 2002. Chemistry of phytotoxins associated with Streptomyces scabies the causal organism of potato common scab. J. Agric. Food Chem. 40:834–837

King, R. R., Lawrence, C. H., and Gray, J. A. 2001. Herbicidal properties of the thaxtomin group of phytotoxins. J. Agric. Food Chem. 49:2298–2301

King, R. R., Lawrence, C. H., Calhoun, L. A., and Ristaino, J. B. 1994. Isolation and Characterization of Thaxtomin-Type Phytotoxins Associated with Streptomyces ipomoeae. J. Agric. Food Chem. 42:1791–1794

King, R. R., Lawrence, C. H., Clark, M. C., and Calhoun, L. A. 1989. Isolation and characterization of phytotoxins associated with Streptomyces scabies. J. Chem. Soc. Chem. Commun. 0:849–850

134

Koh, Y. J., Kim, G. H., Lee, Y. S., Sohn, S. H., Koh, H. S., Kwon, S., Heu, S., and Jung, J. S. 2012. Pectobacterium carotovorum subsp. actinidiae subsp. nov., a new bacterial pathogen causing canker-like symptoms in yellow kiwifruit, Actinidia chinensis. N. Z. J. Crop Hortic. Sci. 40:269–279

Kreuze, J. F., Suomalainen, S., Paulin, L., and Valkonen, J. P. T. 1999. Phylogenetic Analysis of 16S rRNA Genes and PCR Analysis of the nec1 Gene from Streptomyces spp. Causing Common Scab, Pitted Scab, and Netted Scab in Finland. Phytopathology. 89:462–469

Kreuze, J. F., Suomalainen, S., Paulin, L., and Valkonen, J. P. T. 2007. Phylogenetic Analysis of 16S rRNA Genes and PCR Analysis of the nec1 Gene from Streptomyces spp. Causing Common Scab, Pitted Scab, and Netted Scab in Finland. Phytopathology. 89:462–469

Kritzman, G., Shani-Cahani, A., Kirshner, B., Riven, Y., Bar, Z., Katan, J., and Grinstein, A. 1996. Pod wart disease of peanuts. Phytoparasitica. 24:293–304

Kwasiborski, A., Mondy, S., Beury-Cirou, A., and Faure, D. 2013. Genome Sequence of the Pectobacterium atrosepticum Strain CFBP6276, Causing Blackleg and Soft Rot Diseases on Potato Plants and Tubers. Genome Announc. 1:e00374–13

Labeda, D. P. 2011. Multilocus sequence analysis of phytopathogenic species of the genus Streptomyces. Int. J. Syst. Evol. Microbiol. 61:2525–2531

Labeda, D. P. 2016. Taxonomic evaluation of putative Streptomyces scabiei strains held in the ARS Culture Collection (NRRL) using multi-locus sequence analysis. Antonie Van Leeuwenhoek. 109:349–356

Labeda, D. P., Goodfellow, M., Brown, R., Ward, A. C., Lanoot, B., Vanncanneyt, M., Swings, J., Kim, S. B., Liu, Z., Chun, J., Tamura, T., Oguchi, A., Kikuchi, T., Kikuchi, H., Nishii, T., Tsuji, K., Yamaguchi, Y., Tase, A., Takahashi, M., Sakane, T., Suzuki, K. I., and Hatano, K. 2012. Phylogenetic study of the species within the family Streptomycetaceae. Antonie Van Leeuwenhoek. 101:73–104

Lambert, D. H. 1991. First report of additional hosts for the acid scab pathogen Streptomyces acidiscabies. Plant Dis. 75:750

Lambert, D. H., and Loria, R. 1989. Streptomyces acidiscabies sp. nov.†. Int. J. Syst. Evol. Microbiol. 39:393–396

Lauzier, A., Simao-Beaunoir, A.-M., Bourassa, S., Poirier, G. G., Talbot, B., and Beaulieu, C. 2008. Effect of potato suberin on Streptomyces scabies proteome. Mol. Plant Pathol. 9:753– 762

Lawrence, C. H., Clark, M. C., and King, R. R. 1990. Induction of common scab symptoms in aseptically cultured potato tubers by the vivotoxin, thaxtomin. Phytopathology. :606–608

Lee, C. A., Babic, A., and Grossman, A. D. 2010. Autonomous plasmid-like replication of a conjugative transposon. Mol. Microbiol. 75:268–279

135

Lefébure, T., and Stanhope, M. J. 2009. Pervasive, genome-wide positive selection leading to functional divergence in the bacterial genus Campylobacter. Genome Res. 19:1224–1232

Lerat, E., Daubin, V., and Moran, N. A. 2003. From gene trees to organismal phylogeny in prokaryotes: the case of the gamma-Proteobacteria. PLoS Biol. 1:E19

Lerat, S., Simao-Beaunoir, A.-M., and Beaulieu, C. 2009. Genetic and physiological determinants of Streptomyces scabies pathogenicity. Mol. Plant Pathol. 10:579–585

Li, B., Shi, Y., Ibrahim, M., Liu, H., Shan, C., Wang, Y., Kube, M., Xie, G.-L., and Sun, G. 2012. Genome sequence of the rice pathogen Dickeya zeae strain ZJU1202. J Bacteriol. 194:4452–4453

Li, L., Stoeckert, C. J., and Roos, D. S. 2003. OrthoMCL: Identification of ortholog groups for eukaryotic genomes. Genome Res. 13:2178–2189

Li, X. S., Yuan, K. X., Cullis, J., Lévesque, C. A., Chen, W., Lewis, C. T., and De Boer, S. H. 2015a. Draft Genome Sequences for Canadian Isolates of Pectobacterium carotovorum subsp. brasiliense with Weak Virulence on Potato. Genome Announc. 3:e00240–15

Li, Y., Fang, W., Xue, H., Liang, W.-X., Wang, L.-F., Tian, G.-Z., Wang, X.-Z., Lin, C.-L., Li, X., and Piao, C.-G. 2015b. Brenneria populi sp. nov., isolated from symptomatic bark of Populus×euramericana canker. Int. J. Syst. Evol. Microbiol. 65:432–437

Lorenz, M. G., and Wackernagel, W. 1994. Bacterial Gene-Transfer by Natural Genetic- Transformation in the Environment. Microbiol. Rev. 58:563–602

Loria, R., Bignell, D. R. D., Moll, S., Huguet-Tapia, J. C., Joshi, M. V., Johnson, E. G., Seipke, R. F., and Gibson, D. M. 2008. Thaxtomin biosynthesis: the path to plant pathogenicity in the genus Streptomyces. Antonie Van Leeuwenhoek. 94:3–10

Loria, R., Bukhalid, R. A., Fry, B. A., and King, R. R. 1997. Plant pathogenicity in the genus Streptomyces. Plant Dis. 81:836–846

Loria, R., Coombs, J., Yoshida, M., Kers, J., and Bukhalid, R. 2003. A paucity of bacterial root diseases: Streptomyces succeeds where others fail. Physiol Mol Plant Path. 62:65–72

Loria, R., Kers, J., and Joshi, M. 2006. Evolution of Plant Pathogenicity in Streptomyces. Annu. Rev. Phytopathol. 44:469–487

Ma, B., Hibbing, M. E., Kim, H.-S., Reedy, R. M., Yedidia, I., Breuer, J., Breuer, J., Glasner, J. D., Perna, N. T., Kelman, A., and Charkowski, A. O. 2007. Host Range and Molecular Phylogenies of the Soft Rot Enterobacterial Genera Pectobacterium and Dickeya. Phytopathology. 97:1150–1163

MacNeil, D. J., Gewain, K. M., Ruby, C. L., Dezeny, G., Gibbons, P. H., and MacNeil, T. 1992. Analysis of Streptomyces avermitilis genes required for avermectin biosynthesis utilizing a novel integration vector. Gene. 111:61–68

136

Mann, S. G. 2009. The role of the twin arginine secretion system in the virulence of the plant pathogen Streptomyces scabies. Doctoral dissertation, University of East Anglia

Mansfield, J., Genin, S., Magori, S., Citovsky, V., Sriariyanum, M., Ronald, P., Dow, M., Verdier, V., Beer, S. V., Machado, M. A., Toth, I., Salmond, G., and Foster, G. D. 2012. Top 10 plant pathogenic bacteria in molecular plant pathology. Mol. Plant Pathol. 13:614–629

McCormick, J. R., and Flardh, K. 2012. Signals and regulators that govern Streptomyces development. FEMS Microbiol. Rev. 36:206–231

Meier-Kolthoff, J. P., Auch, A. F., Klenk, H.-P., and Goeker, M. 2013. Genome sequence-based species delimitation with confidence intervals and improved distance functions. BMC Bioinformatics. 14:60

Miyajima, K., Tanaka, F., Takeuchi, T., and Kuninaga, S. 1998. Streptomyces turgidiscabies sp. nov. Int. J. Syst. Bacteriol. 48:495–502

Nabhan, S., Boer, S. H., Maiss, E., and Wydra, K. 2012. Taxonomic relatedness between Pectobacterium carotovorum subsp. carotovorum, Pectobacterium carotovorum subsp. odoriferum and Pectobacterium carotovorum subsp. brasiliense subsp. nov. J. Appl. Microbiol. 113:904–913

Nabhan, S., De Boer, S. H., Maiss, E., and Wydra, K. 2013. Pectobacterium aroidearum sp. nov., a soft rot pathogen with preference for monocotyledonous plants. Int. J. Syst. Evol. Microbiol. 63:2520–2525

Natsume, M., Komiya, M., Koyanagi, F., Tashiro, N., Kawaide, H., and Abe, H. 2005. Phytotoxin produced by Streptomyces sp. causing potato russet scab in Japan. 71:364–369

Natsume, M., RYU, R., and Abe, H. 1996. Production of Phytotoxins, Concanamycins A and B by Streptomyces spp. Causing Potato Scab. JPN J Phytopathol. 62:411–413

Natsume, M., TAKI, M., Tashiro, N., and Abe, H. 2001. Phytotoxin Production and Aerial Mycelium Formation by Streptomyces scabies and S. acidiscabies in Vitro. J Gen Plant Pathol. 67:299–302

Natsume, M., Yamada, A., Tashiro, N., and Abe, H. 1998. Differential Production of the Phytotoxins Thaxtomin A and Concanamycins A and B by Potato Common Scab-causing Streptomyces spp. JPN J Phytopathol. 64:202–204

Nes, I. F., and Tagg, J. R. 1996. Novel lantibiotics and their pre-peptides. Antonie Van Leeuwenhoek. 69:89–97

Nikolaichik, Y., Gorshkov, V., Gogolev, Y., Valentovich, L., and Evtushenkov, A. 2014. Genome Sequence of Pectobacterium atrosepticum Strain 21A. Genome Announc. 2:e00935–14

137

Nykyri, J., Niemi, O., Koskinen, P., Nokso-Koivisto, J., Pasanen, M., Broberg, M., Plyusnin, I., Törönen, P., Holm, L., Pirhonen, M., and Palva, E. T. 2012. Revised Phylogeny and Novel Horizontally Acquired Virulence Determinants of the Model Soft Rot Phytopathogen Pectobacterium wasabiae SCC3193 B. Tyler, ed. PLoS Pathog. 8:e1003013

Osborn, A. M., and Boltner, D. 2002. When phage, plasmids, and transposons collide: genomic islands, and conjugative- and mobilizable-transposons as a mosaic continuum. Plasmid. 48:202–212

Panda, P., Fiers, M. W. E. J., Lu, A., Armstrong, K. F., and Pitman, A. R. 2015a. Draft Genome Sequences of Three Pectobacterium Strains Causing Blackleg of Potato: P. carotovorum subsp. brasiliensis ICMP 19477, P. atrosepticum ICMP 1526, and P. carotovorum subsp. carotovorum UGC32. Genome Announc. 3:e00874–15

Panda, P., Lu, A., Armstrong, K. F., and Pitman, A. R. 2015b. Draft Genome Sequence for ICMP 5702, the Type Strain of Pectobacterium carotovorum subsp. carotovorum That Causes Soft Rot Disease on Potato. Genome Announc. 3:e00875–15

Park, D. H., Kim, J.-S., Kwon, S. W., Wilson, C., Yu, Y. M., Hur, J. H., and Lim, C. K. 2003a. Streptomyces luridiscabiei sp. nov., Streptomyces puniciscabiei sp. nov. and Streptomyces niveiscabiei sp. nov., which cause potato common scab disease in Korea. Int. J. Syst. Evol. Microbiol. 53:2049–2054

Park, D. H., Yu, Y. M., Kim, J.-S., Cho, J. M., Hur, J. H., and Lim, C. K. 2003b. Characterization of Streptomycetes Causing Potato Common Scab in Korea. Plant Dis. 87:1290–1296

Park, T. H., Choi, B. S., Choi, A. Y., Choi, I. Y., and Heu, S. 2012. Genome sequence of Pectobacterium carotovorum subsp. carotovorum strain PCC21, a pathogen causing soft rot in Chinese cabbage. J Bacteriol. 194:6345

Pavlovic, G., Burrus, V., Gintz, B., Decaris, B., and Guedon, G. 2004. Evolution of genomic islands by deletion and tandem accretion by site-specific recombination: ICESt1-related elements from Streptococcus thermophilus. Microbiology. 150:759–774

Perombelon, M. C. M., and Kelman, A. 1980. Ecology of the soft rot erwinias. Annu Rev Phytopathol. 18:361–387

Pertry, I., Václavíková, K., Gemrotová, M., Spíchal, L., Galuszka, P., Depuydt, S., Temmerman, W., Stes, E., De Keyser, A., Riefler, M., Biondi, S., Novák, O., Schmülling, T., Strnad, M., Tarkowski, P., Holsters, M., and Vereecke, D. 2010. Rhodococcus fascians Impacts Plant Development Through the Dynamic Fas-Mediated Production of a Cytokinin Mix. Mol Plant Microbe Interact. 23:1164–1174

Pérombelon, M. C. M. 2002. Potato diseases caused by soft rot erwinias: an overview of pathogenesis. Plant Pathol. 51:1–12

138

Pritchard, L., Humphris, S., Baeyen, S., Maes, M., Van Vaerenbergh, J., Elphinstone, J., Saddler, G., and Toth, I. 2013a. Draft Genome Sequences of Four Dickeya dianthicola and Four Dickeya solani Strains. Genome Announc. 1:e00087–12

Pritchard, L., Humphris, S., Saddler, G. S., Elphinstone, J. G., Pirhonen, M., and Toth, I. K. 2013b. Draft Genome Sequences of 17 Isolates of the Plant Pathogenic Bacterium Dickeya. Genome Announc. 1:e00978–13

Puymege, A., Bertin, S., Chuzeville, S., Guedon, G., and Payot, S. 2013. Conjugative transfer and cis-mobilization of a genomic island by an integrative and conjugative element of Streptococcus agalactiae. J Bacteriol. 195:1142–1151

Rajeev, L., Malanowska, K., and Gardner, J. F. 2009. Challenging a paradigm: the role of DNA homology in tyrosine recombinase reactions. Microbiol Mol Biol Rev. 73:300–309

Raoul des Essarts, Y., Mondy, S., Hélias, V., and Faure, D. 2015. Genome Sequence of the Potato Plant Pathogen Dickeya dianthicola Strain RNS04.9. Genome Announc. 3:e00581–15

Richter, M., and Rosselló-Mora, R. 2009. Shifting the genomic gold standard for the prokaryotic species definition. Proc. Natl. Acad. Sci. U.S.A. 106:19126–19131

Rong, X., Liu, N., Ruan, J., and Huang, Y. 2010. Multilocus sequence analysis of Streptomyces griseus isolates delineating intraspecific diversity in terms of both taxonomy and biosynthetic potential. Antonie Van Leeuwenhoek. 98:237–248

Roshan, U., and Livesay, D. R. 2006. Probalign: multiple sequence alignment using partition function posterior probabilities. Bioinformatics. 22:2715–2721

Rosselló-Mora, R. 2006. DNA-DNA Reassociation Methods Applied to Microbial Taxonomy and Their Critical Evaluation. Pages 23–50 in: Springer Berlin Heidelberg, Berlin, Heidelberg.

Rosselló-Mora, R., and Amann, R. 2001. The species concept for prokaryotes. FEMS Microbiol. Rev. 25:39–67

Salichos, L., and Rokas, A. 2013. Inferring ancient divergences requires genes with strong phylogenetic signals. Nature. 497:327–331

Salyers, A. A., Shoemaker, N. B., Stevens, A. M., and LI, L. Y. 1995. Conjugative Transposons - an Unusual and Diverse Set of Integrated Gen-Transfer Elements. Microbiol. Rev. 59:579– 590

Samson, R., Legendre, J. B., Christen, R., Saux, M. F.-L., Achouak, W., and Gardan, L. 2005. Transfer of Pectobacterium chrysanthemi (Burkholder et al. 1953) Brenner et al. 1973 and Brenneria paradisiaca to the genus Dickeya gen. nov. as Dickeya chrysanthemi comb. nov. and Dickeya paradisiaca comb. nov. and delineation of four novel species, Dickeya dadantii sp. nov., Dickeya dianthicola sp. nov., Dickeya dieffenbachiae sp. nov. and Dickeya zeae sp. nov. Int. J. Syst. Evol. Microbiol. 55:1415–1427

139

Sebaihia, M., Bocsanczy, A. M., Biehl, B. S., Quail, M. A., Perna, N. T., Glasner, J. D., DeClerck, G. A., Cartinhour, S., Schneider, D. J., Bentley, S. D., Parkhill, J., and Beer, S. V. 2010. Complete genome sequence of the plant pathogen Erwinia amylovora strain ATCC 49946. J Bacteriol. 192:2020–2021

Seemann, T. 2014. Prokka: rapid prokaryotic genome annotation. Bioinformatics. 30:2068–2069

Seipke, R. F., and Loria, R. 2008. Streptomyces scabies 87-22 possesses a functional tomatinase. J Bacteriol. 190:7684–7692

Seipke, R. F., Kaltenpoth, M., and Hutchings, M. I. 2012. Streptomyces as symbionts: an emerging and widespread theme? FEMS Microbiol. Rev. 36:862–876

Sentausa, E., and Fournier, P.-E. 2013. Advantages and limitations of genomics in prokaryotic taxonomy. Clin. Microbiol. Infect. 19:790–795

Sitkiewicz, I., Green, N. M., Guo, N., Mereghetti, L., and Musser, J. M. 2011. Lateral gene transfer of streptococcal ICE element RD2 (region of difference 2) encoding secreted proteins. BMC Microbiol. 11:65

Smith, M. C. M., and Thorpe, H. M. 2002. Diversity in the serine recombinases. Mol. Microbiol. 44:299–307

Song, J., Lee, S.-C., Kang, J.-W., Baek, H.-J., and Suh, J.-W. 2004. Phylogenetic analysis of Streptomyces spp. isolated from potato scab lesions in Korea on the basis of 16S rRNA gene and 16S–23S rDNA internally transcribed spacer sequences. Int. J. Syst. Evol. Microbiol. 54:203–209

Stamatakis, A. 2014. RAxML version 8: a tool for phylogenetic analysis and post-analysis of large phylogenies. Bioinformatics. 30:1312–1313

Stamatakis, A., Hoover, P., and Rougemont, J. 2008. A Rapid Bootstrap Algorithm for the RAxML Web Servers. Systematic Biol. 57:758–771

Stegmann, E., Rausch, C., Stockert, S., Burkert, D., and Wohlleben, W. 2006. The small MbtH- like protein encoded by an internal gene of the balhimycin biosynthetic gene cluster is not required for glycopeptide production. FEMS Microbiol. Lett. 262:85–92

Stes, E., Francis, I., Pertry, I., Dolzblasz, A., Depuydt, S., and Vereecke, D. 2013. The leafy gall syndrome induced by Rhodococcus fascians. FEMS Microbiol. Lett. 342:187–194

Studholme, D. J. 2016. Genome Update. Let the consumer beware: Streptomyces genome sequence quality. Microb Biotechnol. 9:3–7

Suzui, T., Miyashita, K., and Tashiro, N. 1988. Streptomyces cheloniumii sp. nov., a new species causing russet scab of potato. Pages p 177, Kyoto, Japan.

140

Suzuki, H., Lefébure, T., Bitar, P. P., and Stanhope, M. J. 2012. Comparative genomic analysis of the genus Staphylococcus including Staphylococcus aureus and its newly described sister species Staphylococcus simiae. BMC Genomics. 13:38

Sławiak, M., van Beckhoven, J. R. C. M., Speksnijder, A. G. C. L., Czajkowski, R., Grabe, G., and van der Wolf, J. M. 2009. Biochemical and genetical analysis reveal a new clade of biovar 3 Dickeya spp. strains isolated from potato in Europe. Eur J Plant Pathol. 125:245– 261

Takeuchi, T., Sawada, H., Tanaka, F., and Matsuda, I. 1996. Phylogenetic analysis of Streptomyces spp. causing potato scab based on 16S rRNA sequences. Int. J. Syst. Bacteriol. 46:476–479

Tamura, K., Stecher, G., Peterson, D., Filipski, A., and Kumar, S. 2013. MEGA6: Molecular Evolutionary Genetics Analysis version 6.0. Mol. Biol. Evol. 30:2725–2729

Tegg, R. S., Melian, L., Wilson, C. R., and Shabala, S. 2005. Plant cell growth and ion flux responses to the streptomycete phytotoxin thaxtomin A: Calcium and hydrogen flux patterns revealed by the non-invasive MIFE technique. Plant and Cell Physiology. 46:638–648

Thwaites, R., Wale, S. J., Nelson, D., Munday, D., and Elphinstone, J. G. 2010. Streptomyces turgidiscabies and S. acidiscabies: two new causal agents of common scab of potato (Solanum tuberosum) in the UK. Plant Pathol. 59:804–804

Tomihama, T., Nishi, Y., Sakai, M., Ikenaga, M., Okubo, T., and Ikeda, S. 2016. Draft Genome Sequences of Streptomyces scabiei S58, Streptomyces turgidiscabies T45, and Streptomyces acidiscabies a10, the Pathogens of Potato Common Scab, Isolated in Japan. Genome Announc. 4:e00062–16

Toth, I. K., and Birch, P. 2005. Rotting softly and stealthily. Curr Opin Plant Biol. 8:424–429

Toth, I. K., Pritchard, L., and Birch, P. R. J. 2006. Comparative genomics reveals what makes an enterobacterial plant pathogen. Annu Rev Phytopathol. 44:305–336

Toth, I., Humphris, S., Campbell, E., and Pritchard, L. 2015. Why Genomics Research on Pectobacterium and Dickeya Makes a Difference. Am. J. Potato Res. 92:218–222

Townsend, J. P. 2007. Profiling phylogenetic informativeness. Systematic Biol. 56:222–231

Tóth, L., Maeda, M., Tanaka, F., and al, E. 2005. Isolation and identification of pathogenic strains of Streptomyces acidiscabies from netted scab lesions of potato tubers in Hokkaido (Japan). Acta Microbiologica et Immunologica Hungarica. 48:575–585

141

van der Wolf, J. M., Nijhuis, E. H., Kowalewska, M. J., Saddler, G. S., Parkinson, N., Elphinstone, J. G., Pritchard, L., Toth, I. K., Lojkowska, E., Potrykus, M., Waleron, M., De Vos, P., Cleenwerck, I., Pirhonen, M., Garlant, L., Hélias, V., Pothier, J. F., Pflüger, V., Duffy, B., Tsror, L., and Manulis, S. 2014. Dickeya solani sp. nov., a pectinolytic plant- pathogenic bacterium isolated from potato (Solanum tuberosum). Int. J. Syst. Evol. Microbiol. 64:768–774

VanEtten, H. D., Mansfield, J. W., Bailey, J. A., and Farmer, E. E. 1994. Two Classes of Plant Antibiotics: Phytoalexins versus “Phytoanticipins.” The Plant Cell. 6:1191–1192

Ventura, M., Canchaya, C., Tauch, A., Chandra, G., Fitzgerald, G. F., Chater, K. F., and van Sinderen, D. 2007. Genomics of Actinobacteria: tracing the evolutionary history of an ancient phylum. Microbiol Mol Biol Rev. 71:495–548

Wach, M. J., Krasnoff, S. B., Loria, R., and Gibson, D. M. 2007. Effect of carbohydrates on the production of thaxtomin A by Streptomyces acidiscabies. Arch Microbiol. 188:81–88

Waldee, E. L. 1942. Comparative studies of some peritrichous phytopathogenic bacteria. Iowa State Coll. J. Sci.19:435-484.

Wanner, L. A. 2009. A Patchwork of Streptomyces Species Isolated from Potato Common Scab Lesions in North America. Am. J. Potato Res. 86:247–264

Willey, J. M., and van der Donk, W. A. 2007. Lantibiotics: peptides of diverse structure and function. Annu. Rev. Microbiol. 61:477–501

Wozniak, R. A. F., and Waldor, M. K. 2010. Integrative and conjugative elements: mosaic mobile genetic elements enabling dynamic lateral gene flow. Nat Rev Microbiol. 8:552–563

Xin, X.-F., and He, S. Y. 2013. Pseudomonas syringae pv. tomato DC3000: A Model Pathogen for Probing Disease Susceptibility and Hormone Signaling in Plants. Annu Rev Phytopathol. 51:473–498

Yuan, K. X., Adam, Z., Tambong, J., Lévesque, C. A., Chen, W., Lewis, C. T., De Boer, S. H., and Li, X. S. 2014. Draft Genome Sequence of Pectobacterium wasabiae Strain CFIA1002. Genome Announc. 2:e00214–14

Zhang, J.-X., Lin, B.-R., Shen, H.-F., and Pu, X.-M. 2013. Genome Sequence of the Banana Pathogen Dickeya zeae Strain MS1, Which Causes Bacterial Soft Rot. Genome Announc. 1:e00317–13

Zhang, Y., and Qiu, S. 2015. Examining phylogenetic relationships of Erwinia and Pantoea species using whole genome sequence data. Antonie Van Leeuwenhoek. 108:1037–1046

Zhang, Y., and Qiu, S. 2016. Phylogenomic analysis of the genus Ralstonia based on 686 single- copy genes. Antonie Van Leeuwenhoek. 109:71–82

142

Zhao, W. Q., Liu, D. Q., and Yu, X. M. 2008. First Report of Potato Scab Caused by Streptomyces turgidiscabies in China. Plant Dis. 92:1587–1587

Zhao, W. Q., Yu, X. M., and Liu, D. Q. 2010. First report of Streptomyces acidiscabies causing potato scab in China. Plant Pathol. 59:405–405

Zhou, J., Cheng, Y., Lv, M., Liao, L., Chen, Y., Gu, Y., Liu, S., Jiang, Z., Xiong, Y., and Zhang, L. 2015. The complete genome sequence of Dickeya zeae EC1 reveals substantial divergence from other Dickeya strains and species. BMC Genomics. 16:571

143

BIOGRAPHICAL SKETCH

Yucheng Zhang was born in Penglai, Shandong province, China in 1986. In 2005,

Yucheng Zhang attended Northwest A&F University, and majored in viticulture and enology. He obtained the bachelor’s degree in 2009, and then he was enrolled in a master’s program from the same university to study grape defense mechanisms to abiotic and biotic stresses. He earned the master’s degree in 2012, and after that, he was funded by a Graduate School Fellowship from

University of Florida to pursue a doctorate degree in the Department of Plant Pathology. He worked with Dr. Rosemary Loria to study pathogenicity islands and evolution of Streptomyces.

144