Identification of Erwinia amylovora Genes Induced during ... · (Pyrus communis L. cv. ‘Bartlett’) were surface sterilized with 10% bleach and pricked with a sterile needle as
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Identification of Erwinia amylovora Genes Induced duringInfection of Immature Pear Tissue
Youfu Zhao,† Sara E. Blumer,† and George W. Sundin*Department of Plant Pathology and Center for Microbial Pathogenesis, Michigan State University,
103 Center for Integrated Plant Systems, East Lansing, Michigan 48824
Received 14 July 2005/Accepted 16 September 2005
The enterobacterium Erwinia amylovora is a devastating plant pathogen causing necrotrophic fire blightdisease of apple, pear, and other rosaceous plants. In this study, we used a modified in vivo expressiontechnology system to identify E. amylovora genes that are activated during infection of immature pear tissue,a process that requires the major pathogenicity factors of this organism. We identified 394 unique pearfruit-induced (pfi) genes on the basis of sequence similarity to known genes and separated them into nineputative function groups including host-microbe interactions (3.8%), stress response (5.3%), regulation(11.9%), cell surface (8.9%), transport (13.5%), mobile elements (1.0%), metabolism (20.3%), nutrient acqui-sition and synthesis (15.5%), and unknown or hypothetical proteins (19.8%). Known virulence genes, includinghrp/hrc components of the type III secretion system, the major effector gene dspE, type II secretion, levansu-crase (lsc), and regulators of levansucrase and amylovoran biosynthesis, were upregulated during pear tissueinfection. Known virulence factors previously identified in E. (Pectobacterium) carotovora and Pseudomonassyringae were identified for the first time in E. amylovora and included HecA hemagglutinin family adhesion,Peh polygalacturonase, new effector HopPtoCEA, and membrane-bound lytic murein transglycosylase MltEEA.An insertional mutation within hopPtoCEA did not result in reduced virulence; however, an mltEEA knockoutmutant was reduced in virulence and growth in immature pears. This study suggests that E. amylovora utilizesa variety of strategies during plant infection and to overcome the stressful and poor nutritional environmentof its plant hosts.
Erwinia amylovora is the causative agent of fire blight, adevastating necrotic disease affecting apple, pear, and otherrosaceous plants. Entry of the bacterium into plants can occurvia flower blossoms or actively growing young shoots orthrough wounds. Upon entry, the fire blight pathogen movesthrough intercellular spaces towards the xylem and also thecortical parenchyma (70). Symptoms often appear as water-soaked tissue that rapidly wilts and becomes necrotic, leadingto the characteristic “shepherd’s crook.” As a member of theEnterobacteriaceae, E. amylovora is related to many importanthuman and animal pathogens such as Escherichia coli, Yersiniapestis, Yersinia enterocolitica, Salmonella enterica, and Shigellaflexneri.
Like many other gram-negative plant-pathogenic bacteria,E. amylovora produces a type III Hrp secretion system (TTSS)apparatus that delivers effector proteins into host plants (40).The TTSS in E. amylovora controls the ability of E. amylovorato cause disease in susceptible host plants and to elicit thehypersensitive response in resistant and nonhost plants. Mosthrp genes have been found to encode proteins involved in generegulation or in assembly of the TTSS apparatus (3, 31, 40).
The TTSS of E. amylovora secretes several virulence pro-teins, including HrpA, HrpN, HrpW, and disease-specific pro-tein DspA/E (hereafter referred to as DspE) (13, 14, 28, 40, 41,
74, 75). The HrpA protein is the major structural protein of apilus called the Hrp pilus, which is the extracellular part of theTTSS (37). DspE, HrpN, and HrpW proteins are effector pro-teins of the TTSS and are believed to be injected directly intohost cells (13, 14).
Additional E. amylovora virulence factors that contribute topathogenesis and plant colonization include the exopolysac-charides amylovoran and levan, iron-scavenging siderophoredesferrioxamine, metalloprotease PrtA, multidrug efflux pumpAcrAB, and carbohydrate metabolism genes specifically in-volved in the utilization of sorbitol, sucrose, and galactose (1,15, 17, 51, 80). Transcriptional regulators of the amylovoranand levan biosynthetic operons have also been identified (11,19, 79) and are required for the expression of the biosyntheticmachinery for the exopolysaccharides (10, 22, 39, 79). E. amy-lovora pathogenesis is also subject to global regulation by thesmall regulatory RNA rsmB, which functions by titrating andcountering the activity of the repressor protein RsmA; thissystem is reported to positively regulate exopolysaccharideproduction, motility, and pathogenicity (46). In addition, E.amylovora strains contain a ubiquitous nonconjugative plasmidof 28 to 30 kb designated pEA29; laboratory-derived plasmid-cured strains exhibit a reduction in virulence (49). pEA29encodes several potential virulence genes including a thiamine-biosynthetic operon that is proposed to influence amylovoranproduction (49).
Genetic analysis of virulence genes in E. amylovora has beenperformed mostly through the production and screening ofmutants. Additionally, most of the genes discovered so far havebeen identified from mutant screening under controlled con-
* Corresponding author. Mailing address: Department of Plant Pa-thology and Center for Microbial Pathogenesis, Michigan State Uni-versity, 103 Center for Integrated Plant Systems, East Lansing, MI48824. Phone: (517) 355-4573. Fax: (517) 353-5598. E-mail: [email protected].
† Y. F. Zhao and S. E. Blumer contributed equally to this work.
ditions. However, it is not feasible to mimic all of the nutrientand defense conditions in vitro to characterize all the genesfrom E. amylovora required for infection and colonization ofplants. There is a need, then, for a high-throughput method ofscreening for genes that are involved in virulence and growth inplanta of E. amylovora.
In the last decade, many gene expression technologies in-cluding in vivo expression technology (IVET) have been de-veloped to identify gene expression profiles of organisms dur-ing interactions with various host environments (5, 33, 47).IVET screening theoretically scans the entire genome and,through the use of appropriate environmental conditions anddifferent strategies, can yield large numbers of potentially im-portant genes (59). IVET screens have identified genes up-regulated upon infection with enteric human and animalpathogens such as Salmonella enterica, Shigella flexneri, and Y.enterocolitica (7, 47, 55, 78). IVET systems have also been usedto identify genes expressed during plant infection by Xan-thomonas campestris, Erwinia chrysanthemi, Pseudomonassyringae, and Ralstonia solanacearum (12, 16, 30, 56, 57, 77);phyllosphere colonization by P. syringae (48); rhizosphere col-onization by Pseudomonas putida (60); and saprophytic colo-nization by Pseudomonas fluorescens (58, 67).
Like many plant-pathogenic bacteria, E. amylovora can in-fect different host tissues at different stages of disease devel-opment. E. amylovora infects not only blossoms, leaves, andsucculent shoots but also immature fruits of susceptible hosts.The bacterium also grows epiphytically on stigmas and endo-phytically inside plant tissue. The maintenance of large num-bers of apple trees for study of E. amylovora pathogenesis isquite difficult due to the extensive greenhouse and growthchamber space required. As an alternative, many researchershave utilized immature pear fruits to study E. amylovora infec-tion (10, 14, 29). Immature pear infection is initiated througha wound inoculation; wound colonization is a frequently uti-lized mechanism of E. amylovora infection in nature (70). Im-mature pear assays, using either intact pear fruits or pear slices,have been used successfully to analyze virulence effects ofseveral E. amylovora genes (14, 29, 40).
Although key virulence factors contributing to fire blighthave been identified, little knowledge is available on the globalhost-regulated genes of E. amylovora during infection. To gaina better understanding of the molecular mechanism governingE. amylovora-host plant interactions, we undertook a compre-hensive genome-wide examination of gene expression patternsduring host infection to uncover pathogenesis strategies of theorganism. This IVET screen will also lay the groundwork forfuture studies examining the expression and function of criticalvirulence genes during infection of different host tissues andsurvival within the host. Several known virulence and patho-genesis factors were identified using this modified IVETscreen, along with new potential virulence genes that werepreviously described only in other bacterial pathosystems. Wealso confirmed that infection of immature pear tissue by E.amylovora required the major pathogenicity factors of the bac-terium.
MATERIALS AND METHODS
Bacterial strains and plasmids. The bacterial strains and plasmids utilized inthis study are listed in Table 1. Erwinia amylovora wild type (WT) and mutant
strains and E. coli strains were grown in Luria-Bertani (LB) medium at 28°C and37°C, respectively. Antibiotics were added to the culture medium at the indicatedconcentrations: rifampin, 100 �g/ml; kanamycin (Km), 30 �g/ml; gentamicin, 10�g/ml; and ampicillin (Ap), 100 �g/ml. Oligonucleotide primers used for PCRand sequencing in this study are also listed in Table 1.
DNA manipulation and sequence analysis. Plasmid DNA purification, PCRamplification of genes, isolation of fragments from agarose gels, restrictionenzyme digestion, T4 DNA ligation, and Southern hybridization were performedusing standard molecular procedures (63). Chromosomal DNA was isolatedusing a genomic DNA purification kit (QIAGEN, Valencia, CA). Thermal asym-metric interlaced PCR (TAIL-PCR) was performed using the degenerate prim-ers AD1, AD2, and AD3 as described previously (44), and PCR products fromsecondary and tertiary nested PCR were used for sequencing. DNA sequencingwas performed at the Genomic Technology Support Facility at Michigan StateUniversity. The oligonucleotide primer Aj1585, corresponding to the 5� end ofthe uidA gene, was used for sequencing fragment inserts cloned into the pGCM0plasmid. Sequence management and contig assembly were conducted usingDNAStar software (DNAStar Inc., Madison, WI). Database searches were con-ducted using the BLAST programs at NCBI (www.ncbi.nlm.nih.gov/BLAST).Percent similarity was also calculated using the BLAST program (4). Amino acidalignments were done with ClustalW, v. 1.83 (European Bioinformatics Institute,Cambridge, United Kingdom).
Immature pear infection assays. Immature pears are routinely used to exam-ine the pathogenicity of naturally occurring isolates or bacterial mutants of E.amylovora (14). In order to confirm that infection of immature pear requiredmajor pathogenicity factors as previously reported (14), we inoculated woundedimmature pear fruits with E. amylovora M52 (CFBP1430 dspE) and Ea110 hrpAmutants and monitored them for symptom development and in planta bacterialgrowth. Bacterial suspensions of all strains were grown overnight in LB broth,harvested by centrifugation, and resuspended in 0.5� sterile phosphate-bufferedsaline (PBS) with the cells adjusted to approximately 1 � 104 CFU/�l (opticaldensity at 600 nm of 0.1 and then dilution 100 times) in PBS. Immature pears(Pyrus communis L. cv. ‘Bartlett’) were surface sterilized with 10% bleach andpricked with a sterile needle as described previously (49). Wounded pears wereinoculated with 2 �l of cell suspensions and incubated in a humidified chamberat 28°C. Symptoms were recorded at 2, 4, 6, and 8 days postinoculation. Forbacterial population studies, the pear tissue surrounding the inoculation site wasexcised by using a no. 4 cork borer as described previously (14) and homogenizedin 0.5 ml of 0.5� PBS. Bacterial growth within the pear tissue was monitored bydilution plating of the ground material on LB medium amended with the ap-propriate antibiotics. For each strain tested, fruits were assayed in triplicate, andeach experiment was repeated two to three times.
Construction of the genomic library of transcriptional fusions to uidA. Weused E. amylovora Ea110� (cured of the ubiquitous plasmid pEA29) as thesource of chromosomal DNA for the IVET experiments. We excluded pEA29genes from this study because an analysis of the expression of pEA29-carriedgenes during infection will be presented in a separate report (G. McGhee and G.Sundin, unpublished). To create a library of transcriptional fusions, chromo-somal DNA from E. amylovora Ea110� was partially digested with HaeIII, andfragments between 800 bp and 2 kb in length were separated by electrophoresisand gel purified. The purified fragments were ligated into pGCM0 prepared bySmaI digestion and transformed into WT E. amylovora Ea110 (containingpEA29) by electroporation. The use of WT strain Ea110 was necessary becausethe ubiquitous pEA29 plasmid contributes to E. amylovora virulence (49).
The 6.2-kb pGCM0 reporter vector was constructed by cloning the aacC1 gene(conferring resistance to gentamicin) into the EcoRI and KpnI sites and thepromoterless uidA (�-glucuronidase [GUS]) reporter gene into PstI and HindIIIsites of pGem3zf through multiple cloning steps (Fig. 1A). The aacC1 gene wasamplified from plasmid pX1918GT by PCR using the primer pair Aj1390 andAj1391, whereas the promoterless uidA gene was amplified from plasmidpCAM140 using the primer pair Aj1388 and Aj1389. A transcriptional termina-tor sequence, also from pX1918G, was located immediately downstream of theaacC1 gene, and we added translational termination codons in all three readingframes upstream of the uidA gene. The pGCM0 vector was first digested withSmaI, and the ends were dephosphorylated with calf intestinal alkaline phospha-tase and checked for self-ligation before ligation with E. amylovora chromosomalDNA fragments. After ligation, DNA was introduced into Ea110 by electropo-ration, transformants growing on LB medium amended with gentamicin andampicillin were randomly collected, and plasmids were recovered. The random-ness of the inserts in the IVET collection was confirmed by checking insert sizefrom 30 random colonies through restriction digestion and PCR (data notshown).
VOL. 187, 2005 IVET SCREENING OF THE FIRE BLIGHT PATHOGEN 8089
As a control, we cloned a 570-bp fragment containing the dspE promoter intopGCM0. The fragment was amplified by PCR from strain Ea110 using the primerpair DspE1 and DspE2. The resulting 570-bp product was cleaved with SmaI andligated into pGCM0 in both orientations. The resulting plasmids were designatedpZYF2 (dspErev::pGCM0, dspE promoter in orientation opposite uidA) andpZYF8 (dspEfor::pGCM0, dspE promoter in correct orientation to uidA), respec-tively, and each plasmid was introduced into strain Ea110 by electroporation.
Screening of the E. amylovora IVET library using a GUS-based microtiterplate assay. An in vivo microtiter plate assay was developed for screening of theE. amylovora IVET library (Fig. 1B). Briefly, approximately 19,200 transformantsin strain Ea110 were randomly collected and initially screened for GUS activity
on LB plates containing 5-bromo-4-chloro-3-indolyl-�-D-glucuronide (X-Gluc).After incubation at 25°C for 48 h, bacteria were transferred individually using a48-pin colony transfer apparatus and inoculated onto immature pear disks (3mm) in 96-well microtiter plates. Intact pears were surface sterilized using 10%bleach for 10 min and rinsed three times with sterile water. Disks were cut frompears using a no. 2 cork borer and immediately immersed into microtiter platewells containing 25 �l 0.5� PBS buffer to avoid oxidation. The microtiter plateswere then covered with AirPore tape (QIAGEN, Valencia, CA) after inoculationand incubated in a humidity chamber at 25°C for 48 h. After incubation, aqualitative GUS assay was performed as described below. Transformants show-ing GUS activity on pear tissue but not on LB plates were selected and re-
TABLE 1. Bacterial strains, plasmids, and primers used in this study
Strain, plasmid, or primer Relevant characteristic(s) or sequence(s) (5�-3�)a Referenceor source
StrainsErwinia amylovora
Ea110 Wild type, isolated from apple 49Ea110� Ea110, cured of pEA29 49Ea1189 Wild type, isolated from apple 17CFBP1430 Wild type, isolated from Crataegus 28M52 (Ea dspA) CFBP1430, dspA::uidA-Km, Kmr 28Ea110 hrpA Ea110 �hrpA Kmr 37ZYC1-3 (Ea hopPtoCEA) hopPtoC::Km; partial deletion and Kmr insertional mutant of hopPtoCEA of Ea1189; Kmr This studyZYE3-11 (Ea mltEEA) mltE::Km; partial deletion and Kmr insertional mutant of mltEEA of Ea1189; Kmr This study
S17-1 recA pro hsdR RP4-2-Tc::Mu-Km::Tn7 76S17-1 pir � pir lysogen of S17-1 76
PlasmidspBluescript II SK(�) Apr; cloning vector StratagenepGem3zf� Apr; cloning vector PromegapCAM140 Smr Spr Apr; R6K origin; mTn5SSgusA40 76pCAM140-MCSb Apr; R6K origin; pCAM140 derivative without mini-Tn5; contains the multiple cloning site of
pBluescript II SK(�)17
pX1918GT xylE-Gmr fusion cassette-containing plasmid flanked by inverted repeats of the pUC19 MCS 66pBSL15 Km cassette flanked by inverted repeats of the pUC18 MCS 2pGCM0 Gmr cassette with downstream transcriptional terminator and gusA with upstream
translational stop codons in pGem3zf�This study
pZYF2 570-bp dspE promoter in opposite orientation relative to uidA in pGCM0 This studypZYF8 570-bp dspE promoter in correct orientation relative to uidA in pGCM0 This study
a Kmr, Apr, Gmr, Spr, and Smr, kanamycin, ampicillin, gentamicin, spectinomycin, and streptomycin resistance, respectively. Underlined nucleotides are restrictionsites added, and the restriction enzymes are indicated at the ends of primers. Mixed nucleotides: S, C � G; W, A � T; N, A � T � C � G.
screened on LB plates containing X-Gluc and reinoculated onto pear disks in96-well microtiter plates. Confirmed differentially expressing transformants wereagain selected and stored at �70°C in glycerol stocks for further analysis. Plasmidswere isolated from the consistent differentially expressed transformants and wereend sequenced to identify the genes or promoter regions. Transformants showingGUS activities on both LB plates and pear tissues were assumed to contain consti-tutively expressed fusions and were not analyzed further in this study.
Construction of hopPtoCEA and mltEEA mutants. For the construction ofhopPtoCEA and mltEEA mutants, the sequences of the putative open reading
frames defined by the corresponding clones were determined and used to designprimers to amplify fragments of the gene and its upstream and downstreamsequences. Primer pairs PtoC1-PtoC2 and PtoC3-PtoC4 were used to amplify590-bp and 670-bp fragments from E. amylovora strain Ea1189 corresponding tothe upstream and downstream sequences of the hopPtoCEA gene, respectively.Primer pairs MltE1-MltE2 and MltE3-MltE4 were used to amplify 700-bp and560-bp fragments from E. amylovora strain Ea1189 corresponding to the up-stream and downstream sequences of the mltEEA gene, respectively. The twofragments for each open reading frame were cloned into pBluescript-II SK(�)through multiple cloning steps with corresponding restriction enzyme digestion(SacII-XbaI and EcoRI-XhoI, respectively). The whole fragment was excisedusing SacII and XhoI, gel purified, and cloned into the suicide vector pCAM-MCS (17) digested with the same enzymes. The resulting plasmids were digestedwith SmaI and ligated with a 1.2-kb fragment of the aph gene (conferringkanamycin resistance) released from plasmid pBSL15. The final plasmids weredesignated pZYC8 and pZYE8, respectively, and introduced into E. amylovorastrain Ea1189 by electroporation. Transconjugants resistant to Km were selected.To further exclude mutants resulting from single-crossover events, transformantswere selected on LB plates supplemented with Km and selected onto LB plateswith Ap. Km-resistant and Ap-sensitive colonies were selected, and their geno-types were confirmed by hybridization or PCR analysis.
GUS assays. The GUS reporter gene (uidA) on pGCM0 was used to monitorpromoter activity of IVET clones both in vitro and in vivo. Qualitative GUSactivity of IVET clones was monitored visually by the development of a bluecolor within 48 h of cells on LB medium containing 1 mM X-Gluc. QualitativeGUS activity of IVET clones grown on pear slices in microtiter plates after 48 hat 25°C was also monitored visually by adding 10 �l of 20 mM X-Gluc into thewells followed by incubation for 30 min at 37°C. The development of a blue colorindicated GUS activity.
To monitor the expression of IVET clones in pear tissue, quantitative GUSactivity of bacteria in either culture or pear tissue was determined as describedpreviously (14, 36) using 4-methylumbelliferyl-�-D-glucuronide as a substrateand 0.2 M Na2CO3 as stop buffer. Briefly, E. amylovora strains containing theIVET clones were grown on LB medium, resuspended in 0.5� PBS, and inoc-ulated in immature pear fruits as described above. At 0, 24, and 48 h postinocu-lation, the pear tissue surrounding the inoculation site was excised using a no. 4cork borer and homogenized in 0.5 ml 0.5� PBS. Forty microliters of homoge-nate was mixed with 160 �l of GUS extraction buffer. Reactions were stopped byNa2CO3 addition, and fluorescence was measured using a SAFIRE fluorometer(TECAN Boston, Medford, MA). Bacterial cell numbers in the sample wereestimated by dilution plating, and GUS activity (�mol of 4-methylumbelliferoneproduced per min) was normalized per 109 CFU (14). Three replicate fruits foreach strain were tested, and the experiment was repeated.
Nucleotide sequence accession numbers. Nucleotide sequence data reportedfor the hopPtoCEA and mltEEA genes were deposited in the GenBank databaseunder the accession no. AY887538 and AY887539.
RESULTS
Development of an IVET system and identification ofE. amylovora upregulated genes during immature pear infec-tion. Our immature pear infection results clearly demonstratedthat infection of immature pear by E. amylovora required ma-jor pathogenicity factors (Fig. 2). At 48 h after inoculation,E. amylovora strains CFBP1430 and Ea110 produced water-soaking symptoms in pears with visible bacterial ooze (data notshown). Two different strains were used in this initial experi-ment because of the availability of mutants of these strains.This work also confirmed that growth and symptom develop-ment of WT strains CFBP1430 and Ea110 were similar duringimmature pear inoculation (Fig. 2). Four days after inocula-tion, inoculated immature pears showed necrotic lesions andbacterial ooze formation (Fig. 2A). After 8 days, the entirepears showed necrosis, turning black with copious ooze pro-duction at the inoculation site (Fig. 2A). In contrast, diseasesymptoms were not observed on immature pears inoculatedwith either the E. amylovora hrpA or dspE mutant (Fig. 2A).Disease symptoms caused by WT strains on immature pear
FIG. 1. Overview of the IVET screen for E. amylovora genes in-duced during infection of immature pear disks. (A) Schematic map ofthe IVET vector pGCM0. The 6.2-kb pGCM0 vector was constructedby cloning the aacC1 gene (conferring resistance to gentamicin) intothe EcoRI and KpnI sites and the promoterless uidA (�-glucuroni-dase) reporter gene into PstI and HindIII sites of pGem3zf throughmultiple cloning steps. The symbol represents translational termi-nator codons in all three reading frames upstream of the uidA gene,and the symbol � represents a transcriptional terminator sequenceimmediately downstream of the aacC1 gene. The SmaI site was usedfor ligation of random chromosomal inserts. (B) A library (19,200clones) of SmaI chromosomal DNA fragments (0.8 to 2 kb) fromE. amylovora was constructed in pGCM0, transformed into E. amylo-vora Ea110, and screened individually for GUS activity on LB mediumamended with X-Gluc. A 96-well microplate containing slices of peartissue was inoculated with Ea100 containing random IVET fusionclones and incubated for 48 h at 25°C. Clones exhibiting GUS activityon pear disks but not on LB–X-Gluc medium were selected, and theplasmids were recovered for further analysis. Steps 3 to 5 were re-peated.
VOL. 187, 2005 IVET SCREENING OF THE FIRE BLIGHT PATHOGEN 8091
were correlated with high levels of bacterial growth in peartissue during the 4 days postinoculation (Fig. 2B); however, theCFBP1430 dspE mutant M52 grew only slightly in pears, rep-resenting an approximately 105-fold reduction relative to theWT CFBP1430 strain. Populations of the hrpA mutant de-clined quickly after inoculation, indicating that the hrpA mu-tant was not able to survive in immature pear (Fig. 2B).
Verification that infection of immature pear required majorpathogenicity factors of E. amylovora facilitated the develop-ment of a simple, high-throughput IVET system to identifyE. amylovora genes induced during colonization and infection.We used cores of immature pear tissue in a microtiter plateformat, a system that was conducive to handling large numbersof samples.
To develop an immature pear fruit assay, we used the �-glu-curonidase gene uidA as a reporter (36) in the vector pGCM0,which was constructed as described in Materials and Methods(Fig. 1A). The vector was verified with a control constructcontaining the dspE promoter in both orientations (Table 1).The dspE promoter was previously reported to be stronglyinduced during immature pear infection (14). GUS activity wasnot observed after 2 days of growth in LB medium for eitherEa110(pZYF8) (dspE promoter in correct orientation to uidA)or Ea110(pZYF2) (dspE promoter in opposite orientation touidA). However, GUS activity was observed for strainEa110(pZYF8) in qualitative assays 2 days following inoculationonto immature pear disks but not following inoculation of strainEa110(pZYF2) (data not shown). GUS activity was not observedfor the WT Ea110 strain containing the empty pGCM0 vectoreither on LB medium or in pear disks.
To identify E. amylovora genes expressed during coloniza-tion and infection of pear disks, we constructed a library of 0.8-to 2-kb fragments of genomic DNA of Ea110� (cured ofpEA29) in pGCM0 and introduced the library into WT Ea110by electroporation. In order to screen for differentially ex-pressed promoter fusions, we developed an in planta pear diskmicrotiter plate assay (Fig. 1B). Strain Ea110� containing li-brary clones was first grown on LB–X-Gluc medium for 2 days,visually monitored for GUS activity, and then inoculated ontopear disks in microtiter plates (Fig. 1B). GUS activity wasqualitatively detected after 2 days of incubation at 25°C. Onlyclones that showed high GUS activity in pear disks but no GUSactivity on LB plates were recognized as pear-upregulatedclones. Those differentially expressed clones were againscreened on LB–X-Gluc plates and pear disks to confirm theresults, and the DNA inserts from confirmed clones were sub-jected to further analysis. A total of 19,200 transcriptionalfusion clones were screened on both LB–X-Gluc medium andpear disks, and 498 clones (2.5%) were repeatedly found todifferentially express GUS activity on pear disks in this quali-tative assay.
Sequence analysis of E. amylovora genes upregulated in im-mature pear tissue. We determined the sequence of the insertsfrom the 498 clones and identified the putative genes inducedfollowing BLAST searches of the nonredundant GenBank da-tabase. Of the 498 inserts sequenced, a total of 55 genes wereidentified two or more times and 12 clones contained either anintragenic sequence or a sequence with the putative genepresent in the incorrect orientation. Although it is possible thatthese 12 clones may contain cryptic promoter sequences, as has
FIG. 2. Symptoms and growth of Erwinia amylovora WT strainsand hrpA and dspE mutants in immature pear. (A) Symptoms causedby Erwinia amylovora Ea110, CFBP1430, and corresponding hrpA anddspE mutants in immature pear. DPI, days postinoculation. (B) Growth ofErwinia amylovora WT EA110, CFBP1430, and hrpA and dspE mu-tants during infection of immature pears. The growth of bacterialstrains was monitored at 0, 1, 2, 3, and 4 days after inoculation. Datapoints represent the means of three replicates � standard errors. Similarresults were obtained in a second independent experiment.
been shown in a previous study with P. fluorescens (67), weseparated the clones from the others in the current study anddid not subject them to further analysis. Thus, a total of 394unique putative pear-inducible genes were identified, andthese pear fruit-induced (pfi) genes could be divided into nineputative functional groups, including host-microbe interactions(3.8%), stress response (5.3%), regulation (11.9%), cell surface(8.9%), transport (13.5%), mobile elements-phage (1.0%), me-tabolism (20.3%), nutrient acquisition and synthesis (15.5%), andunknown or hypothetical proteins (19.8%).
The majority of the putative gene products identified asinducible during infection of pear tissue shared high aminoacid similarity with proteins from Yersinia spp., Salmonellaspp., E. coli, Shigella spp., and Erwinia spp. (Table 2). Genesfor several known virulence factors previously reported inE. amylovora such as the TTSS genes hrpGF, hrpL, and hrpXand genes encoding known or new effector proteins DspE andHopPtoCEA were upregulated during pear infection (Table 2).Other known E. amylovora virulence genes identified as up-regulated in this study were genes for levansucrase (lsc), reg-ulator of levansucrase (rlsA), amylovoran regulator (rcsA), andzinc-binding metalloprotease (prtA) (Table 2). In addition,genes encoding polygalacturonase (peh), hemagglutinin familyadhesion (hecA), and membrane-bound lytic murein transgly-cosylase (mltE) were identified for the first time in E. amylo-vora Ea110 (Table 2). Peh and HecA are important virulencefactors in E. chrysanthemi (62), and MltE plays a role in thevirulence of P. syringae (12). A total of 56 upregulated genesidentified were homologs of genes identified in IVET studiesperformed with other bacterial plant or animal pathogens(Table 2).
Type II secretion system (T2SS) genes similar to Yersiniaenterocolitica yts1IJ (34), a known virulence factor and one offive major protein secretion systems in many pathogenic bac-teria, were identified for the first time in E. amylovora. Yts1IJproteins are known type 4 pilin-like proteins (pseudopilins).Interestingly, the type II secretion system is dependent on thegeneral secretory pathway (GSP), i.e., the Sec secretion path-way (34). In our study, the major preprotein translocase SecA(ATPase), molecular chaperone SecB, and membrane proteinsSecDF of the GSP were also induced during infection of pearsalong with the T2SS (Table 2). Furthermore, the peh gene,encoding an enzyme that is known to be secreted by the T2SS,was also upregulated (Table 2), indicating that a functionalT2SS is present in E. amylovora and to a greater extent couldcontribute to the virulence of this pathogen.
Transport genes (pfi 16 to 51) including genes for general,ion, sugar, amino acid, peptide, and nucleotide transport pro-teins were induced in pear tissues (Table 2). Some of thetransporters may belong to the type I secretion system that isknown to be involved in secreting toxins, proteases, and lipasesand are potential virulence factors in E. amylovora. Cell sur-face proteins including inner, periplasmic, and outer mem-brane proteins; lipoproteins; flagella; and polysaccharide pro-teins were also induced during pear tissue infection (pfi 52 to72; Table 2). These membrane proteins may be involved inprotein secretion and membrane maintenance. The sensorcomponent (envZ, pfi 94) of a two-component regulatory sys-tem and cognate outer membrane protein genes that this sys-
tem regulates, ompA (pfi 65) and ompC (pfi 64), were alsodifferentially expressed in pears compared to LB medium.
Under unfavorable conditions such as nutritional stress orexposure to a host defense response, bacterial pathogens re-spond by overexpressing stress response genes. Several stressresponse genes (pfi 75 to 90) were identified in our screen(Table 2). These genes included DNA repair or protection(mutS, recA, and sulA), carbon starvation, heat or phage shock,and antioxidant (such as grpE and ahpC) genes. These resultssuggest that pear tissue at least initially is not a favorablehabitat for E. amylovora growth and/or that DNA damage andthe neutralization of plant-derived reactive oxygen species areinvolved in virulence and in planta growth.
The sensor component of a two-component regulatory sys-tem, grrS (pfi 93), was identified as upregulated in this study.GrrS is a homolog of GacS, which, along with GacA, globallyregulates a network of virulence functions in Erwinia caroto-vora, including the production of quorum-sensing signalingmolecules (23). Besides amylovoran and levansucrase regula-tors (rcsA and rlsA), and genes encoding the sensor componentof a two-component regulatory system (grrS and envZ), ourscreen identified fliZ, a positive regulator of the flagellar bio-synthetic operon in enterobacteria, as upregulated. Other reg-ulatory genes (pfi 96 to 126), phage-related sequences (pfi 73 to74), and metabolism and nutrient-scavenging genes (pfi 127 to157) were identified in our screen and listed in Table 2. Werecovered several metabolic genes that are potential precur-sors for the siderophore desferrioxamine biosynthesis in thisstudy (pfi 153 to 157). It is probable that, under unfavorableconditions, the bacterium itself adjusts and overcomes nutrientand iron deficiencies.
The large number of unknown or hypothetical proteins iden-tified in this IVET screen (78 genes, 19.8%) indicates thefuture possibilities of characterizing novel virulence traits in E.amylovora and assigning functions to these proteins. A com-plete genome sequence of E. amylovora is expected soon.When an annotated genome sequence is released, we willmake a listing available upon request of the gene numbers ofthe unknown or hypothetical proteins identified in this study.
Quantitative expression analysis of selected pear-upregu-lated genes. To verify that the pear-upregulated gene promot-ers identified using the qualitative IVET assay are induced inpear, quantitative GUS activity for six strains containing pfipromoter constructs and for promoter constructs containingthe dspE promoter (in both directions) was monitored 24 and48 h after infection of immature pear. The selected clones werechosen to validate the qualitative modified IVET screen andrepresented the major functional groups identified in thestudy. The positive control dspEfor promoter in pZYF8 washighly induced in pear after infection at both 24 and 48 hpostinoculation (Table 3), whereas the negative control dspErev
promoter in pZYF2 showed very low GUS expression (Table3). Most of the pfi clones tested showed various degrees ofinduction of promoter activity at 24 h and 48 h after infectionof immature pear (Table 3). The pfi 43 clone (oppA) was foundto be highly induced at both 24 and 48 h after inoculation,whereas pfi 5 (hopPtoCEA), pfi 9 (yts1IJ), pfi 91 (rlsA), and pfi93 (grrS) were induced only at 48 h postinoculation. The clonecontaining hypothetical protein (sav2932), on the other hand,
VOL. 187, 2005 IVET SCREENING OF THE FIRE BLIGHT PATHOGEN 8093
TABLE 2. Selected list of Erwinia amylovora genes induced during infection of immature pear tissuea
Function pfino. Geneb Organism Blastx_BestHitc GenBank
accession no.Previouslyreportedd
TTSS and effectors 1 dspE* Erwinia amylovora DspE/pathogenicity factor AAC04850 Ech, 772 hrpGF E. amylovora Type III secretion HrpG/F AAB49178 Pst, 123 hrpL* E. amylovora Type III sensor kinase HrpL AAD246824 hrpX* E. amylovora Type III secretion HrpX AAD246815 hopPtoC Pseudomonas syringae
pv. tomatoType III effector HopPtoC NP_790436 Pst, 12
66 Z1931* Escherichia coli Outer membrane protein 3b (a),protease VII
NP_287408
67 ptrA* Yersinia pestis Protease III precursor NP_40463368 mrdB Escherichia coli Rod shape-determining protein
RodANP_752655
69 flgN Shigella flexneri Flagellum synthesis protein FlgN NP_83677970 fliG Shigella flexneri Flagellar motor switch protein FliG NP_70782471 fliM Escherichia coli Flagellar motor switch protein
FliMNP_754254
72 ycgB Escherichia coli Putative sporulation protein NP_287427
Continued on following page
VOL. 187, 2005 IVET SCREENING OF THE FIRE BLIGHT PATHOGEN 8095
113 rbsR* Shigella flexneri Regulator for rbs operon NP_839112114 Psyr0512 Pseudomonas syringae Rhs family protein ZP_00124230115 STY4601* Salmonella enterica Putative regulator of late gene
expressionNP_458684
116 kdpE Salmonella entericaserovar Typhimurium
Response regulator intwo-component regulatorysystem with KdpD
AAL19646
117 PA2177 Pseudomonasaeruginosa
Probable sensor/responseregulator hybrid
NP_250867
118 y3531* Yersinia pestis Putative kinase NP_670828119 YPO0014 Yersinia pestis Putative type II homoserine
was strongly induced at 24 h postinoculation with expressiontailing off at 48 h (Table 3).
Construction and analysis of knockout mutants. Althoughthe IVET experiments were conducted in the E. amylovoraEa110� background and an Ea110 hrpA mutant was available,we were unsuccessful in subsequent attempts to construct chro-mosomal knockout mutants in this strain. Thus, we utilized thestrain Ea1189 for mutant construction. Although the similarityof genetic backgrounds of these strains is currently unknown,the virulence of the two strains is similar (data not shown) andthe overall genome diversity of E. amylovora is relatively low,and therefore, we hypothesize that the expression of promotersidentified as upregulated in Ea110 would be comparable tothat in Ea1189. We chose the hopPtoCEA and mltEEA genes ascandidates for insertional mutagenesis using allele marker ex-change to investigate the potential roles of those genes invirulence. These genes were chosen because hopPtoCEA was anew putative effector in E. amylovora and mltEEA was previ-ously demonstrated to have an effect on virulence in P. syrin-gae. The full sequences of these genes and their corresponding
upstream and downstream sequences were obtained by variousmethods including fully sequencing available clone inserts orrecovering additional flanking DNA sequences using TAIL-PCR. Sequence analysis showed that the deduced amino acidsequence of the hopPtoCEA gene shared 77% similarity withthat of the hopPtoCPST gene from P. syringae pv. tomato (datanot shown). The deduced amino acid sequence of the E. amy-lovora mltEEA gene showed 75% similarity with that of themltEYP gene from Yersinia pestis (data not shown). As de-scribed in Materials and Methods, we were able to generateinsertional mutants for hopPtoCEA and mltEEA genes andtested the effect of these mutations on pathogenesis and bac-terial growth in immature pear.
We conducted two experiments to evaluate the extent ofsymptom production in immature pears caused by the WTstrain Ea1189, the hopPtoCEA mutant ZYC1-3, and themltEEA knockout mutant ZYE3-11. In experiment 1, meanlesion diameters after 6 days of incubation (measured from 10replicate pears per strain) were 2.15 cm, 1.91 cm, and 1.30 cmfor Ea1189, ZYC1-3, and ZYE3-11, respectively. The mean
TABLE 2—Continued
Function pfino. Geneb Organism Blastx_BestHitc GenBank
a Hypothetical and unknown genes are not listed.b Gene name designations were based on originally reported gene products that shared high similarity to pfi clone sequences. Asterisks after the gene name indicate
that two or more clones were identified for the same gene in the experiment.c Predicted proteins or functions based on similar proteins identified using BlastX searches.d Genes identified to be induced in other bacterium plant or animal systems using a similar IVET screen or other in vivo expression systems. Rso, Ralstonia
a Hours postinoculation (hpi).b GUS activity is shown in �mol of 4-methylumbelliferyl produced min�1 109 CFU�1. Data represent the means of three measurements � standard errors. Similar
results were obtained in a second independent experiment.c Fold induction is shown as GUS activity at 24 or 48 hpi/GUS activity at 0 hpi.d dspE promoter in opposite direction from uidA gene.e dspE promoter in correct direction toward uidA gene.f UL, under detection limit of vector control.g sav2932 encodes a hypothetical protein and is not listed in Table 2.
lesion diameter for ZYE3-11 was significantly smaller (P 0.05) than that of Ea1189 and ZYC1-3 following an analysis ofvariance and least significant difference test. In a second ex-periment utilizing 12 replicate pears per strain, mean lesiondiameters (� the standard error of the mean) were 2.21 � 0.12cm and 1.45 � 0.04 cm for Ea1189 and ZYE3-11, respectively,confirming a small but significant difference in symptom ex-pression.
Quantification of bacterial growth in infected immaturepears indicated that there was no difference in growth betweenEa1189 and ZYE3-11 over the first 2 days after infection;however, ZYE3-11 cell counts were 3- to 10-fold less than thatof Ea1189 at 3 and 4 days after inoculation (Fig. 3B). Threereplicate experiments and a combined total of 9 and 12 indi-vidual pears were analyzed at each time point for Ea1189 andZYE3-11, respectively (Fig. 3B), indicating that this small dif-ference in growth was repeatably observed. In contrast, growthof the hopPtoCEA mutant ZYC1-3 was indistinguishable fromthat of Ea1189 in immature pears over the course of theseexperiments (data not shown).
DISCUSSION
We utilized a simplified qualitative IVET approach to scanthe E. amylovora genome and recovered 394 unique chromo-somal genes with increased expression during infection of pearfruit tissue. Our IVET methodology differed from most IVETsystems utilized previously, which depend on the complemen-tation of an essential gene that is expressed via promotersequences that are induced in the habitat of interest. Thesesystems typically screen pools of clones simultaneously, whichcould potentially prevent the identification of clones due tocompetition effects. Since our system was optimized to beconducted in a 96-well plate format, we were able to screenpromoter clones individually, eliminating the competition ef-fects of clone pools. Also, since our system did not rely oncomplementation of an essential gene, we may have identifiedpromoters either expressed at low levels or delayed in expres-sion following immature pear infection; both of these catego-ries of promoters would likely have been missed in a comple-mentation format. Indeed, our quantitative analysis of expressionfrom a subset of promoters representing major functionalgroups identified in the study revealed interesting differencesin temporal expression patterns. The differences in expressionobserved using this small set of promoters suggests that acomprehensive evaluation of the timing of expression of E.amylovora virulence factors would considerably add to ourknowledge of the fire blight pathosystem.
As expected, this study highlighted the importance of typeIII secretion in E. amylovora pathogenesis with the recovery ofgenes encoding regulatory and structural components of theHrp type III secretion system and effector proteins. While wedid not recover all of the currently known hrp-regulated genesin E. amylovora, our results are similar to those of other IVETstudies with plant-pathogenic bacteria. For example, IVETstudies of E. chrysanthemi and P. syringae pv. tomato identifiedtwo and eight hrp-regulated genes, respectively (12, 77). Thesefindings validated our approach and suggested that a detailedanalysis of the genes recovered in this study would further
reveal additional determinants involved in the pathogenesis ofthe fire blight bacterium.
The dspEF operon, encoding the major effector and patho-genicity factor DspE and its cognate chaperone DspF, wasrecovered multiple times in our analysis and shown by quan-titative expression analysis to be highly expressed during pear
FIG. 3. Symptoms and growth of Erwinia amylovora WT Ea1189and corresponding hopPtoCEA and mltEEA mutants in immature pear.(A) Symptoms caused by Erwinia amylovora Ea1189 and hopPtoCEA(ZYC1-3) and mltEEA (ZYE3-11) mutants in immature pear. W, watercontrol; DPI, days postinoculation. (B) Growth of Erwinia amylovoraWT Ea1189 and the mltEEA (ZYE3-11) mutant during infection ofimmature pears. The growth of strains was monitored at 0, 1, 2, 3, and4 days after inoculation. Data points represent the means of threereplicates � standard errors. Similar results were obtained in twoadditional independent experiments.
VOL. 187, 2005 IVET SCREENING OF THE FIRE BLIGHT PATHOGEN 8099
infection (Table 3). The importance of DspE and its homologsto plant pathogenesis is well known in a number of pathosys-tems (14, 25, 45, 54, 69) although the function(s) of this largeprotein has not been elucidated. DspE was recently shown tocontribute to the suppression of salicylic acid-mediated basalimmunity (20); effector suppression of the host defense re-sponse is rapidly becoming recognized as an important strategyof bacterial plant pathogenesis (3). We identified a new putativeeffector, HopPtoCEA, in this study, an ortholog of HopPtoC fromP. syringae pv. tomato (65). As with many effectors from P.syringae, a knockout mutant of HopPtoCEA in E. amylovoraEa1189 was not reduced in virulence, presumably due to func-tional redundancy with other effectors in the E. amylovoragenome. The other known E. amylovora effectors, HrpN andHrpW, were not identified as upregulated in this study, al-though the roles of hrpN and hrpW in the pathogenicity of E.amylovora were reported to differ during infection of immaturepear fruit (41, 75). It is tempting to speculate that additionaleffector proteins may exist in the genome of E. amylovora andcontribute to the virulence of the bacterium.
The importance of type II secretion in E. amylovora patho-genesis was also highlighted with the identification of the up-regulation of genes of the yts1IJ operon and components of thegeneral secretion pathway. Type II secretion is a cooperativeprocess initially dependent upon the secretion of enzymes intothe periplasm by the general secretion pathway followed bytargeted secretion through the type II apparatus (6, 64). In Y.enterocolitica, the Yts1 protein secretion apparatus is unique tohighly pathogenic species, is important for virulence in a mousemodel, and shares homology with type II secretion clusters fromE. chrysanthemi and E. carotovora (9, 34). Peh (polygalacturo-nase), an enzyme thought to be secreted by the T2SS, was alsoupregulated and recovered in our IVET screen (38). While theimportance of polygalacturonase to virulence in soft-rottingErwinia spp. is well known (38), the role of cell wall-degradingenzymes in E. amylovora pathogenesis is currently still un-known. In addition, the upregulation of MltE, a specialized cellwall-degrading enzyme, was interesting in that the function ofthis enzyme is to generate localized openings of the bacterialpeptidoglycan envelope for export of bulky materials includingpossibly toxins and fimbrial proteins and to allow the efficientassembly and anchoring of supramolecular transport com-plexes such as T2SS and TTSS in the cell envelope (21, 42). Asin P. syringae (12), we found that E. amylovora MltE made asmall contribution to virulence.
We identified three additional upregulated enzymes in ourIVET assay that are potentially secreted from the cell includ-ing levansucrase (Lsc) and the adhesin-like protein HecA,which belongs to a class of external virulence factors that iswidely distributed among plant and animal pathogens. HecAfrom E. chrysanthemi contributes to attachment, aggregation,and epidermal cell killing and is thought to be involved in theearliest stages of E. chrysanthemi pathogenesis (62). Levansu-crase, an enzyme that directs the synthesis of levan from su-crose, has a known effect on the virulence of E. amylovoraduring pear seedling infection (29). The PrtA metalloproteasecontributes to E. amylovora virulence in an apple leaf infectionassay and is apparently dependent upon the type I Prt machin-ery for secretion (38, 80). These results demonstrate the im-
portance of TTSS and T2SS and of other external virulencefactors in E. amylovora infection of fruit tissue.
A total of 5.3% of the IVET genes identified were placed inthe functional category of stress response including genes in-volved in the response to reactive oxygen species, both heatand cold shock, and carbon and sulfate starvation. E. amylo-vora apparently induces an initial host defense response earlyafter infection (71, 72); the bacterium is capable of survivingthis plant oxidative burst, with the initial plant cell death andnutrient leakage being thought to provide the impetus forfurther spreading of the pathogen within the plant. The role ofindividual proteins in oxidative stress survival is currently un-known in E. amylovora; however, the alkyl hydroperoxide re-ductase AhpC is a known virulence factor in several plant-pathogenic bacteria, contributing to protection from oxidativestress from an active plant defense response (53).
We recovered a multitude of transporters functioning in theuptake of iron, sugars, amino acids, and inorganic ions. Theinduction of these systems during infection indicates that E.amylovora elaborates various factors as needed to colonizehost tissues. Iron availability is critical to most bacterial patho-gens, and the siderophore desferrioxamine is a virulence factorin E. amylovora (24). We recovered three upregulated proteinsinvolved in iron transport or storage. It is probable that, underunfavorable conditions, the bacterium itself adjusts and over-comes nutrient and iron deficiencies. Several upregulatedtransport proteins recovered were ABC transporters, which ispotentially significant because ABC transporters both directlyand indirectly affect virulence of bacterial pathogens (27).While most of the transporters were involved in uptake, themultidrug resistance protein EmrB (pfi 31) was also upregu-lated and presumably functions in the efflux of plant-derivedtoxins encountered during infection. The role of the AcrABefflux pump in E. amylovora virulence and tolerance of phy-toalexins including phloretin, naringenin, and quercetin wasrecently reported (17). Thus, it is possible that many of theseABC transporters are involved in the virulence of E. amylo-vora. In conjunction with the number of upregulated transport-ers found, a large proportion of the genes identified in thisstudy were involved in metabolism (20.3%) and nutrient ac-quisition (16%). These frequencies may be associated with thehost tissue (immature pear fruit) chosen for study; however, anumber of genes that we identified were also identified in otherIVET studies involving those of E. coli, P. fluorescens, P. syringae,R. solanacearum, and Vibrio cholerae (Table 2) (12, 16, 47, 77).
About 12% of the genes identified in this study were in-volved in regulation, which is a ratio similar to that identifiedin an IVET examination of E. chrysanthemi infection (77).Previously known E. amylovora transcription factors that wereupregulated included RcsA, an activator (along with RcsB) ofamylovoran production (73), and RlsA, an activator of levanproduction (79), along with the capsular polysaccharide exportprotein KpsC. This further confirms that the production ofboth amylovoran and levan in E. amylovora is induced duringinfection. Another important regulator, GrrS (global responseregulator sensor in a two-component regulatory system), is ahomolog of GacS, which, along with GacA, globally regulatesa network that controls exoenzyme and secondary metabolite(toxin) production in Pseudomonas spp. and virulence func-tions in E. carotovora and also regulates the production of
quorum-sensing signaling molecules (18, 23, 61). GacA/GacS-regulated networks also function by positively controlling thetranscription of small regulatory RNAs, transcriptional activa-tors, and alternative sigma factors such as HrpL (18, 32). In E.amylovora, the small regulatory RNA rsmB titrates the repres-sor RsmA in a system that affects exopolysaccharide produc-tion and, therefore, pathogenicity (46).
EnvZ is the sensor component of the OmpR-EnvZ two-component regulatory system that is very important in regu-lating various cellular components such as outer membraneproteins OmpC and OmpA, which is also upregulated in thisstudy. In Salmonella spp., OmpR-EnvZ regulates another two-component system, SsrA-SsrB, that in turn regulates the typeIII secretion system produced by Salmonella pathogenicity is-land 2 (Spi-2) (43). EnvZ is a transmembrane sensor thatpredominantly responds to acidic pH conditions and subse-quently phosphorylates OmpR, which functions as a transcrip-tional activator in the expression of the ssrAB genes (26). SsrAis a second sensor protein that is responsive to acidic pH andalso detects low-osmolarity conditions and the absence of Ca2�
ions, all environmental conditions within macrophages wherethe Spi-2 type III secretion system is exclusively expressed (26).In E. amylovora, the structural components of the TTSS en-coded by the Hrp regulon are regulated by the two-componentsystem HrpX and HrpY, which direct the expression of the�54-dependent, enhancer-binding protein HrpS (74). BothHrpY and HrpS function in activating the expression of thealternate sigma factor HrpL, thereby regulating the variousgenes and operons of the Hrp regulon, which contains HrpL-dependent promoter sequences (74). The expression of HrpXand HrpS is regulated by low pH, low nutrients, and low-temperature conditions, mimicking the plant apoplast but alsorepresenting conditions that suggest that a two-componentregulatory system such as OmpR-EnvZ could further regulatethe hrpXY operon despite no direct evidence to support thisclaim. Interestingly, both hrpX and hrpL, along with the EnvZgene, were found to be upregulated during infection of imma-ture pear in this study (Table 2).
Among the bacterial cell surface and transmembrane-up-regulated proteins, three flagellar proteins, FliG, FliM, andFlgN, were upregulated. The trait of motility is not requiredfor E. amylovora pathogenesis; however, motility does increaseblossom infectivity, particularly at lower cell concentrations(8). Furthermore, a homolog of Y. pestis FliZ, a positive reg-ulator for the flagellar biosynthetic operon and an alternativesigma factor, was also found to be upregulated in our study. InSalmonella enterica serovar Typhimurium, FliZ upregulatesHilA, which in turn activates production of several invasionproteins encoded within the Salmonella pathogenicity island 1(35). Finally, the contribution of cell shape to virulence wasalso highlighted by the recovery of an E. coli RodA homolog;E. amylovora mutants with TnPhoA insertions within the rodA-pbpA operon were previously reported to be avirulent (52).
In summary, our IVET screen successfully identified a vari-ety of genes upregulated during fruit infection by E. amylovora.We utilized a modified IVET method in this study, which isdifferent from many other IVET studies in that we did notimpose a rigorous selection step, i.e., one that necessitatesrescue of an essential phenotype, in our gene identificationwork. The successful identification of a large number of known
virulence genes of E. amylovora in this study further validatedour approach. However, because of the qualitative nature ofthe gene identification step, through �-glucuronidase stainingand visualization of gene expression on pear slices and agarmedium, it is possible that this methodology may have resultedin some artifacts. Nevertheless, the main goal of this work, asin other IVET analyses, was to identify potentially importantgenes in the E. amylovora infection process that could be sub-jected to further detailed studies to clearly delineate the role ofthese genes in pathogenesis.
We further confirmed that the TTSS and its major effectorprotein DspE are essential for full virulence in E. amylovoraduring infection of immature pear. We also found a completeand functional T2SS and its potential secreted proteins in E.amylovora for the first time. We identified a new putativeeffector and external virulence factors such as HecA whichwere previously unknown in E. amylovora and discovered anumber of putative regulatory proteins that may influence theregulation of virulence factors on a global level and eventuallycontribute to the virulence of the bacterium. We can now askquestions concerning the comparative regulation of criticalgenes identified in this study during infection of other hosttissues, particularly blossoms and shoots. It is possible that E.amylovora may utilize differential virulence strategies depend-ing upon the host tissue encountered. Of interest to us also isthe expression profile of these same genes during infection ofhighly susceptible versus fire blight-tolerant apple varieties.
ACKNOWLEDGMENTS
We thank Gayle McGhee for construction of the pGCM0 vector.We thank three anonymous reviewers whose comments greatly strength-ened the manuscript.
This work was supported by grants from the United States Depart-ment of Agriculture and by the Michigan Agricultural ExperimentStation.
REFERENCES
1. Aldridge, P., M. Metzger, and K. Geider. 1997. Genetics of sorbitol metab-olism in Erwinia amylovora and its influence on bacterial virulence. Mol.Gen. Genet. 256:611–619.
2. Alexeyev, M. F. 1995. Three kanamycin resistance gene cassettes with dif-ferent polylinkers. BioTechniques 18:52–54.
3. Alfano, J. R., and A. Collmer. 2004. Type III secretion system effectorproteins: double agents in bacterial disease and plant defense. Annu. Rev.Phytopathol. 42:385–414.
4. Altschul, S. F., T. L. Madden, A. A. Schaffer, J. Zhang, Z. Zhang, W. Miller,and D. J. Lipman. 1997. Gapped BLAST and PSI-BLAST: a new generationof protein database search programs. Nucleic Acids Res. 25:3389–3402.
5. Angelichio, M. J., and A. Camilli. 2002. In vivo expression technology. Infect.Immun. 70:6518–6523.
6. Arrieta, J. G., M. Sotolongo, C. Menendez, D. Alfonso, L. E. Trujillo, M.Soto, R. Ramirez, and L. Hernandez. 2004. A type II protein secretionpathway required for levansucrase secretion by Gluconacetobacter diazotro-phicus. J. Bacteriol. 186:5031–5039.
7. Bartoleschi, C., M. C. Pardini, C. Scaringi, M. C. Martino, C. Pazzani, andM. L. Bernardini. 2002. Selection of Shigella flexneri candidate virulencegenes specifically induced in bacteria resident in host cell cytoplasm. Cell.Microbiol. 4:613–626.
8. Bayot, R. G., and S. M. Ries. 1986. Role of motility in apple blossominfection by Erwinia amylovora and studies of fire blight control with attract-ant and repellent compounds. Phytopathology 76:441–445.
9. Bell, K. S., M. Sebaihia, L. Pritchard, M. T. G. Holden, L. J. Hyman, M.C. Holeva, N. R. Thomson, S. D. Bentley, L. J. C. Churcher, K. Mungall, R.Atkin, N. Bason, K. Brooks, T. Chillingworth, K. Clark, J. Doggett, A.Fraser, Z. Hance, H. Hauser, K. Jagels, S. Moule, H. Norbertczak, D.Ormond, C. Price, M. A. Quail, M. Sanders, D. Walker, S. Whitehead,G. P. C. Salmond, P. R. J. Birch, J. Parkhill, and I. K. Toth. 2004. Genomesequence of the enterobacterial phytopathogen Erwinia carotovora subsp.atroseptica and characterization of virulence factors. Proc. Natl. Acad. Sci.USA 101:11105–11110.
VOL. 187, 2005 IVET SCREENING OF THE FIRE BLIGHT PATHOGEN 8101
10. Bereswill, S., and K. Geider. 1997. Characterization of the rcsB gene fromErwinia amylovora and its influence on exopolysaccharide synthesis andvirulence of the fire blight pathogen. J. Bacteriol. 179:1354–1361.
11. Bernhard, F., K. Poetter, K. Geider, and D. L. Coplin. 1990. The rcsA genefrom Erwinia amylovora—identification, nucleotide-sequence, and regula-tion of exopolysaccharide biosynthesis. Mol. Plant-Microbe Interact. 3:429–437.
12. Boch, J., V. Joardar, L. Gao, T. L. Robertson, M. Lim, and B. N. Kunkel.2002. Identification of Pseudomonas syringae pv. tomato genes induced dur-ing infection of Arabidopsis thaliana. Mol. Microbiol. 44:73–88.
13. Bogdanove, A. J., D. W. Bauer, and S. V. Beer. 1998. Erwinia amylovorasecretes DspE, a pathogenicity factor and functional AvrE homolog, throughthe Hrp (type III secretion) pathway. J. Bacteriol. 180:2244–2247.
14. Bogdanove, A. J., J. F. Kim, Z. Wei, P. Kolchinsky, A. O. Charkowski, A. K.Conlin, A. Collmer, and S. V. Beer. 1998. Homology and functional similarityof an hrp-linked pathogenicity locus, dspEF, of Erwinia amylovora and theavirulence locus avrE of Pseudomonas syringae pathovar tomato. Proc. Natl.Acad. Sci. USA 95:1325–1330.
15. Bogs, J., and K. Geider. 2000. Molecular analysis of sucrose metabolism ofErwinia amylovora and influence on bacterial virulence. J. Bacteriol. 182:5351–5358.
16. Brown, D. G., and C. Allen. 2004. Ralstonia solanacearum genes inducedduring growth in tomato: an inside view of bacterial wilt. Mol. Microbiol.53:1641–1660.
17. Burse, A., H. Weingart, and M. S. Ullrich. 2004. The phytoalexin-induciblemulti-drug efflux pump, AcrAB, contributes to virulence in the fire blightpathogen, Erwinia amylovora. Mol. Plant-Microbe Interact. 17:43–54.
18. Chatterjee, A., Y. Cui, H. Yang, A. Collmer, J. R. Alfano, and A. K. Chatterjee.2003. GacA, the response regulator of a two-component system, acts as amaster regulator in Pseudomonas syringae pv. tomato DC3000 by controllingregulatory RNA, transcriptional activators, and alternative sigma factors.Mol. Plant-Microbe Interact. 16:1106–1117.
19. Coleman, M., R. Pearce, E. Hitchin, F. Busfield, J. W. Mansfield, and I. S.Roberts. 1990. Molecular cloning, expression and nucleotide sequence of thercsA gene of Erwinia amylovora, encoding a positive regulator of capsuleexpression: evidence for a family of related capsule activator proteins.J. Gen. Microbiol. 136:1799–1806.
20. Debroy, S., R. Thilmony, Y. B. Kwack, K. Nomura, and S. Y. He. 2004. Afamily of conserved bacterial effectors inhibits salicylic acid-mediated basalimmunity and promotes disease necrosis in plants. Proc. Natl. Acad. Sci.USA 101:9927–9932.
21. Dijkstra, A. J., and W. Keck. 1996. Peptidoglycan as a barrier to transen-velope transport. J. Bacteriol. 178:5555–5562.
22. Du, Z., and K. Geider. 2002. Characterization of an activator gene upstreamof lsc, involved in levan synthesis of Erwinia amylovora. Physiol. Mol. PlantPathol. 60:9–17.
23. Ericksson, A. R., R. A. Andersson, M. Pirhonen, and E. T. Palva. 1998.Two-component regulators involved in the global control of virulence inErwinia carotovora subsp. carotovora. Mol. Plant-Microbe Interact. 11:743–752.
24. Expert, D., A. Dellagi, and R. Kachadourian. 2000. Iron and fire blight: rolein pathogenicity of desferrioxamine E, the main siderophore of Erwiniaamylovora, p. 179–195. In J. Vanneste (ed.), Fire blight: the disease and itscausative agent Erwinia amylovora. CABI Publishing, New York, N.Y.
25. Frederick, R. D., M. Ahmad, D. R. Majerczak, A. S. Arroyo-Rodriguez, S.Manulis, and D. L. Coplin. 2001. Genetic organization of the Pantoea stew-artii subsp. stewartii gene cluster and sequence analysis of the hrpA, hrpC,hrpN, and wtsE operons. Mol. Plant-Microbe Interact. 14:1213–1222.
26. Garmendia, J., C. R. Beuzon, J. Ruiz-Albert, and D. W. Holden. 2003. Theroles of SsrA-SsrB and OmpR-EnvZ in the regulation of genes encoding theSalmonella typhimurium SPI-2 type III secretion system. Microbiology 149:2385–2396.
27. Garmory, H. S., and R. W. Titball. 2004. ATP-binding cassette transportersare targets for the development of antibacterial vaccines and therapies.Infect. Immun. 72:6757–6763.
28. Gaudriault, S., L. Malandrin, J. P. Paulin, and M. A. Barny. 1997. DspA, anessential pathogenicity factor of Erwinia amylovora showing homology withAvrE of Pseudomonas syringae, is secreted via Hrp secretion pathway in aDspB-dependent way. Mol. Microbiol. 26:1057–1069.
29. Geier, G., and K. Geider. 1993. Characterization and influence on virulenceof the levansucrase gene from the fire blight pathogen Erwinia amylovora.Physiol. Mol. Plant Pathol. 42:387–404.
30. Guttman, D. S., B. A. Vinatzer, S. F. Sarkar, M. V. Ranall, G. Kettler, andJ. T. Greenberg. 2002. A functional screen for the type III (Hrp) secretomeof the plant pathogen Pseudomonas syringae. Science 295:1722–1726.
31. He, S. Y., K. Nomura, and T. S. Whittam. 2004. Type III protein secretionmechanism in mammalian and plant pathogens. Biochim. Biophys. Acta1694:181–206.
32. Heeb, S., C. Blumer, and D. Haas. 2002. Regulatory RNA as mediator inGacA/RsmA-dependent global control of exoproduct formation in Pseudo-monas fluorescens CHA0. J. Bacteriol. 184:1046–1056.
33. Heithoff, D. M., R. L. Sinsheimer, D. A. Low, and M. J. Mahan. 2000. In vivo
gene expression and the adaptive response: from pathogenesis to vaccinesand antimicrobials. Phil. Trans. R. Soc. Lond. B 355:633–642.
34. Iwobi, A., J. Heesemann, E. Garcia, E. Igwe, C. Noelting, and A. Rakin. 2003.Novel virulence-associated type II secretion system unique to high-pathoge-nicity Yersinia enterocolitica. Infect. Immun. 71:1872–1879.
35. Iyoda, S., T. Kamidoi, K. Hirose, K. Kutsukake, and H. Watanabe. 2001. Aflagellar gene fliZ regulates the expression of invasion genes and virulencephenotype in Salmonella enterica serovar Typhimurium. Microb. Pathog.30:81–90.
36. Jefferson, R. A. 1987. Assaying chimeric genes in plants: the GUS genefusion system. Plant Mol. Biol. Rep. 5:387–405.
37. Jin, Q. L., W. Q. Hu, I. Brown, G. McGhee, P. Hart, A. L. Jones, and S. Y.He. 2001. Visualization of secreted Hrp and Avr proteins along with the Hrppilus during type III secretion in Erwinia amylovora and Pseudomoas syrin-gae. Mol. Microbiol. 40:1129–1139.
38. Kazemi-Pour, N., G. Condemine, and N. Hugouvieux-Cotte-Pattat. 2004.The secretome of the plant pathogenic bacterium Erwinia chrysanthemi.Proteomics 4:3177–3186.
39. Kelm, O., C. Kiecker, K. Geider, and F. Bernhard. 1997. Interaction of theregulator proteins RcsA and RcsB with the promoter of the operon foramylovoran biosynthesis in Erwinia amylovora. Mol. Gen. Genet. 256:72–83.
40. Kim, J. F., and S. V. Beer. 2000. hrp genes and harpins of Erwinia amylovora:a decade of discovery, p. 141–161. In J. Vanneste (ed.), Fire blight: thedisease and its causative agent Erwinia amylovora. CABI Publishing, NewYork, N.Y.
41. Kim, J. F., and S. V. Beer. 1998. HrpW of Erwinia amylovora, a new harpinthat contains a domain homologous to pectate lyases of a distinct class. J.Bacteriol. 180:5203–5210.
42. Koraimann, G. 2003. Lytic transglycosylases in macromolecular transportsystems of gram-negative bacteria. Cell. Mol. Life Sci. 60:2371–2388.
43. Lee, A. K., C. S. Detweiler, and S. Falkow. 2000. OmpR regulates thetwo-component system SsrA-SsrB in Salmonella pathogenicity island 2. J.Bacteriol. 182:771–781.
44. Liu, Y., and R. Whittier. 1995. Thermal asymmetric interlaced PCR: autom-atable amplification and sequencing of insert end fragments from P1 andYAC clones for chromosome walking. Genomics 25:674–681.
45. Lorang, J. M., and N. T. Keen. 1995. Characterization of AvrE from Pseudo-monas syringae pv. tomato—a hrp-linked avirulence locus consisting of atleast two transcriptional units. Mol. Plant-Microbe Interact. 8:49–57.
46. Ma, W., Y. Cui, C. K. Dumenyo, A. Mukherjee, and A. K. Chatterjee. 2001.Molecular characterization of global regulatory RNA species that controlpathogenicity factors in Erwinia amylovora and Erwinia herbicola pv. gyp-sophilae. J. Bacteriol. 183:1870–1880.
47. Mahan, M. J., D. M. Heithoff, R. L. Sinsheimer, and D. A. Low. 2000.Assessment of bacterial pathogenesis by analysis of gene expression in thehost. Annu. Rev. Genet. 34:139–164.
48. Marco, M. L., J. Legac, and S. E. Lindow. 2003. Conditional survival as aselection strategy to identify plant-inducible genes of Pseudomonas syringae.Appl. Environ. Microbiol. 69:5793–5801.
49. McGhee, G. C., and A. L. Jones. 2000. Complete nucleotide sequences ofubiquitous plasmid pEA29 from Erwinia amylovora strain Ea88: gene orga-nization and intraspecies variation. Appl. Environ. Microbiol. 66:4897–4907.
50. Merrell, D. S., and A. Camilli. 2000. Detection and analysis of gene expres-sion during infection by in vivo expression technology. Phil. Trans. R. Soc.Lond. B 355:587–599.
51. Metzger, M., P. Bellemann, P. Bugert, and K. Geider. 1994. Genetics ofgalactose metabolism of Erwinia amylovora and its influence on polysaccha-ride synthesis and virulence of the fire blight pathogen. J. Bacteriol. 176:450–459.
52. Milner, J. S., D. Dymock, R. M. Cooper, and I. S. Roberts. 1993. Penicillin-binding proteins from Erwinia amylovora: mutants lacking Pbp2 are aviru-lent. J. Bacteriol. 175:6082–6088.
53. Mongkolsuk, S., W. Whangsuk, P. Vattanaviboon, S. Loprasert, and M.Fungthong. 2000. A Xanthomonas alkyl hydroperoxide reductase subunit C(ahpC) mutant showed an altered peroxide stress response and complexregulation of the compensatory response of peroxide detoxification enzymes.J. Bacteriol. 182:6845–6849.
54. Mor, H., S. Manulis, M. Zuck, R. Nizan, D. L. Coplin, and I. Barash. 2001.Genetic organization of the hrp gene cluster and dspAE/BF operon inErwinia herbicola pv. gypsophilae. Mol. Plant-Microbe Interact. 14:431–436.
55. Morgan, E., J. D. Campbell, S. C. Rowe, J. Bispham, M. P. Stevens, A. J.Bowen, P. A. Barrow, D. J. Maskell, and T. S. Wallis. 2004. Identification ofhost-specific colonization factors of Salmonella enterica serovar Typhi-murium. Mol. Microbiol. 54:994–1010.
56. Mukaihara, T., N. Tamura, Y. Murata, and M. Iwabuchi. 2004. Geneticscreening of Hrp type III-related pathogenicity genes controlled by the HrpBtranscriptional activator in Ralstonia solanacearum. Mol. Microbiol. 54:863–875.
57. Osbourn, A. E., C. E. Barber, and M. J. Daniels. 1987. Identification ofplant-induced genes of the bacterial pathogen Xanthomonas campestrispathovar campestris using a promoter probe plasmid. EMBO J. 6:23–28.
58. Rainey, P. B. 1999. Adaptation of Pseudomonas fluorescens to the plantrhizosphere. Environ. Microbiol. 1:243–257.
59. Rainey, P. B., and G. M. Preston. 2000. In vivo expression technology strat-egies: valuable tools for biotechnology. Curr. Opin. Biotechnol. 11:440–444.
60. Ramos-Gonzalez, M. I., M. J. Campos, and J. L. Ramos. 2005. Analysis ofPseudomonas putida KT2440 gene expression in the maize rhizosphere: invitro expression technology capture and identification of root-activated pro-moters. J. Bacteriol. 187:4033–4041.
61. Reimmann, C., M. Beyeler, A. Latifi, H. Winteler, M. Foglino, A. Lazdunski,and D. Haas. 1997. The global activator GacA of Pseudomonas aeruginosaPAO positively controls the production of the autoinducer N-butyryl-homo-serine lactone and the formation of the virulence factors pyocyanin, cyanide,and lipase. Mol. Microbiol. 24:309–319.
62. Rojas, C. M., J. H. Ham, W. L. Deng, J. J. Doyle, and A. Collmer. 2002.HecA, a member of a class of adhesions produced by diverse pathogenicbacteria, contributes to the attachment, aggregation, epidermal cell killing,and virulence phenotypes of Erwinia chrysanthemi EC16 of Nicotiana cleve-landii seedlings. Proc. Natl. Acad. Sci. USA 99:13142–13147.
63. Sambrook, J., and D. W. Russell. 2001. Molecular cloning: a laboratorymanual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.
64. Sandkvist, M. 2001. Type II secretion and pathogenesis. Infect. Immun.69:3523–3535.
65. Schechter, L. M., K. A. Roberts, Y. Jamir, J. R. Alfano, and A. Collmer. 2004.Pseudomonas syringae type III secretion system targeting signals and noveleffectors studied with a Cya translocation reporter. J. Bacteriol. 186:543–555.
66. Schweizer, H. P., and T. T. Hoang. 1995. An improved system for genereplacement and xylE fusion analysis in Pseudomonas aeruginosa. Gene 158:15–22.
67. Silby, M. W., and S. B. Levy. 2004. Use of in vivo expression technology toidentify genes important in growth and survival of Pseudomonas fluorescensPf0-1 in soil: discovery of expressed sequences with novel genetic organiza-tion. J. Bacteriol. 186:7411–7419.
68. Snyder, J. A., B. J. Haugen, E. L. Buckles, C. V. Lockatell, D. E. Johnson,M. S. Donnenberg, R. A. Welch, and H. L. T. Mobley. 2004. Transcriptomeof uropathogenic Escherichia coli during urinary tract infection. Infect. Im-mun. 72:6373–6381.
69. Toth, I. K., K. S. Bell, M. C. Oleva, and P. R. Birch. 2003. Pathogen profile,soft rot erwinia: from genes to genomes. Mol. Plant Pathol. 4:17–30.
70. Vanneste, J. L., and S. Eden-Green. 2000. Migration of Erwinia amylovora inhost plant tissues, p. 73–83. In J. Vanneste (ed.), Fire blight: the disease andits causative agent, Erwinia amylovora. CABI Publishing, New York, N.Y.
71. Venisse, J. S., M. A. Barny, J. P. Paulin, and M. N. Brisset. 2003. Involve-ment of three pathogenicity factors of Erwinia amylovora in the oxidativestress associated with compatible interaction in pear. FEBS Lett. 537:198–202.
72. Venisse, J. S., G. Gullner, and M. N. Brisset. 2001. Evidence for the involve-ment of an oxidative stress in the initiation of infection of pear by Erwiniaamylovora. Plant Physiol. 125:2164–2172.
73. Wehland, M., C. Kiecker, D. L. Coplin, O. Kelm, W. Saenger, and F. Bern-hard. 1999. Identification of an RcsA/RcsB recognition motif in the promot-ers of exopolysaccharide biosynthetic operons from Erwinia amylovora andPantoea stewartii subspecies stewartii. J. Biol. Chem. 274:3300–3307.
74. Wei, Z. M., J. F. Kim, and S. V. Beer. 2000. Regulation of hrp genes and typeIII protein secretion in Erwinia amylovora by HrpX/HrpY, a novel twocomponent system, and HrpS. Mol. Plant-Microbe Interact. 13:1251–1262.
75. Wei, Z. M., R. J. Laby, C. H. Zumoff, D. W. Bauer, S. Y. He, A. Collmer, andS. V. Beer. 1992. Harpin, elicitor of the hypersensitive response produced bythe plant pathogen Erwinia amylovora. Science 257:85–88.
76. Wilson, K. J., A. Sessitsch, J. C. Corbo, K. E. Giller, A. D. L. Akkermans, andR. A. Jefferson. 1995. Glucuronidase (GUS) transposons for ecological andgenetic studies of rhizobia and other Gram-negative bacteria. Microbiology141:1691–1705.
77. Yang, S., N. T. Perna, D. A. Cooksey, Y. Okinaka, S. E. Lindow, A. M.Ibekwe, N. T. Keen, and C. H. Yang. 2004. Genome-wide identification ofplant-upregulated genes of Erwinia chrysanthemi 3937 using a GFP-basedIVET leaf array. Mol. Plant-Microbe Interact. 17:999–1008.
78. Young, G. M., and V. L. Miller. 1997. Identification of novel chromosomalloci affecting Yersinia enterocolitica pathogenesis. Mol. Microbiol. 25:319–328.
79. Zhang, Y., and K. Geider. 1999. Molecular characterization of the rlsA generegulating levan production by the fire blight pathogen Erwinia amylovora.Physiol. Mol. Plant Pathol. 54:187–201.
80. Zhang, Y., D. D. Bak, H. Heid, and K. Geider. 1999. Molecular character-ization of a protease secreted by Erwinia amylovora. J. Mol. Biol. 289:1239–1251.
VOL. 187, 2005 IVET SCREENING OF THE FIRE BLIGHT PATHOGEN 8103