The Plant Journal (2015) 82, 352–362
doi: 10.1111/tpj.12812
TECHNICAL ADVANCE
A straightforward and reliable method for bacterial in planta transcriptomics: application to the Dickeya dadantii/ Arabidopsis thaliana pathosystem Emilie Chapelle1,†, Beno^ıt Alunni1,2,†, Pierrette Malfatti1, Lucie Solier1, Jacques Pedron1,3,4, Yvan Kraepiel1,3,4 and Frederique Van Gijsegem1,3,4,* 1 Institut National de la Recherche Agronomique, Universite Pierre et Marie Curie/Universite Paris 06, AgroParisTech, UMR217, Interactions Plantes–Pathoge nes, F-75005 Paris, France, 2 Institute for Integrative Cell Biology, Commissariat a l’Energie Atomique/Centre National de la Recherche Scientifique/Universite Paris-Sud, 1 Avenue de la Terrasse, 91198 Gif sur Yvette, France, 3 Universite Pierre et Marie Curie/Universite Paris 06, UMR 1392, Institute of Ecology and Environmental Sciences, 46 rue d’Ulm, F-75005 Paris, France, and 4 Institut National de la Recherche Agronomique, UMR 1392, Institute of Ecology and Environmental Sciences, 46 rue d’Ulm, F-75005 Paris, France Received 2 January 2015; revised 16 February 2015; accepted 23 February 2015; published online 3 March 2015. *For correspondence (e-mail
[email protected]). † These authors contributed equally to this work.
SUMMARY Transcriptome analysis of bacterial pathogens is a powerful approach to identify and study the expression patterns of genes during host infection. However, analysis of the early stages of bacterial virulence at the genome scale is lacking with respect to understanding of plant–pathogen interactions and diseases, especially during foliar infection. This is mainly due to both the low ratio of bacterial cells to plant material at the beginning of infection, and the high contamination by chloroplastic material. Here we describe a reliable and straightforward method for bacterial cell purification from infected leaf tissues, effective even if only a small amount of bacteria is present relative to plant material. The efficiency of this method for transcriptomic analysis was validated by analysing the expression profiles of the phytopathogenic enterobacterium Dickeya dadantii, a soft rot disease-causing agent, during the first hours of infection of the model host plant Arabidopsis thaliana. Transcriptome profiles of epiphytic bacteria and bacteria colonizing host tissues were compared, allowing identification of approximately 100 differentially expressed genes. Requiring no specific equipment, cost-friendly and easily transferable to other pathosystems, this method should be of great interest for many other plant–bacteria interaction studies. Keywords: soft rot, Erwinia chrysanthemi, Dickeya dadantii, in planta transcriptome, Arabidopsis thaliana, virulence genes, technical Advance.
INTRODUCTION In natural environments, pathogens adopt very different lifestyles, and in particular have to switch from an epiphytic to a parasitic way of life when entering into their host. Under each condition, pathogens tailor their gene expression to adapt to these specific environments. Several key steps during the interaction with the host determine the success of a pathogen infection. However, the early phases of the interaction remain an enigma. Gene 352
expression studies of phytopathogenic bacteria within host tissues have mostly been restricted to a limited number of genes, either in candidate gene studies or through in vivo fishing strategies such as In Vivo Expression Technology (IVET) (Boch et al., 2002; Brown and Allen, 2004; Yang et al., 2004; Yu et al., 2013). Due to the development of powerful genome sequencing methods, characterization of bacterial behaviour in vitro through genome-scale RNA © 2015 The Authors The Plant Journal © 2015 John Wiley & Sons Ltd
Transcriptome of D. dadantii in planta 353 profiling has provided some insights into bacterial lifestyle adaptation mechanisms. For example, in various pathogenic bacteria, transcriptomic analyses allowed identification of stress tolerance or metabolic pathways activated by abiotic factors such as temperature, O2 availability or nutrient sources (Koide et al., 2006; Bronstein et al., 2008; Serrania et al., 2008; Babujee et al., 2012). Adapted RNA harvesting protocols recently allowed the first in vivo transcriptomic virulence expression analysis of a few animal and insect pathogens (Audia et al., 2008; Camejo et al., 2009; Toledo-Arana et al., 2009; Costechareye et al., 2013). However, pathogen transcriptomes have barely been studied during interactions of bacteria with their host plants. Such analyses have only been reported in legume/rhizobia systems, taking advantage of the high bacterial population present in nodules (Capela et al., 2006; Pessi et al., 2007), or for vascular pathogens, where bacteria may easily be recovered by simple exudation (Soto-Suarez et al., 2010; Jacobs et al., 2012). In the case of bacterial foliar pathogens, analysis of the early stages of virulence expression at the genome scale is lacking with respect to understanding of the regulatory pathways enabling adaptation to the host microenvironment and in elucidating the switch to a pathogenic lifestyle. There are several stumbling blocks in plant pathosystems that explain the lack of specific techniques allowing transcriptomic studies of foliar bacterial pathogens in planta (Hinton et al., 2004; La et al., 2008). First, during the early stages of bacterial colonization, the proportion of prokaryotic RNA in crude samples is very low, making it difficult to perform dynamic studies of the bacterial transcriptome under these conditions. Second, plant, especially chloroplastic, RNA is a contaminant that is not efficiently removed by techniques used in animal pathosystem analysis, such as bacterial RNA enrichment or eukaryotic RNA subtraction (Graham and Clark-Curtiss, 1999; La et al., 2008). One way to overcome these problems is to purify bacteria from infected plant tissues before RNA extraction. We tackled this difficulty during the early stages of plant infection by the phytopathogenic enterobacterium Dickeya dadantii, a causal agent of soft rot disease on numerous crops and ornamental plants (Ma et al., 2007). The D. dadantii infection process may be divided in two phases: asymptomatic host colonization and disease occurrence. After penetrating its host plant, D. dadantii resides latently in the plant intercellular spaces without provoking any symptoms during the first hours of infection, and disease occurs only when the environmental conditions are favourable for production of virulence factors and massive bacterial multiplication (Perombelon, 2002). Mutagenesis and targeted gene expression analyses allowed identification of many bacterial virulence factors and related regulators (Barras et al., 1994; Charkowski et al., 2012). Like many other gram-negative pathogenic bacteria, D. dadantii relies
upon several systems of protein secretion for its virulence (Charkowski et al., 2012). Soft rot symptoms are mainly caused by the production and secretion of a battery of plant cell wall-degrading enzymes (e.g. pectinases, cellulase and proteases) that disrupt the integrity of plant tissues and subsequently trigger cell death. D. dadantii also possesses an Hrp type III secretion system (T3SS) that has been shown to play a role in pathogenesis both at low infection inocula (Bauer et al., 1994, 1995) and in semi-tolerant saintpaulia plants (Yang et al., 2002). Furthermore, several reports have highlighted the importance in virulence of various other virulence factors, including bacterial cell envelope components such as surface polysaccharides (lipopolysaccharides and exopolysaccharides) and adhesins (Schoonejans et al., 1987; Condemine et al., 1999; Rojas et al., 2002), or efficient iron uptake systems (Expert, 1999; Franza et al., 2005). D. dadantii is also able to tolerate the chemical and environmental stresses encountered in planta, and to escape from infection-induced plant defence responses by accumulating osmoprotectants (Gouesbet et al., 1996) or antioxidants (Reverchon et al., 2002). It also produces several factors involved in the detoxification of antibacterial compounds or in cellular repair (Lopez-Solanilla et al., 1998; El-Hassouni et al., 1999; Santos et al., 2001). Bacterial attack and spreading rely upon fine genetic regulation, allowing adaptation to the plant environment at the onset of infection and enabling the switch to disease development. Candidate gene expression analysis after foliar inoculation of the model plant Arabidopsis using total RNA of infected plants revealed concerted activation of several virulence genes controlled by a set of master regulators at 12 hours post-inoculation (hpi), i.e. when bacterial populations in the leaves stabilized before symptom occurrence (Lebeau et al., 2008; Reverchon et al., 2010; Mhedbi-Hajri et al., 2011). However, data on gene expression modulation at the whole-genome level during the very early stages of infection are still lacking. To overcome these problems, we have devised a reliable and straightforward method for bacterial cell purification from infected leaf tissues, allowing transcriptomic analysis even when bacterial populations are very low relative to plant material. We validated our method by studying the transcriptomic changes occurring in D. dadantii at the earliest stages of Arabidopsis thaliana leaf infection. It is anticipated that this protocol will be of great help for many other plant–bacteria interaction studies. RESULTS Localization and growth kinetics of D. dadantii during the early steps of Arabidopsis infection In several plants, D. dadantii infects leaves via wounds and natural openings such as stomata or water pores of hydathodes. To mimic these natural inoculation
© 2015 The Authors The Plant Journal © 2015 John Wiley & Sons Ltd, The Plant Journal, (2015), 82, 352–362
354 E. Chapelle et al. conditions, our infection kinetics studies were performed using a non-invasive inoculation method that involves briefly immersing 6-week-old Arabidopsis plantlets into a D. dadantii bacterial suspension as previously described (Lebeau et al., 2008). Microscopic observations using a bacterial strain constitutively expressing the fluorescent GFP protein revealed that, until 6 hpi, the bacterial population was mainly scattered and freely swimming on the leaf surface (Figure S1). At this early step of the interaction, bacteria may thus be recovered simply by rinsing the inoculated plants in a buffer and collecting the bacteria by centrifugation. Between 8 and 12 hpi, bacteria were mainly found in microcolonies deeply nestled at the junctions between epidermal cells and in the parenchyma apoplast (Figure S1). The evolution of bacterial population sizes on/in the leaves during infection was quantified by plating dilution series of bacteria (Figure S2). During the first 12 hpi, there was no massive bacterial multiplication, such that 105 to 106 bacteria were present per leaf and no symptoms were visible. It was only at 24 hpi, when the first maceration symptoms appeared, that the bacterial population rapidly multiplied in macerated plant tissue to reach approximately 107 bacteria per leaf. Thus, to identify the genes important for plant colonization at the very start of the infection process, bacterial RNA must be extracted under conditions where only 105 to 106 bacteria are present per leaf.
Isolation of bacterial cells from leaf tissues As the amount of RNA in bacteria has been estimated to be approximately 0.05–0.1 pg/cell (La et al., 2008), it was necessary to devise a protocol allowing isolation of bacteria from several hundred leaves (70–100 Arabidopsis plants in our pathosystem) in order to obtain sufficient amounts of highly purified bacterial RNA for transcriptomic studies. Furthermore, bacterial RNA turnover occurs within a few minutes (Rauhut and Klug, 1999; Gingeras and Rosenow, 2000), so the bacterial transcriptome must be fixed at the time of harvesting and during the whole purification procedure. This was achieved by performing the whole procedure in the presence of the RNA-stabilizing reagent GenelockTM, now renamed AssayAssureTM (Audia et al., 2008). To establish the purification procedure, inoculations were performed using a D. dadantii strain constitutively expressing the fluorescent GFP protein. After collection of infected rosettes, disruption of the infected leaves and filtration to remove plant cell debris, bacteria were enriched from the homogenate by differential centrifugations (Figure 1a, see Experimental procedures). However, the pellets were still heavily contaminated by green plant material, consisting mainly of chloroplasts as determined by microscopic observations (Figure 1b). To remove the plant plastids, the samples were further purified on a digressive density gradient. When bacteria were abundant enough to be extracted from a few infected plantlets (5–10, from 24
(a) Harvesting of infected leaves in GenelockTM RNA protecting reagent
(b) Blender homogenization and plant tissues removal (10 µm filtration + differential pelleting) X 1000
(c) 10% 25% 30%
Isolation of bacterial cells through density gradient centrifugation (GentodenzTM layers)
Figure 1. Recovery of pathogenic bacterial cells from infected plant tissues. (a) Workflow of the isolation procedure. See Experimental procedures for a detailed description. This method relies upon the use of density gradient centrifugation in the presence of a RNA protecting reagent. For illustrative purposes, the experiment presented here was performed using a GFP-tagged D. dadantii strain. (b,d) Microscopic observations of the samples before density gradient centrifugation (b) (left, visible; right, UV) and after density gradient centrifugation (d) (left, visible; right, UV). (c) The gradient after centrifugation, either after treatment of six infected plants collected at 24 hpi (left) or 70 infected plants collected at 12 hpi (right).
Plant Contaminants
50%
Bacterial cells
(d) Recovery, clean up, and –80°C storage of bacterial pellet prior to RNA extraction
X 1000
© 2015 The Authors The Plant Journal © 2015 John Wiley & Sons Ltd, The Plant Journal, (2015), 82, 352–362
Transcriptome of D. dadantii in planta 355 hpi), bacteria were visibly separated from green contaminants in the gradient (Figure 1c, left). Under these conditions, the bacterial layer was simply removed by pipetting, and bacteria were recovered by centrifugation. However, at earlier time points, many more plants had to be treated to recover enough bacteria (>108 cells). Under such conditions, plant material was so abundant that the bacterial ring was not easily visible (Figure 1c, right). Nevertheless, bacteria were recovered by collecting the part of the gradient just below the green ring of plant material. Under these conditions, after centrifugation, the bacterial pellet was still covered by a densely packed green layer comprising remaining plastid contaminants. However, this green layer is loose and may easily be detached by gently pipetting the buffer up and down, and then discarded (see Experimental procedures). Microscopic examination of the final bacterial pellet obtained under these conditions shows that chloroplastic contamination was insubstantial (Figure 1d and Table 1). Bacterial RNA quality and purity assessments To validate our purification protocol, RNAs were extracted from isolated bacteria by the hot phenol method, and tested for (i) the efficiency of bacterial transcriptome preservation by Genelock treatment, and (ii) the efficiency of plant contamination removal. To check the efficiency of the bacterial transcriptome stabilizing during the purification process, we used quantitative RT-PCR to compare the expression profiles of known virulence genes in total RNA (plant + bacteria) extracted directly from infected plants and in RNA extracted from bacterial samples purified from plant material by our protocol. We analysed the kinetics of activation of the pelBCD genes, encoding three major pecTable 1 Enrichment of bacterial RNA in RNA extracted from purified bacteria compared to total RNA isolated from infected plants.
Experiment 1 8 hpi 24 hpi Experiment 2 8 hpi 24 hpi Experiment 3 8 hpi 24 hpi
lpxC
TUB6
23S
131 100
76 359
82 260
9891 100
911 359
146 260
986 2139
14 115
26 65
Transcript accumulations for the constitutively expressed bacterial lpxC gene, the constitutively expressed tubulin-encoding plant nuclear gene (TUB6) and the chloroplastic 23S rRNA gene (23S) were analysed by quantitative RT-PCR during Arabidopsis infection. The values represent the purification factor for bacterial RNA (positive values; ratio of bacterial RNA after/before purification) and depletion factors for plant nuclear RNA and chloroplastic rRNA (negative values; ratio of plant nuclear or chloroplastic RNA after/before purification).
tate lyases, the prtC gene, encoding a secreted protease, and the hrpN gene, encoding the T3SS HrpN effector. The pattern of expression of these five virulence genes during infection is well documented (Lebeau et al., 2008; Reverchon et al., 2010; Mhedbi-Hajri et al., 2011). The transcript accumulation was followed by quantitative RT-PCR from 0 to 24 hpi. The data in Figure 2 show that the kinetics of virulence gene activation was retained in purified bacterial samples, confirming that the bacterial RNA steady-state status is not altered during the purification process. To estimate the enrichment efficiency of our purification protocol, the levels of bacterial RNA purification and contaminant plant RNA depletion were analysed by quantitative RT-PCR using primers designed to amplify constitutively expressed bacterial lpxC gene transcripts (Hommais et al., 2011), TUB6 tubulin-encoding nuclear plant gene transcripts, and chloroplastic 23S rRNA gene transcripts. The increase in the relative bacterial RNA content (approximately 100 to 10 000-fold compared to crude total RNA samples) and the depletion of the relative plant nuclear-encoded and chloroplastic RNA content (14 to 900fold, approximately, compared to crude total RNA samples) varied greatly among bacterial purifications (Table 1). Nevertheless, in all cases, our bacteria purification protocol resulted in a 104 to 105-fold enrichment of bacterial RNA. The purification procedure was efficient even at an early time point after inoculation (8 hpi), where the amount of bacteria was very low compared to plant material. Transcriptomic analysis of D. dadantii virulence regulation in planta during the first stages of infection Our microscopic analysis revealed the first occurrences of bacterial entry inside leaves at 6 hpi, and previous transcriptomic analyses showed that, under our infection conditions, most tested virulence genes were already activated at 12 hpi (Lebeau et al., 2008; Reverchon et al., 2010; Mhedbi-Hajri et al., 2011). To assess the transcriptomic changes during the very early stages of infection, we compared the transcriptome profiles of bacteria present on leaves at 4 hpi after an epiphytic phase of acclimation with those of bacteria isolated from infected whole plants at 8 hpi when bacteria are already in intimate association with plant tissues (see Figure S1). Bacteria were purified from 50 6-week-old plantlets at each time point. Three independent experiments were performed at various times using independent seed sowings. RNA was harvested, and gene transcription was measured using NimbleGen microarray chips (see Experimental procedures). Differentially expressed genes were identified using the ANAIS web interface (Simon and Biot, 2010; http://anais.versailles.inra.fr/), and a total of 487 genes that exhibited significant differential expression (adjusted P values < 0.05) were detected. We focused on the 107 genes whose expression varied more than twofold (Table S1), of which
© 2015 The Authors The Plant Journal © 2015 John Wiley & Sons Ltd, The Plant Journal, (2015), 82, 352–362
356 E. Chapelle et al.
Relative fold change
(a)
Total RNA 256 64
pelB
16
pelC
4
pelD
1
prtC
0.25
hrpN 0
6
12
24
Relative fold change
Hours post inoculation (hpi) RNA extracted from purified bacteria 64 32 16 8 4 2 1 0.5 0 6 12 24
pelC pelD prtC hrpN
Hours post inoculation (hpi)
(b) Relative fold change (RNA extracted from purified bacteria)
pelB
8
R2 = 0.87
6 4 2 0 –2 –2
0
2
4
6
8
Relative fold change (total RNA) Figure 2. Comparison of virulence gene transcript accumulation in total RNA isolated from infected plants and RNA extracted from purified bacteria. (a) Transcript accumulations for three pel genes, encoding the major pectate lyases pelB, pelC and pelD, prtC, encoding a major protease, and hrpN, encoding the type III harpin effector, were quantitatively analysed during Arabidopsis infection. Transcript accumulations were normalized against transcript accumulation of the constitutive lpxC gene, and were expressed as relative fold change using the transcript level measured at 0 hpi (set to 1) for each gene as reference. (b) Comparison of gene expression measurements in total RNA isolated from infected plants and RNA extracted from purified bacteria. The fold changes calculated as in (a) were compared in two independent experiments. The correlation factor (R2) is given.
62 were up-regulated and 45 were down-regulated. These genes were classified into seven categories according to their function (Figure 3). Approximately 40% of these genes have no known function, corresponding to the percentage of genes of unknown function found in the total genome (Glasner et al., 2011), and, notably, the majority of them (68%) were down-regulated at 8 hpi. Eleven genes (10%) are related to metabolism, which is lower than metabolic genes present in the genome (28%). Other than unknown and metabolism, genes in three categories (transport, protein secretion and responses to stresses) were the most represented (Figure 3 and Table 2).
Interestingly, among the 15 genes encoding transport functions, four of the down-regulated genes (togM, togN, kdgN and kduD) are related to pectin catabolism. The 13 modulated genes encoding proteins that are either secreted or involved in a secretion system include only one up-regulated gene (pel), indicating that the pectin catabolic pathway was not activated at this early colonization step of infection before onset of the maceration process. By contrast, six hrp/hrc genes involved in the type III secretion system were activated at 8 hpi infection. These included the hrpN gene (encoding a harpin) and the genes clustered in the hrpAB hrcJ hrpDE operon. Four of the down-regulated genes involved in protein secretion encode components of the secondary type II secretion system Stt (Ferrandez and Condemine, 2008). Of the ten genes related to stress responses, eight were up-regulated at 8 hpi. These include the three ind genes involved in biosynthesis of the antioxidant molecule indigoidine, the emrAB operon, encoding a multidrug resistance system, and two genes implicated in the SOS response and DNA repair (Table 2). To validate the microarray data, the activation of 17 genes belonging to the three main categories (transport, protein secretion and responses to stresses) was assessed by quantitative RT-PCR on total RNA of infected plants from two new independent infection experiments using non-enriched samples (Table 3). The expression of all genes was at least twofold higher at 8 hpi compared to 4 hpi, except for one putative transporter gene (ID15684) for which the fold change was between 1.7- and 1.9-fold. It should be noted that, even though up-regulation was always observed, the intensity of the activation varied from one experiment to the other due to the variability encountered in biological replicates. By contrast, no large variations were observed when comparing transcript accumulations of these genes at 4 hpi and on plant leaves collected just after immersion in the bacterial suspension (Table 3). DISCUSSION Transcriptome analysis of bacterial pathogens is a powerful approach to identify and analyse the expression patterns of genes that are differentially expressed during infection. However, studying pathogens during early steps of the interaction with their hosts is challenging, as the amount of pathogen RNA is often vanishingly small compared to host RNA, especially at early time points after infection. Study of foliar plant pathogens poses a further complication due to the presence of large amounts of chloroplastic contamination. Here, we present a method that allows transcriptomic analysis of bacterial pathogen gene expression at the onset of infection of a host plant after foliar infection. Our protocol is based on bacterial cell isolation prior to RNA
© 2015 The Authors The Plant Journal © 2015 John Wiley & Sons Ltd, The Plant Journal, (2015), 82, 352–362
Transcriptome of D. dadantii in planta 357 Table 2 Selected categories of genes identified in the transcriptome as up- or down-regulated at 8 hpi compared with 4 hpi ASAP ID
Name
Gene description
Fold change (8 hpi/4 hpi)
P value (ANOVA)
Metabolism ABF-0015532 ABF-0015533 ABF-0015679 ABF-0017046 ABF-0015004 ABF-0047120 ABF-0020854 ABF-0014533 ABF-0014883 ABF-0014882 ABF-0016568 Transport ABF-0018996 ABF-0015135 ABF-0015975 ABF-0015977 ABF-0015684
trpG trpE kdsC accC hemD fadD plsX N/A cysI cysJ pyrI
Anthranilate synthase, amidotransferase component Component I of anthranilate synthase 3-deoxy-D-manno-octulosonate 8-phosphate phosphatase Acetyl CoA carboxylase, biotin carboxylase subunit Uroporphyrinogen III synthase Acyl CoA synthetase, long-chain fatty acid–CoA ligase Fatty acid/phospholipid synthesis protein Hypothetical protein Sulfite reductase, b subunit, NAD(P)-binding Sulfite reductase, a subunit, flavoprotein Aspartate carbamoyltransferase, regulatory subunit
6.3 4.8 2.6 2.2 2.1 2.1 2.1 2.0 2.4 2.5 3.4
3.07E-03 1.14E-03 3.26E-05 2.64E-03 9.40E-07 3.93E-07 2.16E-04 4.32E-04 4.87E-04 1.30E-03 4.35E-03
N/A nhaA ygaZ ygaY yrbF
5.7 3.4 3.3 2.9 3.0
4.16E-03 1.64E-03 2.01E-07 2.66E-07 2.35E-05
ABF-0017318 ABF-0020490 ABF-0016875 ABF-0015523 ABF-0174089 ABF-0019625 ABF-0018199 ABF-0020754 ABF-0015566
N/A ffh N/A kdgN togN togM ganC ynfM araF
2.6 2.2 2.1 2.2 2.2 3.3 2.3 2.4 2.5
1.36E-05 2.26E-04 4.47E-03 1.07E-03 2.37E-03 6.50E-04 2.29E-03 1.21E-07 3.22E-03
ABF-0015475 Protein secretion ABF-0020837 ABF-0018117 ABF-0020784 ABF-0019590
kefG
ABC-type Fe3+ transport system, periplasmic component Sodium–proton antiporter Putative amino acid transporter Major facilitator superfamily permease protein Predicted toluene transporter subunit: ATP-binding component of ABC superfamily Ferrichrome-iron receptor Signal recognition particle (SRP) component with 4.5S RNA ABC transporter, permease protein Oligogalacturonate-specific porin Oligogalacturonide ABC transporter, permease component Oligogalacturonide ABC transporter, permease component Galactan ABC transporter Predicted transporter Periplasmic-binding component of an ABC superfamily Larabinose transporter Component of potassium effux complex with KefB
2.4
7.02E-05
3.6 2.8 3.3 2.7
7.85E-04 8.12E-04 2.11E-07 2.04E-04
ABF-0019593 ABF-0019587 ABF-0019592 ABF-0019588 ABF-0020728 ABF-0020727 ABF-0047167 ABF-0020342 ABF-0018595 Stress ABF-0016084 ABF-0016083
hrpA hrpE hrpB hrpD sttF sttG sttE sttD N/A
Pectate lyase precursor Preprotein translocase membrane subunit Harpin hrpN Type III secretion bridge between inner and outer membrane lipoprotein (YscJ, HrcJ, EscJ, PscJ) Hrp pili protein HrpA HrpE HrpB HrpD General secretion pathway protein F General secretion pathway protein G General secretion pathway protein E General secretion pathway protein D Putative exported protein
2.6 2.5 2.3 2.1 2.0 2.1 2.2 2.2 2.5
2.44E-03 8.49E-04 3.61E-03 4.49E-04 9.25E-05 2.13E-04 2.15E-06 1.37E-05 2.60E-03
3.9 3.8
8.96E-05 1.18E-04
ABF-0016081 ABF-0015971
indC emrA
3.2 3.4
3.76E-05 1.22E-12
ABF-0015970
emrB
3.0
8.00E-08
ABF-0018823 ABF-0019803
sulA uvrA
2.3 2.1
3.23E-04 3.25E-03
ABF-0019309
aer2
Indigoidine biosynthesis protein Similar to phosphoglycolate phosphatase, clustered with ribosomal large-subunit pseudouridine synthase C Indigoidine synthase Membrane fusion component of tripartite multi-drug resistance system Inner membrane component of tripartite multi-drug resistance system SOS cell division inhibitor ATPase and DNA damage recognition protein of nucleotide excision repair excinuclease UvrABC Aerotaxis sensor receptor protein
2.4
2.85E-04
pelC secG hrpN hrcJ
indA indB
(continued) © 2015 The Authors The Plant Journal © 2015 John Wiley & Sons Ltd, The Plant Journal, (2015), 82, 352–362
358 E. Chapelle et al. Table 2. (continued) ASAP ID
Name
Gene description
Fold change (8 hpi/4 hpi)
P value (ANOVA)
ABF-0014520 ABF-0014651 Regulation ABF-0015133 ABF-0015973
cspC ftnA
Stress protein, member of the CspA family Ferritin iron storage protein (cytoplasmic)
2.2 2.2
1.41E-03 6.06E-05
nhaR mprA/emrR
3.7 3.6
1.20E-04 4.96E-08
ABF-0017604 ABF-0019531 ABF-0017487 ABF-0020794 ABF-0020352 ABF-0019270
N/A acrR N/A cueR rpoE pspB
DNA-binding transcriptional activator DNA-binding transcriptional repressor of microcin B17 synthesis and multi-drug efflux GntR family transcriptional regulatory protein DNA-binding transcriptional repressor Organic hydroperoxide resistance transcriptional regulator Copper-responsive transcription regulator RNA polymerase, r24 (rE) factor DNA-binding transcriptional regulator of the psp operon
3.4 2.3 2.2 2.0 2.0 2.0
2.22E-05 1.73E-05 3.54E-03 1.71E-03 2.23E-03 3.99E-06
Gene transcription was measured using NimbleGen microarrays, with two and three replicates for the 4 hpi and 8 hpi time points, respectively.
Protein secretion 13 (8+/5–) Unknown 40 (13+/27–)
Transport 15 (9+/6–) Total: 107 genes (62+/45–) Metabolism 11 (7+/4–)
Others 10 (9+/1–)
Stress 10 (8+/2–)
Figure 3. Functional classification of D. dadantii genes whose expression is modulated at 8 hpi compared with 4 hpi. For each category, the total number of genes is indicated, with the number of up and down-regulated genes in parentheses.
extraction. It provides high-purity bacterial RNA by drastically reducing plant RNA contamination (Table 1) even under conditions where large amounts of plant material must be handled to isolate sufficient bacteria to perform transcriptomic analyses. To ensure that the bacterial transcriptome is not altered during the purification procedure, the whole procedure is performed in presence of a RNA stabilizing reagent. Using quantitative RT-PCR, we assessed the efficiency of RNA stabilization by comparing the accumulation of transcripts of well-known virulence genes either in total RNA extracted from whole infected plants (mixed bacterial and plant RNA) or in RNA extracted from bacteria isolated from infected tissue using our purification protocol (Figure 2). The most common inoculation mode of foliar bacterial pathogens reported in the literature is via syringe
infiltration of a highly concentrated bacterial suspension. This method of inoculation bypasses the initial steps of phyllosphere acclimation/host penetration, and results in a very dense bacterial population residing in a buffer-saturated mesophyll apoplast. In contrast, our immersion-based inoculation procedure mirrors more accurately the natural interaction, and is thus particularly relevant to study the early steps of plant colonization. Accordingly, we used this protocol to analyse the very early steps of Arabidopsis thaliana colonization by D. dadantii. A preliminary microarray analysis comparing bacterial samples recovered by centrifugation either from the bacterial suspension used for inoculation (0 hpi) or the epiphytic population at 6 hpi showed that a quarter of the bacterial genome was differentially expressed between these two conditions (twofold cut-off, false discovery rate < 0.05), including a lot of housekeeping genes (Figure S3). This resembled the huge re-programming reported for Salmonella typhimurium grown either in broth or on agar surfaces, where differential regulation of almost one-third of the functional genome was observed (Wang et al., 2004). Results obtained in this way may therefore reflect the markedly different physiological state of bacteria in the liquid inoculation buffer versus bacteria present on leaves, rather than modulation of gene expression related to the infection process sensu stricto. These data, coupled with the fact that, under our conditions of infection, most tested virulence genes were already activated at 12 hpi (Lebeau et al., 2008; Reverchon et al., 2010; Mhedbi-Hajri et al., 2011), bring us to compare the transcriptome profiles of bacteria present at the leaf surface at 4 hpi, after an epiphytic phase of acclimation, with those of bacteria at the onset of infection but before symptom occurrence at 8 hpi. However, we are aware that this comparison does not allow us to detect genes that may be important in the
© 2015 The Authors The Plant Journal © 2015 John Wiley & Sons Ltd, The Plant Journal, (2015), 82, 352–362
Transcriptome of D. dadantii in planta 359 Table 3 Quantitative RT-PCR analysis of the expression of a subset of D. dadantii genes up-regulated in planta at 8 hpi
Gene
Function
Protein secretion pelC Pectate lyase hrpN Type III secretion hrpA Type III secretion hrpB Type III secretion hrcJ Type III secretion hrpD Type III secretion hrpE Type III secretion Resistance to stress indA Indigoidine biosynthesis indB Indigoidine biosynthesis indC Indigoidine biosynthesis emrR Regulation emrA Tripartite multi-drug resistance emrB Tripartite multi-drug resistance Transport ID18996 ABC transporter ID15684 Predicted toluene transport ID15977 Major facilitator superfamily (MFS) permease protein aer2 Aerotaxis sensor protein
Experiment 1
Experiment 2
Microarrays 8/4 hpi
8/4 hpi
8/4 hpi
0/4 hpi
3.58 3.28 2.62 2.26 2.73 2.10 2.47
4.3 11.7 7.2 5.3 2.0 3.1 3.0
3.3 11.9 12.6 12 6.5 7.3 2.4
0.94 0.42 0.79 1.38 0.77
3.85 3.84 3.21 3.63 3.38 3.04
3.6 4.9 5.9 2.5 5.0 6.0
1.18 0.38 0.91 0.76 0.60
12.2 14.3 * 2.2 4.8 4.3
5.73 3.02 2.89 2.36
18.9 1.7 2.3 2.7
0.45 0.94 1.34 1.13
50.6 1.8 3.0 2.4
0/4 hpi
1.42 0.46 0.47 1.19 1.10
1.21 0.68 0.51 0.65 0.46 0.80 0.61 0.59
Expression data are expressed as the ratio of transcript accumulation at 8 hpi versus transcript accumulation at 4 hpi, or the ratio of transcript accumulation on plant leaves collected just after inoculation (0 hpi) versus transcript accumulation at 4 hpi, normalized against transcript accumulation of the constitutive lpxC gene. *No indC transcripts were detected at 4 hpi.
bacteria–plant interaction already taking place during contact in the phyllosphere. Our analysis revealed that approximately 100 genes were differentially expressed at the onset of infection, with 60% of the genes being up-regulated and 40% being down-regulated. Remarkably, among the 18 down-regulated genes with known function, eight were related to the pectin degradation pathway. They included four stt genes encoding components of a secondary type II secretion system that allows secretion of the pectin lyase PnlH to the outer face of the outer membrane (Ferrandez and Condemine, 2008), and four genes (togM, togN, kdgN and kduD) whose products are involved in transport into the bacterial cell of oligosubunits derived from pectin (Hugouvieux-Cotte-Pattat and Reverchon, 2001; Condemine and Ghazi, 2007; Delangle et al., 2007). These data indicate that genes important for pectin degradation are repressed at the onset of infection. However, we detected up-regulation of other virulence genes very soon after the onset of infection, such as six T3SS hrp genes, including the hrpAB hrcJ hrpDE operon, which encodes HrpA, the major component of the Hrp pilus, and the HrpB and HrpD proteins that have been shown to be T3SS substrates in Pseudomonas syringae and contribute to translocation and secretion of the model AvrPto effector (Ramos et al., 2007). The sixth hrp gene to be up-regulated was the hrpN gene, which encodes the harpin type III effector that has also been shown to be implicated in translocation of
type III effectors into plant cells in Erwinia amylovora (Bocsanczy et al., 2008; Boureau et al., 2011). These results highlight the very early activation of the T3SS in D. dadantii at the start of plant colonization, which may be linked to the well-known role of type III effectors in repression of plant defence responses in many pathosystems (Deslandes and Rivas, 2012). Our transcriptomic analysis also indicated a stressful environment encountered by D. dadantii at the onset of infection. In particular, the bacterium responded to oxidative stress by activating the production of indigoidine, an antioxidant molecule that was shown to protect D. dadantii from H2O2 damage in vitro (Reverchon et al., 2002). It may also protect itself from as yet unknown antibacterial compounds by activating the production of the EmrAB multidrug resistance efflux pump that was shown to confer resistance to hydrophobic toxins both in Escherichia coli and Salmonella typhimurium (Lomovskaya and Lewis, 1992; Nishino et al., 2006). In conclusion, the method described here enables simple, reliable and cost-effective isolation of bacterial cells from plant leaves by differential centrifugation in the presence of a RNA stabilizing reagent, allowing analysis of early steps of the plant–bacteria interaction. Subsequent RNA extraction yields purified bacterial RNA that may be used for microarray-based transcriptomic studies. Our microarray dataset revealed already known virulence
© 2015 The Authors The Plant Journal © 2015 John Wiley & Sons Ltd, The Plant Journal, (2015), 82, 352–362
360 E. Chapelle et al. or adaptation functions that are modulated during the D. dadantii–host plant interaction, confirming the accuracy of our method, and led to identification of candidate genes whose involvement in the interaction will be studied further. Preliminary assays confirmed that our method may be broadened to other plant/bacteria interactions. Thus, this technical breakthrough is of great interest to investigate the mechanisms that operate at the onset of infection and during asymptomatic colonization by bacteria in various plant–bacteria interactions. EXPERIMENTAL PROCEDURES Bacteria, plant material and infection procedure Dickeya dadantii 3937 strain or its GFP+ derivative (Asselbergh et al., 2008) were grown at 30°C overnight on Luria–Bertani (LB) plates. Bacteria were resuspended in 50 mM KPO4 buffer, pH 6.8, containing 0.01% v/v Silwet L-77 surfactant (van Meeuwen Chemicals, www.vanmeeuwen.com/van-meeuwen-chemicals/). Col-0 Arabidopsis plants were grown for 6 weeks under short-day conditions at 24/19°C (8 h light/16 h dark). Six-week-old plants (50–70 per time point) were inoculated by rapid immersion in a bacterial suspension at 108 colony-forming units per ml, and were incubated at 24/ 19°C (day/night) under long-day conditions (16 h light) in small transparent enclosed containers (length 51 cm, width 38.5 cm, height 36 cm; Bouillard Freres SA, www.bouillard.fr/) with abundant watering to achieve 100% humidity. The long-day conditions allowed us to collect bacteria at all time points under the same conditions of temperature and light.
Isolation of bacteria from infected plants At various time points after inoculation, plant aerial parts were collected in 50 mM KPO4 buffer, pH 6.8, supplemented with the GenelockTM/AssayAssureTM RNA stabilizing reagent (Sierra Molecular Inc., www.sierramolecular.com/) to a final 20% concentration, hereafter referred to as RNA freezing buffer. Plant tissues were disrupted using a blender, and the homogenate was filtered through 25 lm and then 10 lm cheesecloth to remove large plant tissue debris. The flow throughs were further cleaned by centrifugation at 500 g for 10 min, and the bacterial cells were collected by centrifuging the resulting supernatant at 10 000 g for 15 min. The pellet, consisting of bacteria contaminated with plant material, was resuspended in 3 ml RNA freezing buffer supplemented with GentodenzTM resin (Gentaur France SARL, www.gentaur-worldwide.com) to a final concentration of 50% w/v. A Gentodenz density gradient was mounted above the sample by slowly depositing various layers of Gentodenz-containing RNA freezing buffer (2 ml of 30% Gentodenz, 3 ml of 25% Gentodenz, 2 ml of 10% Gentodenz), and the gradient was centrifuged for 1 h at 12 000 g in swinging bucket rotor like Sorvall HB-6. Bacteria, appearing as an opalescent whitish band, were recovered by pipetting. When many infected plants were treated, the bacterial ring was sometimes not easy to visualize (see Figure 1c). In these cases, we used a gradient in which the bacterial ring is clearly visible as a reference to determine the part of the gradient to be collected. These reference gradients consisted either of samples at late time points (Figure 1c, left) or gradients comprising a suspension of pure bacteria mixed with limited plantlet material. After centrifugation at 13 000 g for
5 min, the Gentodenz resin was removed by two rinses in RNA freezing buffer. If the bacterial pellet was still covered by a densely packed green layer formed by remaining plant contaminants, this layer was loosened by gently pipetting the buffer up and down, and then discarding it. Finally, bacteria were resuspended in 200 ll of ethanol/phenol stop solution (5% water-saturated phenol in ethanol), recovered by quick centrifugation, (16 000 g in a table-top centrifuge) frozen in liquid nitrogen and stored at 80°C. Bacteria present on leaf surfaces were recovered by rinsing the inoculated leaves in RNA freezing buffer, filtering the bacterial suspension through 10 lm cheesecloth to remove contaminating soil, and recovering the bacterial pellet by centrifugation at 10 000 g for 15 min. For a better comparison with the samples extracted from plant tissues, these samples were then processed through a Gentodenz gradient as described above.
RNA extraction RNA from purified bacteria was isolated by a hot phenol procedure. The bacterial pellet was resuspended in 600 ll of ice-cold buffer A (20 mM sodium acetate, pH 5.2, 1 mM EDTA). Cells were immediately lysed by addition of 40 ll of 10% SDS and 600 ll of hot acidic phenol (65°C), vortexed, and incubated for 6–8 min at 65°C. RNA was further cleaned by two additional phenol/chloroform extractions and ethanol-precipitated. The isolation of total RNA of infected plants was performed as described by Lebeau et al. (2008). Briefly, aerial plant tissues were collected at various time points post-inoculation and ground in liquid N2 to a fine powder. RNAs were extracted in guanidium isothiocyanate extraction buffer and pelleted by centrifugation (420 000 g for 5 h at 20°C) on a caesium chloride cushion. Pelleted RNAs were washed twice with ice-cold 70% ethanol and resuspended in diethylpyrocarbonate-treated H2O.
Expression analysis RNA samples were treated with RNase-free DNase I (Invitrogen, https://www.lifetechnologies.com/uk/en/home/brands/invitrogen. html) to remove any DNA contamination. First-strand cDNAs were then synthesized from 3 lg total RNA using Superscript II RNAse H reverse transcriptase (Invitrogen), and random hexamer primers (Invitrogen) according to the manufacturer’s instructions. For quantitative RT-PCR analysis, cDNAs were amplified using Maximaâ SYBR Green/ROX qPCR Master Mix (Fermentas, https:// www.lifetechnologies.com/fr/fr/home/brands/thermo-scientific. html) according to the manufacturer’s instructions in an Applied Biosystems (http://www.lifetechnologies.com/fr/fr/ home/brands/applied-biosystems.html) 7300 real-time PCR system under the following conditions: 10 min at 95°C, followed by 40 amplification cycles each consisting of 15 sec at 95°C and 60 sec at 60°C. The results were analysed using Applied Biosystems sequence detection software version 1.3.1. Specific PCR primers are listed in Table S2. The constitutive lpxC gene encoding an UDP-N-acetylglucosamine deacetylase involved in lipid A biosynthesis was used as internal control to normalize the expression data (Hommais et al., 2011). No amplification of this gene was detected by quantitative PCR using cDNA derived from RNA of uninoculated plants. The comparative quantification method (DDCT ) was used to compare the various conditions (Livak and Schmittgen, 2001) as described previously (Mhedbi-Hajri et al., 2011). All assays were performed in duplicate (biological replicates) to control for overall variability.
© 2015 The Authors The Plant Journal © 2015 John Wiley & Sons Ltd, The Plant Journal, (2015), 82, 352–362
Transcriptome of D. dadantii in planta 361 Microarray design and analysis The microarrays used in this study were custom designed and produced by NimbleGen Systems Inc. (http://www.nimblegen. com/) and based on the annotated sequence (version 6) of D. dadantii (Glasner et al., 2011; available at https://asap.ahabs. wisc.edu/asap/home.php). These arrays comprised 4753 coding sequences. The microarrays consisted of five probes of 60 nucleotides per coding sequence and three copies of each probe per array separated in three blocks. For microarray analyses, double-strand cDNA was synthesized according to the NimbleGen protocol using random hexamer primers for first-strand cDNA synthesis. cDNAs were labelled and hybridized by NimbleGen Systems Inc. One-colour NimbleGen expression data were processed using the ANAIS web interface (Simon and Biot, 2010; http://anais.versailles.inra.fr/). Normalization was performed using the robust multi-array analysis background correction process. To identify differentially expressed genes, fold change was calculated for each experimental condition, as compared to the referent experimental condition, and a one-way ANOVA test was performed on all log10-transformed normalized data. The raw P values were adjusted using the false discovery rate/Bonferroni method (type I error = 5%) to control for family-wise error rate and to drastically limit false positives in a multiple comparison context (Benjamini and Hochberg, 1995).
ACKNOWLEDGEMENTS We would like to thank the members of the Lyon Dickeya group (UMR5240 Microbiologie, Adaptation et Pathogenie) for their helpful discussions, Marie-Anne Barny for her critical reading of the manuscript and suggestions, and Caroline Kunz for the English corrections. This work was supported by a REGUPATH grant from the French ‘ANR blanc 2007’ Programme (ANR-07-BLAN-0212).
SUPPORTING INFORMATION Additional Supporting Information may be found in the online version of this article. Figure S1. Microscopic visualization of D. dadantii infection kinetics. Figure S2. Growth kinetics of D. dadantii upon Arabidopsis infection. Figure S3. Functional classification of D. dadantii genes whose expression is modulated at 6 hpi compared with 0 hpi. Table S1. D. dadantii genes whose expression is modulated at least twofold 8 at hpi compared with 4 hpi. Table S2. Primers used for expression studies.
REFERENCES € fte, M. and Van Gijsegem, F. (2008) Asselbergh, B., Enow Achuo, A., Ho Abscisic acid-deficiency leads to rapid activation of tomato defence responses upon infection with Erwinia chrysanthemi. Mol Plant Pathol. 9, 11–24. Audia, J.P., Patton, M.C. and Winkler, H.H. (2008) DNA microarray analysis of the heat shock transcriptome of the obligate intracytoplasmic pathogen Rickettsia prowazekii. Appl. Environ. Microbiol. 74, 7809–7812. Babujee, L., Apocada, J., Balakrishnan, V., Liss, P., Kiley, P.J., Charkowski, A.O., Glasner, J.D. and Perna, N.T. (2012) Evolution of metabolic and regulatory networks associated with oxygen availability in two phytopathogenic enterobacteria. BMC Genom. 13, 110. Barras, F., Van Gijsegem, F. and Chatterjee, A.K. (1994) Extracellular enzymes and pathogenesis of soft-rot Erwinia. Ann Rev Phytopathol. 32, 201–234.
Bauer, D.W., Bogdanove, A.J., Beer, S.V. and Collmer, A. (1994) Erwinia chrysanthemi hrp genes and their involvement in soft rot pathogenesis and elicitation of the hypersensitive response. Mol Plant-Microbe Interact. 7, 573–581. Bauer, D.W., Wei, Z.M., Beer, S.W. and Collmer, A. (1995) Erwinia chrysanthemi harpinEch: an elicitor of the hypersensitive response that contributes to soft-rot pathogenesis. Mol Plant-Microbe Interact. 8, 484–491. Benjamini, Y. and Hochberg, Y. (1995) Controlling the false discovery rate: a practical and powerful approach to multiple testing. J. Roy. Stat. Soc. B 57, 289–300. Boch, J., Joardar, V., Gao, L., Robertson, T.L., Lim, M. and Kunkel, B.N. (2002) Identification of Pseudomonas syringae pv. tomato genes induced during infection of Arabidopsis thaliana. Mol. Microbiol. 44, 73–88. Bocsanczy, A.M., Nissinen, R.I., Oh, C.S. and Beer, S.V. (2008) HrpN of Erwinia amylovora functions in the translocation of DspA/E into plant cells. Mol Plant Pathol. 9, 425–434. Boureau, T., Siamer, S., Perino, C., Gaubert, S., Patrit, O., Degrave, A., Fagard, M., Chevreau, E. and Barny, M.A. (2011) The HrpN effector of Erwinia amylovora, which is involved in type III translocation, contributes directly or indirectly to callose elicitation on apple leaves. Mol PlantMicrobe Interact. 24, 577–584. Bronstein, P., Filiatrault, M., Myers, C., Rutzke, M., Scneider, D. and Cartinhour, S. (2008) Global transcriptional responses of Pseudomonas syringae DC3000 to changes in iron bioavailability in vitro. BMC Microbiol. 8, 209. Brown, D.G. and Allen, C. (2004) Ralstonia solanacearum genes induced during growth in tomato: an inside view of bacterial wilt. Mol. Microbiol. 53, 1641–1660. Camejo, A., Buchrieser, C., Couve, E., Carvalho, F., Reis, O., Ferreira, P., Sousa, S., Cossart, P. and Cabanes, D. (2009) In Vivo transcriptional profiling of Listeria monocytogenes and mutagenesis identify new virulence factors involved in infection. PLoS Pathog. 5, e1000449. Capela, D., Filipe, C., Bobilk, C., Batut, J. and Bruand, C. (2006) Sinorhizobium meliloti differentiation during symbiosis with alfalfa: a transcriptomic dissection. Mol Plant-Microbe Interact. 19, 363–372. Charkowski, A., Blanco, C., Condemine, G. et al. (2012) The role of small molecules in soft rot Enterobacteriaceae pathogenicity. Annu. Rev. Phytopathol. 50, 425–449. Condemine, G. and Ghazi, A. (2007) Differential regulation of two oligogalacturonate outer membrane channels, KdgN and KdgM, of Dickeya dadantii (Erwinia chrysanthemi). J. Bacteriol. 189, 5955–5962. Condemine, G., Castillo, A., Passeri, F. and Enard, C. (1999) The PecT repressor coregulates synthesis of exopolysaccharides and virulence factors in Erwinia chrysanthemi. Mol Plant-Microbe Interact. 12, 45–52. Costechareye, D., Chich, J.F., Strub, J.M., Rahbe, Y. and Condemine, G. (2013) Transcriptome of Dickeya dadantii infecting Acyrthosiphon pisum reveals a strong defense against antimicrobial peptides. PLoS ONE, 8, e54118. Delangle, A., Prouvost, A.-F., Cogez, V., Bohin, J.-P., Lacroix, J.-M. and Hugouvieux Cotte-Pattat, N. (2007) Characterization of the Erwinia chrysanthemi gan Locus, Involved in Galactan Catabolism. J. Bacteriol. 189, 7053–7061. Deslandes, L. and Rivas, S. (2012) Catch me if you can: bacterial effectors and plant targets. Trends Plant Sci. 17, 644–655. El-Hassouni, M.E., Chambost, J.P., Expert, D., Van Gijsegem, F. and Barras, F. (1999) The minimal gene set member msrA, encoding peptide methionine sulfoxide reductase, is a virulence determinant of the plant pathogen Erwinia chrysanthemi. Proc. Natl Acad. Sci. USA, 96, 887–892. Expert, D. (1999) Withholding and exchanging iron: interactions between Erwinia spp. and their plant hosts. Annu. Rev. Phytopathol. 37, 307– 334. Ferrandez, Y. and Condemine, G. (2008) Novel mechanism of outer membrane targeting of proteins in Gram-negative bacteria. Mol. Microbiol. 69, 1349–1357. Franza, T., Mahe, B. and Expert, D. (2005) Erwinia chrysanthemi requires a second iron transport route dependent of the siderophore achromobactin for extracellular growth and plant infection. Mol. Microbiol. 55, 261– 275. Gingeras, T.R. and Rosenow, C. (2000) Studying microbial genomes with high-density oligonucleotide arrays. ASM News, 8, 463–469.
© 2015 The Authors The Plant Journal © 2015 John Wiley & Sons Ltd, The Plant Journal, (2015), 82, 352–362
362 E. Chapelle et al. Glasner, J.D., Yang, C.-H., Reverchon, S. et al. (2011) Genome sequence of the plant-pathogenic bacterium Dickeya dadantii 3937. J. Bacteriol. 193, 2076–2077. Gouesbet, G., Trautwetter, A., Bonnassie, S., Wu, L.F. and Blanco, C. (1996) Characterization of the Erwinia chrysanthemi osmoprotectant transporter gene ousA. J. Bacteriol. 178, 447–455. Graham, J.E. and Clark-Curtiss, J.E. (1999) Identification of Mycobacterium tuberculosis RNAs synthesized in response to phagocytosis by human macrophages by selective capture of transcribed sequences (SCOTS). Proc. Natl Acad. Sci. USA, 96, 11554–11559. Hinton, J.C.D., Hautefort, I., Eriksson, S., Thompson, A. and Rhen, M. (2004) Benefits and pitfalls of using microarrays to monitor bacterial gene expression during infection. Curr. Opin. Microbiol. 7, 277–282. Hommais, F., Zghidi-Abouzid, O., Oger-Desfeux, C., Pineau-Chapelle, E., Van Gijsegem, F., Nasser, W. and Reverchon, S. (2011) lpxC and yafS are the most suitable internal controls to normalize real time RT-qPCR expression in the phytopathogenic bacterium Dickeya dadantii. PLoS ONE, 6, e20269. Hugouvieux-Cotte-Pattat, N. and Reverchon, S. (2001) Two transporters TogT and TogMNAB are responsible for oligogalacturonide uptake in Erwinia chrysanthemi 3937. Mol. Microbiol. 41, 1125–1132. Jacobs, J.M., Babujee, L., Meng, F., Milling, A. and Allen, C. (2012) The in planta transcriptome of Ralstonia solanacearum: conserved physiological and virulence strategies during bacterial wilt of Tomato. mBio, 3: e00114-12. Koide, T., Vencio, R. and Gomes, S. (2006) Global gene expression analysis of the heat shock response in the phytopathogen Xylella fastidiosa. J. Bacteriol. 188, 5821–5830. La, M.-V., Raoult, D. and Renesto, P. (2008) Regulation of whole bacterial pathogen transcription within infected hosts. FEMS Microbiol. Rev. 32, 440–460. Lebeau, A., Reverchon, S., Gaubert, S., Kraepiel, Y., Simond-Cote, E., Nasser, W. and Van Gijsegem, F. (2008) The GacA global regulator is required for the appropriate expression of Erwinia chrysanthemi 3937 pathogenicity genes during plant infection. Environ. Microbiol. 10, 545– 559. Littlejohn, G.R., Gouveia, J.D., Edner, C., Smirnoff, N. and Love, J. (2010) Perfluorodecalin enhances in vivo confocal microscopy resolution of Arabidopsis thaliana mesophyl. New Phytol. 186, 1018–1025. Livak, K.J. and Schmittgen, T.D. (2001) Analysis of relative gene expression data using real-time quantitative PCR and the 2 DDCT method. Methods, 25, 402–408. Lomovskaya, O. and Lewis, K. (1992) Emr, an Escherichia coli locus for multidrug resistance. Proc. Natl Acad. Sci. USA, 89, 8938–8942. Lopez-Solanilla, E., Garcia-Olmedo, F. and Rodriguez-Palenzuela, P. (1998) Inactivation of the sapA to sapF locus of Erwinia chrysanthemi reveals common features in plant and bacterial pathogenesis. Plant Cell, 10, 917–924. Ma, B., Hibbing, M.E., Kim, H.-S. et al. (2007) The host range and molecular phylogenies of the soft rot enterobacterial genera Pectobacterium and Dickeya. Phytopathology, 97, 1150–1163. Mhedbi-Hajri, N., Malfatti, P., Pedron, J., Gaubert, S., Reverchon, S. and Van Gijsegem, F. (2011) PecS is an important player in the regulatory network governing the coordinated expression of virulence genes during the interaction between Dickeya dadantii 3937 and plants. Environ. Microbiol. 13, 2901–2914. Nishino, K., Latifi, T. and Groisman, E.A. (2006) Virulence and drug resistance roles of multidrug efflux systems of Salmonella enterica serovar Typhimurium. Mol. Microbiol. 59, 126–141.
Perombelon, M.C.M. (2002) Potato diseases caused by soft rot erwinias: an overview of pathogenesis. Plant. Pathol. 51, 1–12. Pessi, G., Ahrens, C.H., Rehrauer, H., Lindemann, A., Hauser, F., Fischer, H.M. and Hennecke, H. (2007) Genome-wide transcript analysis of Bradyrhizobium japonicum bacteroids in soybean root nodules. Mol PlantMicrobe Interact. 20, 1353–1363. Ramos, A.R., Morello, J.E., Ravindran, S., Deng, W.-L., Huang, H.-C. and Collmer, A. (2007) Identification of Pseudomonas syringae pv. syringae 61 Type III Secretion System Hrp proteins that can travel the Type III pathway and contribute to the translocation of effector proteins into plant cells. J. Bacteriol. 189, 5773–5778. Rauhut, R. and Klug, G. (1999) mRNA degradation in bacteria. FEMS Microbiol. Rev. 23, 353–370. Reverchon, S., Rouanet, C., Expert, D. and Nasser, W. (2002) Characterization of indigoidine biosynthetic genes in Erwinia chrysanthemi and role of this blue pigment in pathogenicity. J. Bacteriol. 184, 654–665. Reverchon, S., Van Gijsegem, F., Effantin, G., Zghidi-Abouzid, O. and Nasser, W. (2010) Systematic targeted mutagenesis of the MarR/SlyA family members of Dickeya dadantii 3937 reveals a role for MfbR in the modulation of virulence gene expression in response to acidic pH. Mol. Microbiol. 78, 1018–1037. Rojas, C.M., Ham, J.H., Deng, W.L., Doyle, J.J. and Collmer, A. (2002) HecA, a member of a class of adhesins produced by diverse pathogenic bacteria, contributes to the attachment, aggregation, epidermal cell killing, and virulence phenotypes of Erwinia chrysanthemi EC16 on Nicotiana clevelandii seedlings. Proc. Natl Acad. Sci. USA, 99, 13142–13147. Santos, R., Franza, T., Laporte, M.L., Sauvage, C., Touati, D. and Expert, D. (2001) Essential role of superoxide dismutase on the pathogenicity of Erwinia chrysanthemi strain 3937. Mol Plant-Microbe Interact. 14, 758–767. Schoonejans, E., Expert, D. and Toussaint, A. (1987) Characterization and virulence properties of Erwinia chrysanthemi lipopolysaccharide-defective, phi EC2-resistant mutants. J. Bacteriol. 169, 4011–4017. Serrania, J., Vorho€ lter, F.-J., Niehaus, K., P€uhler, A. and Becker, A. (2008) Identification of Xanthomonas campestris pv. campestris galactose utilization genes from transcriptome data. J. Biotechnol. 135, 309–317. Simon, A. and Biot, E. (2010) ANAIS: Analysis of NimbleGen Arrays Interface. Bioinformatics, 26, 2468–2469. Soto-Suarez, M., Bernal, D., Gonzalez, C., Szurek, B., Guyot, R., Tohme, J. and Verdier, V. (2010) In planta gene expression analysis of Xanthomonas oryzae pathovar oryzae, African strain MAI1. BMC Microbiol. 10, 170. Toledo-Arana, A., Dussurget, O., NIkitas, G. et al. (2009) The Listeria transcriptional landscape from saprophytism to virulence. Nature, 459: 950– 956. Wang, Q., Frye, J.G., McClelland, M. and Harshey, R.M. (2004) Gene expression patterns during swarming in Salmonella typhimurium: genes specific to surface growth and putative new motility and pathogenicity genes. Mol. Microbiol. 52, 169–187. Yang, C.-H., Gavilanes-Ruiz, M., Okinaka, Y., Vedel, R., Berthuy, I., Boccara, M., Chen, J.W., Perna, N.T. and Keen, N.T. (2002) hrp genes of Erwinia chrysanthemi 3937 are important virulence factors. Mol Plant-Microbe Interact. 15, 472–480. Yang, S., Perna, N.T., Cooksey, D.A., Okinaka, Y., Lindow, S.E., Ibekwe, A.M., Keen, N.T. and Yang, C.-H. (2004) Genome-wide identification of plant-upregulated genes of Erwinia chrysanthemi 3937 using a GFPbased IVET leaf array. Mol Plant-Microbe Interact. 17, 999–1008. Yu, X., Lund, S.P., Scott, R.A., Greenwald, J.W., Records, A.H., Nettleton, D., Lindow, S.E., Gross, D.C. and Beattie, G.A. (2013) Transcriptional responses of Pseudomonas syringae to growth in epiphytic versus apoplastic leaf sites. Proc. Natl Acad. Sci. USA, 110, E425–E434.
© 2015 The Authors The Plant Journal © 2015 John Wiley & Sons Ltd, The Plant Journal, (2015), 82, 352–362
