An image classification approach to analyze the suppression of plant immunity by the human pathogen SalmonellaTyphimurium
© Schikora et al.; licensee BioMed Central Ltd. 2012
Received: 31 January 2012
Accepted: 11 May 2012
Published: 19 July 2012
The enteric pathogen Salmonella is the causative agent of the majority of food-borne bacterial poisonings. Resent research revealed that colonization of plants by Salmonella is an active infection process. Salmonella changes the metabolism and adjust the plant host by suppressing the defense mechanisms. In this report we developed an automatic algorithm to quantify the symptoms caused by Salmonella infection on Arabidopsis.
The algorithm is designed to attribute image pixels into one of the two classes: healthy and unhealthy. The task is solved in three steps. First, we perform segmentation to divide the image into foreground and background. In the second step, a support vector machine (SVM) is applied to predict the class of each pixel belonging to the foreground. And finally, we do refinement by a neighborhood-check in order to omit all falsely classified pixels from the second step. The developed algorithm was tested on infection with the non-pathogenic E. coli and the plant pathogen Pseudomonas syringae and used to study the interaction between plants and Salmonella wild type and T3SS mutants. We proved that T3SS mutants of Salmonella are unable to suppress the plant defenses. Results obtained through the automatic analyses were further verified on biochemical and transcriptome levels.
This report presents an automatic pixel-based classification method for detecting “unhealthy” regions in leaf images. The proposed method was compared to existing method and showed a higher accuracy. We used this algorithm to study the impact of the human pathogenic bacterium Salmonella Typhimurium on plants immune system. The comparison between wild type bacteria and T3SS mutants showed similarity in the infection process in animals and in plants. Plant epidemiology is only one possible application of the proposed algorithm, it can be easily extended to other detection tasks, which also rely on color information, or even extended to other features.
Numerous bacteria, pathogenic to humans and other mammals, are found to thrive also on plants, Salmonella enterica, Pseudomonas aeruginosa, Burkholderia cepacia, Erwinia spp., Staphylococcus aureus, Escherichia coli O157:H7, and Listeria monocytogenes are able to infect both animal and plant organisms [1–5]. Among these, Salmonella, a genus of Gram-negative enteropathogenic bacteria, are the causal agents of both gastroenteritis and typhoid fever. They are responsible for an estimated one million casualties and about 100 million human infections annually. Not only in developing countries in Africa or South-East Asia, where typhoid and paratyphoid fever are unfortunately still prevalent, but also in developed communities salmonellosis is still not vanquished. The most common mode of infection in humans is by ingestion of contaminated food or water.
Plants can be the source of infection
Many reports have linked food poisoning with the consumption of Salmonella-contaminated raw vegetables and fruits (for review see [2, 6]). A large study conducted in the European Union revealed that in 2007, 0.3% of products were infected with Salmonella bacteria , during the same time in UK, the Netherlands, Germany, and Ireland 0.1 to 2.3% of pre-cut products were contaminated . In the USA, the proportion of raw food-associated salmonellosis outbreaks increased from 0.7% in the 1960s to 6% in the 1990s , and crossed 25% in recent years . In order to monitor the molecular subtype pattern of the outbreak strains a national program (PulseNet) was created in the USA . This program significantly improved the identification of outbreaks and their sources. Most studies on Salmonella-plant interactions suggested an epiphytic lifestyle of Salmonella on plants. However, a growing body of evidence points to an active process in which bacteria infect various plants and use them as viable hosts [11–20]. In this report we developed an automatic algorithm to quantify the symptoms caused by Salmonella infection on Arabidopsis plants. The algorithm is designed to attribute image pixels into one of the two classes: healthy and unhealthy. We show that it outperforms other algorithms developed for this task. It was tested on infection with the non-pathogenic E. coli and the plant pathogen Pseudomonas syringae and subsequently used to study the interaction between plant host and Salmonella wild type and T3SS mutants. We proved that T3SS mutants of Salmonella are unable to suppress the plant defense mechanisms. Results obtained through the automatic analyses were further verified on biochemical and transcriptome levels.
Automatic classification as key concept to objective analysis
During the last few years, image classification has proved increasingly useful in biology, as numerous tasks have been simplified with the help of automated image classification [21–23]. Plant diseases need to be controlled for at least two reasons: to maintain the quality of food produced by farmers around the world and in order to reduce the food-borne illnesses originated from infected plants . Thus, automatic identification of “unhealthy” regions in leaf images is a useful tool for various biological research projects aiming the control of diseases or characterization of plant defense mechanisms [25, 26]. There is a wide variety of plant diseases caused by either environmental factors (nutrition, moisture, temperature, etc.) or by other organisms (fungi, bacteria, viruses). However, in most cases the common symptom is the change of the leaf color. A good color variation model can be employed to distinguish “healthy” and “unhealthy” regions in leaf images. A probabilistic algorithm, employing a Gaussian mixture model (GMM) and a Bayesian classifier to classify disease symptoms in Arabidopsis plants was presented in . However, because the estimation of a robust GMM is not always possible from the real data, results from Bayes-like classifiers can be inaccurate. To overcome this limitation we propose a different classification strategy. The algorithm described in this report uses color feature space as input for learning algorithm (Support Vector Machine (SVM)) which classifies the pixels of leaf images.
Type III secretion system is responsible for effectors delivery
Salmonellosis develops after the bacteria enter epithelial cells of the intestine . Studies of the infection mechanisms in animals have shown that Salmonella actively remodel the host cells physiology and architecture, and suppress the host immune system by injecting a cocktail of effectors delivered by Type III Secretion Systems (T3SSs). Salmonella enterica subsp. enterica has two distinct T3SSs, T3SS-1 and T3SS-2, encoded by the Salmonella Pathogenicity Islands (SPI) SPI-1 and SPI-2, respectively [29, 30]. T3SS-1 secretes at least 16 proteins of which 6 were shown to interact with the host signaling cascades and the cytoskeleton. T3SS-2 secretes at least 19 Salmonella enterica-specific effector proteins that are involved in survival and multiplication within the host cell [31, 32]. The expression and the secretion of SPI-1 and SPI-2 encoded effectors are tightly regulated. Recently, a sorting platform for T3SS effectors was reported that determines the appropriate hierarchy for protein secretion . In this study, the authors identified the cytoplasmic SpaO-OrgA-OrgB complex, which enables the sequential delivery of translocases before the secretion of the actual effectors. Furthermore, the authors described the role of specific chaperones in the recognition and loading of effectors into the sorting SpaO-OrgA-OrgB complex. In conclusion, it was postulated that similar sorting platforms might exist in other T3SSs as their components are widely conserved. Many recent reports suggest that the mechanisms used by Salmonella to infect animal and plant hosts might be similar [20, 34].
Effector proteins defeat immune system
In the battle between pathogen and its host, the pathogen needs to suppress the host immune system in order to establish a successful infection. The early line of immunity relies on the recognition of conserved pathogen-associated molecular patterns (PAMPs) by host-encoded pattern recognition receptors (PRRs) and thereby the activation of an array of defense responses called PAMP-triggered immunity (PTI). The best-studied PAMP in plants is flg22, a conserved 22 amino acid peptide from the bacterial flagellar protein flagellin, recognized by the PRR FLAGELLIN INSENSITIVE 2 (FLS2) . During infection, pathogens secrete effectors with the aim to suppress PTI and cause effector-triggered susceptibility (ETS). In a second layer of defense, intracellular resistance proteins (R-proteins) recognize pathogen effectors and activate effector-triggered immunity (ETI). The plant pathogen Pseudomonas syringae injects about 40 effectors into plant cells. Among these, AvrPto, AvrPtoB and HopAI1 attenuate the flg22-induced defense responses [36–38]. Strikingly, HopAI1 is also present in animal/human pathogens such as Shigella spp. (OspF) [39, 40] and Salmonella spp. (SpvC) , where it interacts with the mitogen-activated protein kinases (MAPKs) ERK1/2 and p38. The role of multiple Salmonella effectors in animal infection has been described (reviewed in ), but a functional proof of Salmonella effector action in plants is still missing. Nonetheless, several lines of evidence point to an active interaction between these bacteria and plant hosts.
Salmonellasuppresses plant defenses
Two very recent studies report the suppression of the plant immune system by Salmonella[34, 43]. The authors showed that in contrast to wild type living bacteria, dead or chloramphenicol treated bacteria elicited an oxidative burst and pH changes in tobacco cells. A similar response was provoked by the inv A− mutant, which has no functional SPI-1 T3SS . Those results suggest that Salmonella depends on the secretion of effectors during plant infection and actively suppresses the immune response. We observed similar phenomena during infection of Arabidopsis. Salmonella T3SS mutants were compromised in virulence towards the wild type Col-0 plants. Comparison between global transcriptome profiles of Arabidopsis plants infected with wild type Salmonella or the prg H−(T3SS-1) mutant revealed 649 genes, which are upregulated upon challenge with prg H− mutant but not with the wild type Salmonella. GO term enrichment analysis (AmiGO version 1,7)  of these 649 prg H−-specific genes showed an overrepresentation of genes related to responses to biotic stress, relations with other organisms and defense mechanisms . Moreover, challenge with T3SS mutants provoked stronger symptoms on Arabidopsis plants suggesting that those mutants are not able to suppress plant defenses. Those symptoms could be, at least to some extent, part of the hypersensitivity response (HR). HR is a common defense mechanism against biotrophic and hemibiotrophic pathogens, resulting in localized cell death and therefore arresting the proliferation of pathogen. However, successful pathogenic bacteria evolved mechanisms to suppress this resistance mechanism. In a simplified manner one could describe a very fast and strong occurrence of chlorotic and dead tissues after infection with Salmonella as resistance mechanism. On the other hand, necrotic and lysed tissues suggest no resistance capabilities. This distinction served as the base for an automatic analysis of infection symptoms caused by wild type Salmonella and four distinct T3SS mutants as well as the plant pathogenic Pseudomonas syringae and the nonpathogenic E. coli.
denoting a weighted squared sum of the individual channels. For the results presented in this paper we used wI1 = 0.1 and wI2 = wI3 = 0.45. As an additional input we used mean values for the foreground μobjand background μbgd and a smoothing parameter . [I123(x)]In denotes the value of pixel x for the color channel I n . The desired segmentation is a binary image . We minimize (1) for real-valued u using successive over-relaxation (SOR), as in [49, 51] and binarize the solution to obtain the globally optimal segmentation.
Having obtained a binary image , we classified each pixel belonging to Ωobj into “unhealthy” or “healthy” regions. For this purpose we use a state-of-the-art machine-learning algorithm, support vector machine (SVM), that have found a wide acceptance in recent years due to its ability to classify linear and non-linear data. SVMs have been applied with great success in many challenging classification problems processing large data sets. The basic concept was introduced in . In our work we will use a modified maximum margin idea, called Soft Margin, which allows the handling of not perfectly linear separable data. It is based on learning from examples, which means, it requires a separate set of training and testing data. The training algorithm builds a model that predicts the class of unknown input data.
We needed a labeled training data, which serves as an input for the learning function. For training we chose 40.000 pixels of leaf images randomly from all available images. Then we hand-labeled every chosen pixel into one of three classes: healthy, unhealthy and background. Like many other pixel-based classification methods, we exploit the color variation property of image co-ordinates in order to form a decision model. Since the components of I1I2I3 color space  are uncorrelated, statistically it is the best way to detect color variations. While I1 contains the illumination information, I2 and I3 mainly contain color information. Hence, we used only I2 and I3 in order to provide invariance to illumination changes. Thus the training data comprise of 2D color values, selected from “healthy” and “unhealthy” leaf images and labeled into the two different classes.
Training phase - offline
Where S denotes the set of indices of the support vectors. S is determined by finding the indices i where α i > 0. Instead of using an arbitrary support vector x s , it is better to take an average of the support vectors in S. Thus, the training phase of SVM gives w and b which is used later to compute the class of unknown vectors. Since the training phase is time consuming, it is done offline.
Prediction phase - online
where w and b are obtained from the training part of the SVM algorithm.
In addition, we split 9797 data points from the labeled training set and classified this data to get an objective performance measure. The GMM approach reached a correct classification rate of 91.5%. The proposed SVM approach could improve the results, so that a correct classification rate of 95.8% could be achieved.
Photo-based analysis of symptoms caused by different bacteria in Arabidopsis
T3SS mutants cause stronger symptoms than the wild type bacteria
T3SS mutants cannot suppress the induction of the pathogenesis-related gene PDF1.2
Infection with T3SS mutant results in longer activation of MAP kinases
Plants have sophisticated mechanisms by which they recognize pathogen-originated signals. In case of pathogen attack, plants might initiate a rapid and intense activation of defense reactions known as hypersensitive response (HR). HR occurs within few hours and results in localized cell death. Very often HR is the consequence of effector-triggered immunity (ETI), which occurs when the plant recognizes the effectors injected by the pathogen into the plant cells. Rapid cell death or HR prevents the bacteria from spreading systematically. Salmonella uses diverse effectors to manipulate the cellular signals leading to the host defense response . Salmonella enterica subsp. enterica used in this study possesses two different T3SS, encoded by Salmonella Pathogenicity Island 1 (SPI-1) and SPI-2. Both T3SSs secret different yet overlapping sets of effector proteins tat function at different stages of the infection. However, many of the secreted effectors can by translocated via both T3SSs. The stronger symptoms seen in the leaves treated with the T3SSs mutants if compared to the wild type Salmonella, indicates the inability of Salmonella mutants to inhibit the molecular mechanisms that finally lead to HR, and in consequence it suggests the necessity of such effectors (and both functional T3SSs) for the infection of vegetal hosts. It is probable that both T3SSs are needed for the immune suppression, however the effectors translocated by the remaining T3SS in a mutant are sufficient to elicit ETI. Giving the importance for human health, the suppression of the animal immune system by Salmonella is very intensely studied. We know already 44 effectors which are injected by Salmonella into animal host cells, and for many of them we know the function and the target proteins . Interestingly, very often bacterial effectors target the MAPK cascades, which are important regulators of the immune response in animals and plants. SpvC from Salmonella spp. encodes a phosphothreonine lyase that dephosphorylates the pTXpY double phosphorylated activation loop in the ERK1/2 kinases [61–63]. Another effector from Salmonella spp. the SptP inhibits phosphorylation and membrane localization of Raf kinase and therefore the activation of the downstream ERK kinases . Although several Salmonella effectors have homologues in plant pathogenic bacteria, the SpvC is present in the Pseudomonas spp. as HopAI1, HopAO1 also from Pseudomonas spp. on the other hand, is the homologue of SptP, the function of Salmonella proteins in the inactivation of the plant immune system remains unknown. It is however very tempting to speculate that biochemical features of those effectors are conserved between animal and plant hosts, providing Salmonella (and other pathogenic bacteria) with efficient tools for suppression of the host immune system. Such suppression was reported in two recent reports. Shirron and Yaron studied infection of tobacco plants with S. Typhimurium . The authors showed that in contrast to wild type living bacteria, dead bacteria elicited an oxidative burst and pH changes in tobacco cells. Similar response was provoked by the inv A− mutant, which has no functional SPI-1 T3SS . Those results suggest that Salmonella depends on the secretion of effectors during infection of tobacco leaves to actively suppress their immune responses. A general transcriptome analysis performed in our laboratory suggests a similar scenario . Infection with the prg H− mutant, but not with the 14028 s wild type, induces about 640 genes, the majority of which are related to defense responses. Moreover, we showed that mutants impaired in their T3SSs are less virulent towards Arabidopsis plants then wild type bacteria . Taken together, recently published and presented results build a growing body of evidences indicating that Salmonella, similarly to the infection in animals, actively suppresses the plant defense mechanisms. Whether this bacterium uses the same or different effectors in order to achieve this goal is not yet clear, it seems however to be acceptable to conclude that Salmonella uses the same T3SSs in plant and animal infections.
This report presents an automatic pixel-based classification method for detecting “unhealthy” regions in leaf images. This method has been tested extensively with very promising results. Linear SVM has been used to classify each pixel. We have also shown how the results from SVM could be remarkably improved by using the neighborhood-check technique. The proposed method was compared to existing method and showed a higher accuracy. We used this algorithm to study the impact of the human pathogenic bacterium Salmonella Typhimurium on plants immune system. The comparison between wild type bacteria and T3SS mutants showed similarity in the infection process in animals and in plants. The result obtained with the proposed algorithm and also transcriptome and biochemical analyses suggest that T3SSs are necessary for a successful infection of plants. Plant epidemiology is only one possible application of the proposed algorithm, it can be easily extended to other detection tasks, which also rely on color information, or even extended to other features.
Arabidopsis thaliana wild type Col-0 (NASC ID: N70000) seeds were germinated on MS media for around 2 weeks. The seedlings were then transferred to soil and grown in short day chamber (7 hours of light) at 24°C for additional 4 weeks.
Salmonella enterica subsp. enterica serovar Typhimurium (ATCC 14028s), Salmonella T3SS mutants (all in the 14028s genetic background) and Escherichia coli K12 strain DH5αwere grown on LB agar and liquid media with required antibiotics. Pseudomonas syringae pathovar tomato DC3000 was grown in King’s B medium containing required antibiotics. prg H− and ssa V−mutants were obtained from Prof. David Holden, Imperial College, London. inv A− and ssa J− mutants were constructed in the INRA Tours laboratory by Dr. Isabelle Virlogeux-Payant.
Around 6-week-old Arabidopsis plants were chosen for infiltration experiment. The cultured bacteria were spun down, washed with 10 mM MgCl2 solution. Final optical density (OD600) of infiltration solution was 0.1. Infiltration was done via syringe on the abaxial surface of the leaves.
Analysis of lesions in leaves
For the analysis, images of leaves were captured at 5 consecutive days after infiltration. At least 5 leaves were photographed per each time point and infiltration variant. This experiment was repeated 5 times. Lesions in leaves were analyzed with the help of an automated program calculating the changed color in a proportion to the normal color of the leaves. The diseased portion were calculated in percentage and evaluated, cf. Section Image-Based Classification. Altogether over 1200 images were evaluated.
Around 2-week-old Arabidopsis plants were transferred to MS liquid media and left undisturbed overnight. Bacteria were washed in 10 mM MgCl2, and the liquid medium was inoculated with bacteria at OD600 = 0.1. Whole plants were collected at regular intervals for further analysis.
RNA extraction and reverse transcription
Extraction of total RNA was done with Trizol Ⓡ (Invitrogen) accordingly to manufacturer instructions. Whole plants were collected in liquid nitrogen and homogenized. Total RNA was extracted. All RNA samples were treated with DNase I (Fermentas International Inc.). Complementary DNA (cDNA) was prepared with the help of reverse transcriptase (qScript, Quanta Biosciences) accordingly to manufacturer protocol. Equal amount of 2 μg RNA from all samples was taken to ensure the best possible gene expression levels analysis.
After the preparation of cDNA, quantitative PCR was performed in the Applied Biosystems 7500 FAST real-time PCR system. SYBR green was used as a fluorescence dye for the PCR reactions. 20 μl total volume reaction was used and three repetitions were made for each of the sample. qPCR was done with the following primers: UBQ4: forward primer: GCT TGG AGT CCT GCT TGG ACG, reverse primer: CGC AGT TAA GAG GAC TGT CCG GC; PDF1.2: forward primer: GTT TGC TTC CAT CAT CAC CC, reverse primer: GGG ACG TAA CAG ATA CAC TTG.
Western blot analysis
Whole plants were collected in liquid nitrogen, homogenized in a tissue homogenizer and total protein were extracted in 200 μl of lysis buffer (25 mM TRIS (pH = 7.8), 10 mM MgCl2, 15 mM EGTA, 75 mM NaCl, 1 mM DTT, 0.5 mM NaVO4, 1 mM NaF, 15 mM β-glycerophosphate (Sigma-Aldrich), 15 mM 4-nitrophenyl phosphate (Sigma), 0.5 mM PMSF, 5 μg/ml leupeptine (Roche), 5 μg/ml aprotinin (Roche), 0.1% Tween 20). After vigorous vortexing, samples were centrifuged at 14,000 rpm and supernatant, containing the proteins was collected. Bio-Rad mini format 1-D electrophoresis system was used for sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE). 12% resolving gel and 3.2% stacking gel were used. Equal amount of proteins (20 μg) was used for each sample. Primary antibodies: α-phospho-ERK 1/2 (Sigma-Aldrich), AtMPK6 (Biolabs). Secondary antibody: Anti-Rabbit IgG HRP-conjgate (Sigma-Aldrich).
The authors would like to thank Prof. David Holden (Imperial College London) for the prg H− and ssa V−Salmonella mutants, Dr. Isabelle Virlogeux-Payant (INRA Tours) for the inv A−and ssa J−mutants and Prof. Katja Becker (JLU Gießen) for permission to work in the S2 laboratory.
- Heaton J, Jones K: Microbial contamination of fruit and vegetables and the behaviour of enteropathogens in the phyllosphere: a review. J Appl Microbiol. 2008, 104: 613-626. 10.1111/j.1365-2672.2007.03587.x.View ArticlePubMedGoogle Scholar
- Holden N, Pritchard L, Toth I: Colonization outwith the colon: plants as an alternative environmental reservoir for human pathogenic enterobacteria. FEMS Microbiol Rev. 2009, 33: 689-703. 10.1111/j.1574-6976.2008.00153.x.View ArticlePubMedGoogle Scholar
- Plotnikova J, Rahme L, Ausubel F: Pathogenesis of the human opportunistic pathogen Pseudomonas aeruginosa PA14 in Arabidopsis. Plant Physiol. 2000, 124: 1766-1774. 10.1104/pp.124.4.1766.PubMed CentralView ArticlePubMedGoogle Scholar
- Prithiviraj B, Bais H, Jha A, Vivanco J: Staphylococcus aureus pathogenicity on Arabidopsis thaliana is mediated either by a direct effect of salicylic acid on the pathogen or by SA-dependent, NPR1-independent host responses. Plant J. 2005, 42: 417-432. 10.1111/j.1365-313X.2005.02385.x.View ArticlePubMedGoogle Scholar
- Milillo S, Badamo J, Boor K, Wiedmann M: Growth and persistence of Listeria monocytogenes isolates on the plant model Arabidopsis thaliana. Food Microbiol. 2008, 25: 698-704. 10.1016/j.fm.2008.03.003.View ArticlePubMedGoogle Scholar
- Brandl M: Fitness of human enteric pathogens on plants and implications for food safety. Annu Rev Phytopathol. 2006, 44: 367-392. 10.1146/annurev.phyto.44.070505.143359.View ArticlePubMedGoogle Scholar
- Westrell T, Ciampa N, Boelaert F, Helwigh B, Korsgaard H, Chriel M, Ammon A, Makela P: Zoonotic infections in Europe in 2007: a summary of the EFSA-ECDC annual report. Euro Surveill. 2009, 14: 1-3.Google Scholar
- Sivapalasingam S, Friedman C, Cohen L, Tauxe RV: Fresh produce: a growing cause of outbreaks of foodborne illness in the United States, 1973 through 1997. J Food Prot. 2004, 67: 2342-2353.PubMedGoogle Scholar
- Rangel J, Sparling P, Crowe C, Griffin P, Swerdlow D: Epidemiology of Escherichia coli O157:H7 outbreaks, United States, 1982-2002. Emerg Infect Dis. 2005, 11: 603-609. 10.3201/eid1104.040739.PubMed CentralView ArticlePubMedGoogle Scholar
- Gerner-Smidt P, Hise K, Kincaid J, Hunter S, Rolando S, Hyytia-Trees E, Ribot E, Swaminathan B: PulseNet USA: a five-year update. Foodborne Pathog Dis. 2006, 3: 9-19. 10.1089/fpd.2006.3.9.View ArticlePubMedGoogle Scholar
- Barak J, Gorski L, Liang A, Narm K: Previously uncharacterized Salmonella enterica genes required for swarming play a role in seedling colonization. Microbiology. 2009, 155: 3701-3709. 10.1099/mic.0.032029-0.View ArticlePubMedGoogle Scholar
- Barak J, Gorski L, Naraghi-Arani P, Charkowski AO: Salmonella enterica virulence genes are required for bacterial attachment to plant tissue. Appl Environ Microbiol. 2005, 71: 5685-5691. 10.1128/AEM.71.10.5685-5691.2005.PubMed CentralView ArticlePubMedGoogle Scholar
- Barak J, Kramer L, Hao L: Colonization of tomato plants by Salmonella enterica is cultivar dependent, and type 1 trichomes are preferred colonization sites. Appl Environ Microbiol. 2011, 77: 498-504. 10.1128/AEM.01661-10.PubMed CentralView ArticlePubMedGoogle Scholar
- Golberg D, Kroupitski Y, Belausov E, Pinto R, Sela S: Salmonella Typhimurium internalization is variable in leafy vegetables and fresh herbs. Int J Food Microbiol. 2011, 145: 250-257. 10.1016/j.ijfoodmicro.2010.12.031.View ArticlePubMedGoogle Scholar
- Iniguez A, Dong Y, Carter H, Ahmer B, Stone J, Triplett E: Regulation of enteric endophytic bacterial colonization by plant defenses. Mol Plant Microbe Interact. 2005, 18: 169-178. 10.1094/MPMI-18-0169.View ArticlePubMedGoogle Scholar
- Klerks M, Franz E, van Gent-Pelzer M, Zijlstra C, van Bruggen A: Differential interaction of Salmonella enterica serovars with lettuce cultivars and plant-microbe factors influencing the colonization efficiency. ISME J. 2007, 1: 620-631. 10.1038/ismej.2007.82.View ArticlePubMedGoogle Scholar
- Kroupitski Y, Golberg D, Belausov E, Pinto R, Swartzberg D, Granot D, Sela S: Internalization of Salmonella enterica in leaves is induced by light and involves chemotaxis and penetration through open stomata. Appl Environ Microbiol. 2009, 75: 6076-6086. 10.1128/AEM.01084-09.PubMed CentralView ArticlePubMedGoogle Scholar
- Noel JT, Arrach N, Alagely A, McClelland M, eplitski M: Specific responses of Salmonella enterica to tomato varieties and fruit ripeness identified by in vivo expression technology. PLoS One. 2010, 5: e12406-10.1371/journal.pone.0012406.PubMed CentralView ArticlePubMedGoogle Scholar
- Saggers E, Waspe C, Parker M, Waldron K, Brocklehurst T: Salmonella must be viable in order to attach to the surface of prepared vegetable tissues. J Appl Microbiol. 2008, 105: 1239-1245. 10.1111/j.1365-2672.2008.03795.x.View ArticlePubMedGoogle Scholar
- Schikora A, Carreri A, Charpentier E, Hirt H: The dark side of the salad: Salmonella typhimurium overcomes the innate immune response of Arabidopsis thaliana and shows an endopathogenic lifestyle. PLoS One. 2008, 3: e2279-10.1371/journal.pone.0002279.PubMed CentralView ArticlePubMedGoogle Scholar
- Warfield SK, Kaus M, Jolesz FA, Kikinis R: Adaptive, Template Moderated, Spatially Varying Statistical Classification. Medical Image Analysis. 2000, 4 (1): 43-55. 10.1016/S1361-8415(00)00003-7.View ArticlePubMedGoogle Scholar
- Cocosco CA, Zijdenbos AP, Evans AC: A Fully Automatic and Robust Brain MRI Tissue Classification Method. Medical Image Analysis. 2003, 7 (4): 513-527. 10.1016/S1361-8415(03)00037-9.View ArticlePubMedGoogle Scholar
- Kamber M, Shinghal R, Collins D, Francis G, Evans A: Model-based 3-D segmentation of multiple sclerosis lesions in magnetic resonance brain images. IEEE Trans Med Imaging. 1995, 14 (3): 442-453. 10.1109/42.414608.View ArticlePubMedGoogle Scholar
- Schikora A, Carreri A, Charpentier E, Hirt H: The dark side of salad: Salmonella typhimurium overcomes the innate immune response of Arabidopsis thaliana and shows an endophatogenic lifestyle. PLoS ONE. 2008, 3 (5): e2279-10.1371/journal.pone.0002279.PubMed CentralView ArticlePubMedGoogle Scholar
- Hashim H, Haron MA, Osman FN, Junid SAMA: Classification of Rubber Tree Leaf Disease Using Spectrometer. Proceedings of the 2010 Fourth Asia International Conference on Mathematical/Analytical Modelling and Computer Simulation. 2010, Washington, DC, USA: IEEE Computer Society, 302-306.View ArticleGoogle Scholar
- Díaz G, Romero E, Boyero JR, Malpica N: Recognition and Quantification of Area Damaged by Oligonychus Perseae in Avocado Leaves. Proceedings of the 14th Iberoamerican Conference on Pattern Recognition: Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications, CIARP ’09. 2009, Guadalajara, Jalisco, Mexico: Springer-Verlag, 677-684.View ArticleGoogle Scholar
- Schikora M, Schikora A, Kogel KH, Koch W, Cremers D: Probabilistic Classification of Disease symptoms caused by Salmonella on Arabidopsis Plants. GI Jahrestagung. 2010(2), 874-879.Google Scholar
- Patel J, Rossanese O, Galan J: The functional interface between Salmonella and its host cell: opportunities for therapeutic intervention. Trends Pharmacol Sci. 2005, 26: 564-570. 10.1016/j.tips.2005.09.005.View ArticlePubMedGoogle Scholar
- Collazo C, Galan J: The invasion-associated type-III protein secretion system in Salmonella. Gene. 2997, 192: 51-59.View ArticleGoogle Scholar
- Hensel M: Salmonella pathogenicity island 2. Mol Microbiol. 2000, 36: 1015-1023. 10.1046/j.1365-2958.2000.01935.x.View ArticlePubMedGoogle Scholar
- Kuhle V, Hensel M: Cellular microbiology of intracellular Salmonella enterica: functions of the type III secretion system encoded by Salmonella pathogenicity island 2. Cell Mol Life Sci. 2004, 61: 2812-2826. 10.1007/s00018-004-4248-z.View ArticlePubMedGoogle Scholar
- Waterman S, Holden D: Functions and effectors of the Salmonella pathogenicity island 2 type III secretion system. Cell Microbiol. 2003, 5: 501-511. 10.1046/j.1462-5822.2003.00294.x.View ArticlePubMedGoogle Scholar
- Lara-Tejero M, Kato J, Wagner S, Liu X, Galan J: A sorting platform determines the order of protein secretion in bacterial type III systems. Science. 2011, 331: 1188-1191. 10.1126/science.1201476.View ArticlePubMedGoogle Scholar
- Shirron N, Yaron S: Active Suppression of Early Immune Response in Tobacco by the Human Pathogen Salmonella Typhimurium. PLoS One. 2011, 6: e18855-10.1371/journal.pone.0018855.PubMed CentralView ArticlePubMedGoogle Scholar
- Gomez-Gomez L, Boller T: FLS2: an LRR receptor-like kinase involved in the perception of the bacterial elicitor flagellin in Arabidopsis. Mol Cell. 2000, 5: 1003-1011. 10.1016/S1097-2765(00)80265-8.View ArticlePubMedGoogle Scholar
- Gohre V, Spallek T, Haweker H, Mersmann S, Mentzel T, Boller T, de Torres M, Mansfield J, Robatzek S: Plant pattern-recognition receptor FLS2 is directed for degradation by the bacterial ubiquitin ligase AvrPtoB. Curr Biol. 2008, 18: 1824-1832. 10.1016/j.cub.2008.10.063.View ArticlePubMedGoogle Scholar
- Shan L, He P, Li J, Heese A, Peck S, Martin TNG, Sheen J: Bacterial effectors target the common signaling partner BAK1 to disrupt multiple MAMP receptor-signaling complexes and impede plant immunity. Cell Host Microbe. 2008, 4: 17-27. 10.1016/j.chom.2008.05.017.PubMed CentralView ArticlePubMedGoogle Scholar
- Zhang J, Shao F, Li Y, Cui H, Chen L, Li H, Zou Y, Long C, Lan L, Chai J, Chen S, Tang X, Zhou J: A Pseudomonas syringae effector inactivates MAPKs to suppress PAMP-induced immunity in plants. Cell Host Microbe. 2007, 1: 175-185. 10.1016/j.chom.2007.03.006.View ArticlePubMedGoogle Scholar
- Li H, Xu H, Zhou Y, Zhang J, Long C, Li S, Chen S, Zhou J, Shao F: The phosphothreonine lyase activity of a bacterial type III effector family. Science. 2007, 315: 1000-1003. 10.1126/science.1138960.View ArticlePubMedGoogle Scholar
- Zhu Y, Li H, Long C, Hu L, Xu H, Liu L, Chen S, Wang D, Shao F: Structural insights into the enzymatic mechanism of the pathogenic MAPK phosphothreonine lyase. Mol Cell. 2007, 28: 899-913. 10.1016/j.molcel.2007.11.011.View ArticlePubMedGoogle Scholar
- Mazurkiewicz P, Thomas J, Thompson J, Liu M, Arbibe L, Sansonetti P, Holden D: SpvC is a Salmonella effector with phosphothreonine lyase activity on host mitogen-activated protein kinases. Mol Microbiol. 2008, 67: 1371-1383. 10.1111/j.1365-2958.2008.06134.x.PubMed CentralView ArticlePubMedGoogle Scholar
- McGhie E, Brawn L, Hume P, Humphreys D, Koronakis V: Salmonella takes control: effector-driven manipulation of the host. Curr Opin Microbiol. 2009, 12: 117-124. 10.1016/j.mib.2008.12.001.PubMed CentralView ArticlePubMedGoogle Scholar
- Schikora A, Virlogeux-Payant I, Bueso E, Garcia A, Nilau T, Charrier A, Pelletier S, Menanteau P, Baccarini M, Velge P, Hirt H: Conservation of Salmonella, infection mechanisms in plants and animals. PLoS ONE. 2011, xx: 1-8.Google Scholar
- Carbon S, Ireland A, Mungall C, Shu S, Marshall B, Lewis S: AmiGO: online access to ontology and annotation data. Bioinformatics. 2009, 25: 288-289. 10.1093/bioinformatics/btn615.PubMed CentralView ArticlePubMedGoogle Scholar
- Madhogaria S, Schikora M, Koch W, Cremers D: Pixel-Based Classification Method for Detecting Unhealthy Regions in Leaf Images. Proc. of 6th. Workshop on Sensor Data Fusion (SDF). 2011Google Scholar
- Mumford D, Shah J: Optimal approximations by piecewise smooth functions and associated variational problems. Commun Pur Appl Math. 1989, 42: 577-685. 10.1002/cpa.3160420503.View ArticleGoogle Scholar
- Chan T, Vese L: Active contours without edges. IEEE Trans Image Process. 2001, 10 (2): 266-277. 10.1109/83.902291.View ArticlePubMedGoogle Scholar
- Unger M, Pock T, Cremers D, Bischof H: TVSeg - Interactive Total Variation Based Image Segmentation. British Machine Vision Conference (BMVC). 2008Google Scholar
- Schikora M, Häge M, Ruthotto E, Wild K: A convex formulation for color image segmentation in the context of passive emitter localization. 12th International Conference of Information Fusion. 2009, 1424-1431.Google Scholar
- Hafner W: Segmentierung von Video-Bildfolgen durch Adaptive Farbklassifikation. 1999, Herbert Utz Verlag GmbH ISBN 3-89675-483-1Google Scholar
- Kolev K, Klodt M, Brox T, Cremers D: Continuous Global Optimization in Multiview 3D Reconstruction. Int J Comput Vis. 2009, 84: 80-96. 10.1007/s11263-009-0233-1.View ArticleGoogle Scholar
- Cortes C, Vapnik V: Support Vector Networks. In Machine Learning. Volume. 1995, 20: 273-297.Google Scholar
- Burges C: A tutorial on support vector machines for pattern recognition. Data Min Knowl Disc. 1998, 2: 121-167. 10.1023/A:1009715923555.View ArticleGoogle Scholar
- John Shawe-Taylor NC: Kernel Methods for Pattern Analysis. 2004, New York, NY, USA: Cambridge University PressView ArticleGoogle Scholar
- Hensel M, Shea J, Baumler A, Gleeson C, Blattner F, Holden D: Analysis of the boundaries of Salmonella pathogenicity island 2 and the corresponding chromosomal region of Escherichia coli K-12. J Bacteriol. 1997, 179: 1105-1111.PubMed CentralPubMedGoogle Scholar
- Galan J, Collmer A: Type III secretion machines: bacterial devices for protein delivery into host cells. Science. 1999, 284: 1322-1328. 10.1126/science.284.5418.1322.View ArticlePubMedGoogle Scholar
- Behlau I, Miller S: A PhoP-repressed gene promotes Salmonella typhimurium invasion of epithelial cells. J Bacteriol. 1993, 175: 4475-4484.PubMed CentralPubMedGoogle Scholar
- Kubori T, Sukhan A, Aizawa S, Galan J: Molecular characterization and assembly of the needle complex of the Salmonella typhimurium type III protein secretion system. Proc Natl Acad Sci U S A. 2000, 97: 10225-10230. 10.1073/pnas.170128997.PubMed CentralView ArticlePubMedGoogle Scholar
- Marlovits T, Kubori T, Lara-Tejero M, Thomas D, Unger V, Galan J: Assembly of the inner rod determines needle length in the type III secretion injectisome. Nature. 2006, 441: 637-640. 10.1038/nature04822.View ArticlePubMedGoogle Scholar
- Pitzschke A, Schikora A, Hirt H: MAPK cascade signalling networks in plant defence. Curr Opin Plant Biol. 2009, 12: 421-426. 10.1016/j.pbi.2009.06.008.View ArticlePubMedGoogle Scholar
- Arbibe L, DW DK, Batsche E, Pedron T, Mateescu B, Muchardt C, Parsot C, Sansonetti P: An injected bacterial effector targets chromatin access for transcription factor NF-kappaB to alter transcription of host genes involved in immune responses. Nat Immunol. 2007, 8: 47-56.View ArticlePubMedGoogle Scholar
- Li H, Xu H, Zhou J, Zhang J, Long C, Li S, Chen S, Zhou J, Shao F: The phosphothreonine lyase activity of a bacterial type III effector family. Science. 2007, 315: 1000-1003. 10.1126/science.1138960.View ArticlePubMedGoogle Scholar
- Mazurkiewicz P, Thomas J, Thompson J, Liu M, Arbibe L, Sansonetti P, Holden D: SpvC is a Salmonella effector with phosphothreonine lyase activity on host mitogen-activated protein kinases. Mol Microbiol. 2008, 67: 1371-1383. 10.1111/j.1365-2958.2008.06134.x.PubMed CentralView ArticlePubMedGoogle Scholar
- Lin S, Le T, Cowen D: SptP, a Salmonella typhimurium type III-secreted protein, inhibits the mitogen-activated protein kinase pathway by inhibiting Raf activation. Cell Microbiol. 2003, 5: 267-275. 10.1046/j.1462-5822.2003.t01-1-00274.x.View ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.