Global impact of Salmonella type III secretion effector SteA on host cells

! 1! Global impact of Salmonella type III secretion effector SteA on host cells Elena Cardenal-Muñoza, Gabriel Gutiérrezb and Francisco Ramos-Morales* Departamento de Genética, Facultad de Biología, Universidad de Sevilla, Apartado 1095, 41080 Sevilla, Spain a[email protected] b[email protected] *Corresponding author. Mailing address: Departamento de Genética, Facultad de Biología, Universidad de Sevilla, Avda Reina Mercedes, 6, 41012 Sevilla, Spain. Phone: 34 95 455 7107. Fax: 34 95 455 7104. E-mail: [email protected] ! 2! Abstract Salmonella enterica is a Gram-negative bacterium that causes gastroenteritis, bacteremia and typhoid fever in several animal species including humans. Its virulence is greatly dependent on two type III secretion systems, encoded in pathogenicity islands 1 and 2. These systems translocate proteins called effectors into eukaryotic host cell. Effectors interfere with host signal transduction pathways to allow the internalization of pathogens and their survival and proliferation inside vacuoles. SteA is one of the few Salmonella effectors that are substrates of both type III secretion systems. Here, we used gene arrays and bioinformatics analysis to study the genetic response of human epithelial cells to SteA. We found that constitutive synthesis of SteA in HeLa cells leads to induction of genes related to extracellular matrix organization and regulation of cell proliferation and serine/threonine kinase signaling pathways. SteA also causes repression of genes related to immune processes and regulation of purine nucleotide synthesis and pathway-restricted SMAD protein phosphorylation. In addition, a cell biology approach revealed that epithelial cells expressing steA show altered cell morphology, and decreased cytotoxicity, cell-cell adhesion and migration. Keywords Salmonella enterica; type III secretion system; SteA; microarray; human epithelial cell; cellcell adhesion. Abbreviations T3SS, type three secretion system; SPI, Salmonella pathogenicity island; NCBI, National Center for Biotechnology Information; LB, Luria-Bertani; FDR, false discovery rate; LDH, lactate dehydrogenase; qPCR, quantitative real-time PCR; HEMA, 2-hydroxyethyl methacrylate; GO, gene ontology. ! 3! 1. Introduction Many pathogenic Gram-negative bacteria possess type III secretion systems (T3SSs) for their interaction with the host. These systems allow delivery into eukaryotic host cells of effector proteins that direct the different stages of the infection at the cellular level [1]. Salmonella enterica possesses two distinct virulence-related T3SSs, T3SS1 and T3SS2, that are encoded by genes located in Salmonella pathogenicity islands 1 and 2 (SPI1 and SPI2), respectively [2]. T3SS1 is necessary for the invasion of non-phagocytic cells [3], whereas T3SS2 is induced after invasion and is essential for survival and replication within macrophages [4,5]. S. enterica injects more than thirty T3SS effectors to their host cells and some of them have been shown to manipulate cellular processes such as actin cytoskeleton organization, tight junction alterations, biogenesis of the Salmonella-containing vacuole and cell death [6]. However, the functions for many effectors are still unknown. SteA was identified as a S. enterica serovar Typhimurium T3SS effector [7] that can be secreted to culture media and translocated into epithelial cells and macrophages through T3SS1 and T3SS2, depending on culture conditions, infected cell types and infection times. The first 10 amino acids of SteA act as a signal sequence for its translocation into the eukaryotic cell [8]. The gene steA is located outside SPI1 and SPI2, and its low GC content (43%) suggests horizontal acquisition, common in virulence-associated genes. We have previously shown that its expression is transcriptionally controlled by the bacterial redox status in a PhoQ/PhoP-dependent manner [9]. A steA null mutant is three-fold attenuated for BALB/c mice virulence after intraperitoneal infection [7] and SteA seems to be involved in the bacterial persistence during long time infection in 129X1/SvJ mice [10]. In the host cell, SteA localizes to the trans-Golgi network (TGN) and to Salmonella-induced membrane tubules containing the trans-Golgi marker GalT-mCherry [7,11]. A very recent report has shown that SteA participates in the control of Salmonella-containing vacuole membrane ! 4! dynamics![12]. This study also suggests that SteA should have additional roles, which remain to be elucidated, at earlier times of host cell infection. Similarity to proteins with known activities has been useful in some cases to determine the function of a specific effector [13], but in the case of SteA no sequence similarities have been detected. Expression of individual effectors, like AvrA or SopB, in the budding yeast Saccharomyces cerevisiae, that serves as a simplified heterologous model [14], has been a productive approach to study the effects of these proteins on host cells [15,16,17], thereby helping in the discovery of their activities. Here, we use mammalian cells as a relevant model to specifically analyze the effect on the host transcriptome of the effector SteA. We show that in epithelial HeLa cells SteA leads to changes in the expression of genes related to extracellular matrix organization, cell proliferation, serine/threonine kinase signaling pathways, immune processes, regulation of purine nucleotide synthesis and pathwayrestricted SMAD protein phosphorylation, and produces significant changes in cell death, adhesion, and migration. ! 5! 2. Materials and methods 2.1. Bacterial strains and bacterial culture Bacterial strains used were Escherichia coli DH5α [18] and S. enterica serovar Typhimurium SV5846![8], a derivative of strain 14028 carrying a steA::3xFLAG chromosomal fusion. The standard bacterial culture medium was Luria-Bertani (LB) broth. Solid LB contained agar 1.5% final concentration. Antibiotics were used at the following concentrations: kanamycin, 50 µg/ml; ampicillin, 100 µg/ml. 2.2 Plasmid construction, DNA amplification, and sequencing Plasmid pIZ1963 is a derivative of pBABEpuro [19] coding for SteA with a C-terminal 3xFLAG tag. To construct this plasmid, steA::3xFLAG was amplified from strain SV5846 using primers steApcdnadir and steAflagxho3, digested with BamHI and XhoI and ligated together with vector pBABEpuro previously digested with BamHI and SalI. Amplification reactions and confirmation by sequencing were carried out as previously described [8]. Primers are listed in Table S1. 2.3 Mammalian cell culture, transfection and lysis HeLa cells (ECAC no. 93021013) were cultured, transfected, and lysed as previously described [20]. For stable transfection, HeLa cells were electroporated with pBABEpuro or its derivative and selection was started 24 h after electroporation in medium containing 1 µg/ml puromycin (InvivoGen). 2.4 Bacterial infection of cultured cells and analysis of SteA translocation HeLa cells were plated in 6-well plates at 6x105 cells per well and incubated for 24 h. Salmonella infections were carried out as previously described![8]. The cell culture was ! 6! washed twice with PBS 1 h post-infection and overlaid with DMEM containing 100 µg/ml gentamicin. One hour later the concentration of gentamicin was lowered to 16 µg/ml after an additional wash with PBS. Infected mammalian cells were lysed 6 h post-infection with Nonidet P-40 buffer as described![8]. The extract was centrifuged at 13000 rpm for 20 min and the supernatant was filtered and analyzed by immunoblot. 2.5 Electrophoresis, immunoblot, and antibodies Proteins in cell extracts were resolved by SDS-PAGE. The gel was blotted onto a nitrocellulose membrane (Amersham) and probed with mouse monoclonal anti-FLAG M2 primary antibodies (1:5000; Sigma) and mouse monoclonal anti-β-actin C4 (1:5000; Santa Cruz Biotech) as loading control. Goat anti-mouse horseradish peroxidase-conjugated antibodies (1:5000; BioRad) were used as secondary antibodies. Detection was via chemiluminescence procedures (Pierce). 2.6 RNA preparation, gene array processing, and statistical analysis Total RNA from HeLa cells stably transfected with pBABEpuro or its derivative encoding the SteA-3xFLAG fusion was isolated in triplicate using 1 ml of TRIzol reagent (Invitrogen) according to the protocol supplied by the manufacturer. An additional purification step was carried out by using the RNeasy Min Elute Cleanup Kit (Qiagen). Biotinylated singlestranded cDNA was prepared from 100 ng per sample of total intact RNA extracted from 9 independent samples (3 from HeLa pBABEpuro, 3 from HeLa pBABEpuro-SteA-3xFLAG L2, and 3 from HeLa pBABEpuro-SteA-3xFLAG L4). Labelled cDNA was hybridized to GeneChip Human Gene 1.0 ST Arrays (Affymetrix) following the manufacturer’s instructions. The arrays were scanned in a 3000 7G Scanner from Affymetrix. Image analysis, fluorescent data quantification and quality control was carried out with Affymetrix software. All procedures and preliminary data analysis, including fluorescent data processing, ! 7! normalization using Robust Multi-array Average (RMA) algorithms, and annotations, were performed at the Genomics Unit of the Andalusian Center for Molecular Biology and Regenerative Medicine (CABIMER, Seville, Spain). Fold change was calculated for each cell line relative to the control line transfected with pBABEpuro. Statistical significance (p value) was calculated by empirical Bayes moderated t-test based on the results of three arrays per condition. The false discovery rate (FDR) for each p value was also calculated. Genes that changed with an FDR-adjusted p value higher than 0.05 were removed from subsequent analysis. The microarray data used in this analysis is available from NCBI’s Gene Expression Omnibus at http://www.ncbi.nlm.nih.gov/geo/ under accession number GSE51043. 2.7 Function prediction of gene sets To predict the functional association networks of the activated and repressed genes from the microarray, we used the web-based software application tool GeneMANIA [21]. GO terms associated to genes in Table 1 were searched using AmiGO version 1.8 [22]. 2.8 Quantitative real-time PCR (qPCR) The protocol for qPCR was previously described [9]. DNA primers are indicated in Table S1. Gene expression levels were normalized to transcripts of BCAT1, gene that appeared not affected by the presence of SteA in the microarray. 2.9 Proliferation assays HeLa cells were seeded in 6-well plates at a density of 6x104 cells per well and grown at 37ºC, 5% CO2 in DMEM. At 24 h intervals, the cells from a well per cell line were trypsinized and counted on a haemocytometer 2.10 Cell death, cell adhesion, and cell migration assays ! 8! Cell death was measured using a previously described protocol based on lactate dehydrogenase (LDH) release in the cultures [20]. Cell adhesion was determined as previously described [23]. For cell-cell adhesion assays, 6x104 cells/ml were seeded onto 6- well plates coated with poly-(2-hydroxyethyl methacrylate) (poly-HEMA) (Sigma) in order to prevent cell-matrix interactions. After 48 h of culture in suspension, cells were photographed on an inverted microscope. For the migration assay (wound closure assay), we followed the protocol previously described [24]. 2.11 Statistical analysis Student’s t test was used to analyze differences in growth, cytotoxicity percentage, and relative migration. p values of 0.05 or less were considered significant. ! 9! 3. Results and Discussion 3.1$cDNA$microarray$analysis$of$gene$expression$in$HeLa$cells$in$response$to$SteA$ To investigate the consequences of the presence of the Salmonella T3SS effector SteA in human epithelial cells, we generated three stably transfected HeLa cell lines, one with the empty vector pBABEpuro and two with a derivative of the same vector encoding a SteA- 3xFLAG fusion (lines SteA L2 and and SteA L4). The expression level of tagged SteA in these cell clones was determined by Western blot analysis and compared to the level of SteA translocated into HeLa cells during in vitro infections with S. enterica serovar Typhimurium strain SV5846. As seen in Figs. 1A and 1B, the level of SteA was higher in L4 than in L2. Importantly, in both cell lines the levels were lower than after translocation from Salmonella. This result suggests that the effects of SteA studied in this work are not due to a nonphysiological level of expression of the effector. Next, we performed microarray hybridization with RNA from stably transfected cells. Biotin-labelled cDNAs prepared from total RNA were hybridized to microarray chips containing 28,869 sequenced human genes. The number of genes that changed in response to SteA expression in a statistically significant manner (FDR-adjusted p value < 0.05) is represented in Fig. 1C (upregulated genes) and Fig. 1D (downregulated genes) using Venn diagrams. As a result of this analysis we decided to focus on 58 genes (0.2 % of the whole human gene number analyzed) that showed differential expression in both of the cell lines expressing SteA (L2 and L4) when compared to the control cell line. Of these, 13 genes showed higher expression while 45 genes showed lower expression. These genes are listed in Table S2. A general conclusion that can be drawn from the microarray analysis is that the presence of SteA within HeLa cells causes more gene repression than activation. This is true for both stably transfected cell lines, L2 and L4. The ! 16! [20] J. Bernal-Bayard, F. Ramos-Morales, Salmonella type III secretion effector SlrP is an E3 ubiquitin ligase for mammalian thioredoxin, J Biol Chem 284 (2009) 27587-27595. [21] K. Zuberi, M. Franz, H. Rodriguez, J. Montojo, C.T. Lopes, G.D. Bader, Q. Morris, GeneMANIA prediction server 2013 update, Nucleic Acids Res 41 (2013) W115-122. [22] S. Carbon, A. Ireland, C.J. Mungall, S. Shu, B. Marshall, S. Lewis, AmiGO: online access to ontology and annotation data, Bioinformatics 25 (2009) 288-289. [23] S. Mueller, Cadenas, E., Schönthal, A. H., p21WAF1 regulates anchorage-independent growth of HCT116 colon carcinoma cells via E-cadherin expression., Cancer Res 60 (2000) 156-163. [24] W. Kleeberger, Bova, G. S., Nielsen, M. E., Herawi, M., Chuang, A. Y., Epstein, J. I., Berman, D. M., Roles for the stem cell associated intermediate filament Nestin in prostate cancer migration and metastasis., Cancer Res 67 (2007) 9199-9206. [25] O. Steele-Mortimer, L.A. Knodler, S.L. Marcus, M.P. Scheid, B. Goh, C.G. Pfeifer, V. Duronio, B.B. Finlay, Activation of Akt/protein kinase B in epithelial cells by the Salmonella typhimurium effector SigD, J Biol Chem 275 (2000) 37718-37724. [26] H. Wu, R.M. Jones, A.S. Neish, The Salmonella effector AvrA mediates bacterial intracellular survival during infection in vivo, Cell Microbiol 14 (2012) 28-39. [27] J.E. Galán, Common themes in the design and function of bacterial effectors, Cell Host Microbe 5 (2009) 571-579. ! 17! Figure legends Fig. 1. Expression of SteA-3xFLAG in stably transfected HeLa cells (A) HeLa cells were stably transfected with plasmids pBABEpuro (Vector) and pBABEpuro-SteA-3xFLAG (SteA L2 and SteA L4) or were infected with S. enterica serovar Typhimurium strain SV5846 (Infected) and lysates were immunoblotted with monoclonal anti-FLAG antibodies and with antibodies against β-actin as a loading control. (B) Quantification of SteA-3xFLAG with respect to β-actin and normalization regarding the SteA L2 quantification. Means and standard deviations from two independent experiments are represented. (C and D) Venn diagrams of genes with differential expression between HeLa cells expressing SteA-3xFLAG and control HeLa cells. Number of significantly (FDR-adjusted p value < 0.05) upregulated (C) or downregulated (D) genes in clones expressing SteA-3xFLAG compared to the control (L2/Control and L4/Control) is shown. Fig. 2 Microarray and qPCR results for selected differentially expressed genes. (A) Microarray results of 7 differentially expressed genes in HeLa pBABEpuro-SteA-3xFLAG (SteA L2 and SteA L4) compared to HeLa pBABEpuro. (B) qPCR results showing mRNA relative amount of the same genes in the same strains. Means and standard deviations from three experiments performed in triplicate are represented. Gene names are indicated. M (log) is the log2 of the ratio L2/control or L4/control for each gene Fig. 3. Effect of SteA on cell death. (A) Proliferation of HeLa cells expressing SteA- 3xFLAG (SteA L4) compared to control cells (Vector) during a period of 5 days. (B) HeLa cells stably transfected with pBABEpuro (Vector) and pBABEpuro-SteA-3xFLAG (SteA L4) were cultured in serum-free medium for 24 h and cell death was determined by LDH release. Means from representative experiments carried out in triplicate are represented. Error bars ! 18! represent the standard deviations. The asterisk indicates a significant difference (p < 0.01) by the Student’s t test. Fig. 4. HeLa cells expressing SteA have a defect in cell-cell adhesion and cell migration. (A) HeLa cells stably transfected with pBABEpuro (Vector) and pBABEpuro-SteA-3xFLAG (SteA L4) were observed under an inverted microscope to detect differences in cellular shape (upper panels) and in cell colony shape (lower panels). (B) HeLa cells were transferred to poly-HEMA-coated plates. After 48 h of culture in suspension, cells were photographed under an inverted microscope. Depicted are control cells HeLa pBABEpuro (Vector), which form dense aggregates, and HeLa pBABEpuro-SteA-3xFLAG (SteA L4), which adhere to each other only very weakly. (C) Representative images of wound-healing assays showing that HeLa pBABEpuro (Vector) closed the wound 4 days after wounding, whereas HeLa pBABEpuro-SteA-3xFLAG (SteA L4) did not. Wounds are outlined. (D) Wound widths measured 48 h after wounding and normalized to the value obtained for HeLa pBABEpuro are graphed, showing means and standard deviations from six experiments. The asterisk indicates a significant difference (p < 0.01). Scale bars, 200 µm. ! 19! Table 1 Selected biological functions associated to a subset of differentially expressed genes in steA- transfected HeLa cells Upregulated genes Biological process or molecular function Proliferation Death/apoptosis ECMa Adhesion Migration Endocytosis CAV1 X X X CCND1 X CTGF X X X X X C5AR1 X X C5orf23 (NPR3) X EFEMP1 X LUM X MAP2K6 X SDC2 X Downregulated genes Biological process or molecular function Proliferation Death/apoptosis ECMa Adhesion Migration Endocytosis AKR1C2 X ALDH3A1 X CNTNAP3 X COL15A1 X X X CYFIP2 X X FURIN X X IER3 X MPDZ X MYH10 X X NFE2L2 X NOG X X RARRES1 X SLC7A11 X TGFB1 X X X X X TNS1 X UNC13B X aECM: extracellular matrix regulation. X indicates that this GO term appears associated to this gene. ! 20! Figure 1 ! 21! Figure 2 ! 22! Figure 3 ! 23! Figure 4 ! 24! Figure S1