Abstract
Human enteropathogenic bacteria has been reported to be transmitted by fresh vegitables. Shigella, which infects primates, is reportedly transmitted by fresh vegetables; however, its molecular interactions with plants have not been elucidated. Here, we show that four Shigella strains, S. boydii (S. b), S. sonnei, S. flexneri 2a (S. f 2a), and S. flexneri 5a (S. f 5a), proliferated at different levels in Arabidopsis thaliana. Microscopic studies revealed that these bacteria were present inside leaves and damaged plant cells. Green fluorescent protein (GFP)-tagged S. b and S. f 5a colonized in leaves only, and S. f 2a colonized both leaves and roots. Using mutants lacking type III secretion systems (T3SS), we found that T3SS of Shigella that regulate the pathogenesis of Shigellosis in humans also play a central role in proliferation in Arabidopsis. Strikingly, the immunosuppressive activity of two T3S effectors, OspF and OspG, were needed for the proliferation of Shigella in Arabidopsis. Of note, delivery of OspF or OspG effectors inside of plant cells upon Shigella inoculation was confirmed by using a split GFP system, respectively. These findings demonstrate that the human pathogen Shigella can proliferate in plants by adoption of immunosuppressive machinery for its original host human.
Introduction
Human hosts are primarily infected by the ingestion of water or food contaminated with bacteria from feces (the fecal-oral route), including ingestion of undercooked contaminated meats and improperly washed contaminated vegetables (Mead et al., 1999; Weir, 2002; Gupta et al., 2004). Fresh fruits and vegetables, such as lettuce, tomatoes, and green peppers, are responsible for the widespread transmission of food-borne infections by Salmonella or Shigella (Guchi and Ashenafi, 2010; Semenov et al., 2010; Gu et al., 2013). These observations suggest that human pathogenic bacteria use plants as alternative hosts as a stable environmental niche. Several human pathogenic bacteria, including Salmonella enterica serovar Typhimurium, Escherichia coli O157:H7, and Pseudomonas aeruginosa, are known to use plants as alternative hosts (Plotnikova et al., 2000; Semenov et al., 2010). These bacteria can attach to the plant surface, and actively invade and proliferate in plant tissues (Cevallos-Cevallos et al., 2012; Martínez-Vaz et al., 2014). In particular, several enteropathogenic bacteria, including E. coli O157:H7 and Salmonella, spread within plants through vascular tissues after infection via contaminated water (Solomon et al., 2002).
Shigella is a human-adapted pathogen that infects the host via multiple transmission routes. It is a non-motile, rod-shaped, facultative intracellular and invasive pathogen, very closely related to Escherichia coli. Based on the carbohydrate composition of the O-antigen, i.e., the polysaccharide component of the lipopolysaccharide molecule that is the major bacterial surface antigen, Shigella is classified into four serogroups. These have been given species designations, namely, S. dysenteriae 1 (serogroup A), S. flexneri (serogroup B), S. boydii (serogroup C), and S. sonnei (serogroup D) (Lindberg et al., 1991; Schroeder and Hilbi, 2008). Shigella spp. are important epidemic pathogens and serious public health concern in developed and developing countries. About 164,300 deaths in all age groups, and 54,900 deaths in children younger than 5 years, were reported globally in 2015 (Mortality and Causes of Death, 2016). However, the actual number of infections might be higher because mild symptoms are not reported (Mortality and Causes of Death, 2016), suggesting that the microorganism may employ a variety of survival strategies not only for human intestinal infection but also for survival in a non-adapted host. Although Shigella contamination has also been reported in plants (Naimi et al., 2003; Ohadi et al., 2013), it is not yet known whether the bacterium actively invades and/or proliferates inside the plant.
Unlike animals, plants have no adaptive immune system; instead, each cell possesses an innate immune system. The innate immune system in plants and animals recognizes and suppresses pathogens and has common features that are preserved throughout evolution. Plant pattern recognition receptors recognize conserved microbial or pathogen-associated molecular patterns; the pattern-triggered immunity (PTI) is activated via the mitogen-activated protein kinase (MAPK) cascades (Jones and Dangl, 2006). To suppress PTI, bacteria inject effector proteins into plant cells using type III secretion systems (T3SS). To counteract this PTI evasion response, the plant nucleotide binding-leucine rich repeat proteins recognize the pathogen effectors; effector-triggered immunity is then activated to accompany the hypersensitive response (Jones and Dangl, 2006). For human or animal intestinal bacteria to infect plants, the PTI must first be disabled. S. enterica serovar Typhimurium, similar to its activity in the mammalian host, uses T3SS to suppress plant immune responses (Schikora et al., 2011; Schikora et al., 2012). In particular, one of the T3S effector proteins of S. enterica, SpvC targets the MAPK signaling system as in animals to suppress the host PTI, when expressed in plants under the control of plant binary vector (Neumann et al., 2014).
Here, we examined the ability of four Shigella strains (S. s (Holt et al., 2012), S. b and S. f 2a (Wei et al., 2003), and S. f 5a (Onodera et al., 2012)) to proliferate in Arabidopsis plants. We found that the four strains invaded and proliferated differently in plant tissues. A Proliferation of mutants lacking the T3S effectors, i.e., noninvasive human strains, was reduced in planta. Reverse genetics and molecular biology experiments demonstrated that the immunosuppressive function of Shigella T3S effectors OspF and OspG was essential for Shigella proliferation in plants. Notably, we first observed the delivery of Shigella Type III effector proteins inside of plant cells by Shigella inoculation using the split GFP technology. These observations indicate that Arabidopsis may be useful as a model host for studying the pathogenesis of Shigella.
Materials and Methods
Plant materials and growth
Arabidopsis thaliana accession Columbia (Col-0) was used for Shigella infection. Briefly, Arabidopsis seeds were surface-sterilized in 70% (v/v) ethanol for 2 min followed by an incubation in 50% household bleach for 10 min. Seeds were washed extensively with sterile deionized water before placed on 1/2 Murashige and Skoog (MS) medium (Duchefa Biochemie, Haarlem, Netherlands) supplemented with 1% sucrose and solidified with 0.6% (w/v) agar (Murashige and Skoog, 1962). Nicotiana benthamiana plants were grown at 22 ± 3°C under a 16 h light/8 h dark cycle in plastic pots containing steam-sterilized mixed soil (2:1:1, v/v/v, soil/vermiculite/perlite) (Moon et al., 2016). To measure the plant immune response in terms of MAPK activity, 1 μM flg22 peptide (Cat No: FLG22-P-1; Alpha Diagnostics, Inc., San Antonio, TX, USA) was used as a positive control (Bethke et al., 2009).
Bacterial strains, growth conditions, and plasmids
The bacterial strains and plasmids used in the study are described in Table S1. Shigella and Pseudomonas strains harboring a plasmid pDSK-GFPuv were generated by electroporation, as described previously (Wang et al., 2007; Hong et al., 2016).
Shigella spp. were grown at 37°C in Luria-Bertani (LB) medium or tryptic soy agar containing 0.003% (w/v) Congo red dye (Sigma-Aldrich, St. Louis, MO, USA) (Runyen-Janecky and Payne, 2002). Pseudomonas syringae strains were grown at 28°C (with shaking at 200 rpm) in King’s B liquid medium (Sigma-Aldrich) containing appropriate antibiotics (King et al., 1954). As a non-pathogenic control, Escherichia coli DH5α was grown in LB medium at 37°C with shaking (Kennedy, 1971). Agrobacterium tumefaciens GV2260 was grown at 28°C in LB broth with shaking at 200 rpm (Shamloul et al., 2014).
The coding region of ospF or ospG was PCR-amplified using attB-containing PCR primers (Table S2). The PCR fragments were cloned into the pDONR™207 vector by BP recombination using the Gateway® BP Clonase™ II kit (Invitrogen, Carlsbad, CA, USA). The products were then transferred to pBAV178 (for AvrRpt2 fusion) or pBAV179 (for HA fusion) vectors by LR recombination (Gateway® LR Clonase™ II, Invitrogen). pBAV178, pBAV179, and pME6012 (empty vector control) were kindly provided by Dr. Jean T. Greenberg (University of Chicago) (Vinatzer et al., 2005).
Bacterial inoculation assay in planta
Arabidopsis seedlings (2 weeks old) grown in 1/2 MS medium were used for flood-inoculation (Ishiga et al., 2011). Briefly, 10 Arabidopsis seedlings in one container were incubated for 3 min with 35 ml of each bacterial strain suspension (5 × 106 or 5 × 105 cfu/ml) containing 0.02% Silwet L-77 (Lehle Seeds, Round Rock, TX, USA) or buffer. After the bacterial suspensions were removed by decantation, plates containing inoculated plants were incubated in a growth room (23 ± 2°C, 16 h light/8 h dark). Bacterial cell counts from inoculated plants were monitored as described previously (Ishiga et al., 2011). Three inoculated seedlings in one container were harvested by cutting the hypocotyls, and total fresh weight was measured. The cfu were normalized to cfu/mg using sample weight. The cfu of seedlings in three separate containers (as biological replicates) were measured. In addition, the bacterial population was evaluated in more than three independent experiments conducted successively under the same conditions.
To assess root invasion, 10-d-old Arabidopsis seedlings grown vertically in 1/2 MS medium were inoculated by dropping 2.0 μl of bacterial suspension (5 × 107 cfu/ml) onto the root tips. Symptoms were observed under white light, and bacterial proliferation was monitored at 5 dpi by observation of GFP-expressing bacteria under UV light. Three biological replicates were generated in separate plates, and three independent experiments were conducted under the same conditions.
To analyze the effector secretion, pBAV178 was used to transform P. syringae pv. tomato DC3000 (Pst) by electroporation (Cadoret et al., 2014). Bacterial suspensions (2.5 × 108 cfu/ml) were used to syringe-infiltrate Arabidopsis leaves; 2 d after the infiltration, the hypersensitive response was assessed by trypan blue staining (Koch and Slusarenko, 1990).
To test the virulence of Shigella effectors via Pst, plasmid pBAV179 was used to transform Pst; 5 × 107 cfu/ml bacterial suspensions containing 0.02% (w/v) Silwet L-77 were used to spray-infect Arabidopsis leaves. After infection, the plants were covered with a clear lid to maintain humidity and transferred to a growth room (22 ± 3 °C, 16 h light/8 h dark). The symptoms and bacterial proliferation were assessed at 4 dpi.
Shigella infection to visualize of effector secretion into plant cells
Transgenic Arabidopsis expressing Nucleus-targeted Nu-sfGFP1-10OPT were flood inoculated (5 × 105 cfu/ml) with S. f 5a, S. f 5a::OspF-sfGFP11 or S. f 5a::OspG-sfGFP11. To produce the sfGFP11-fused Shigella effectors, OspF or OspG was transferred to pEP119T by LR recombination. At specific time points, Arabidopsis leaf discs were observed under a confocal microscope (Park et al., 2017). Each experiment included at least three independent plants.
Microscopy
For SEM, flood-inoculated Arabidopsis leaves were fixed in 4% (w/v) paraformaldehyde and dehydrated in an ethanol series (30%, 50%, 70%, 96%, and 100%). The fixed leaves were then dried, coated with gold-palladium, and visualized using a scanning electron microscope (LEO 1455VP, Oberkochen, Germany) (Plotnikova et al., 2000). For TEM, flood-inoculated Arabidopsis leaves were cut off, fixed overnight in 2.5% (w/v) glutaraldehyde, post-fixed in 2% (w/v) osmium tetroxide, dehydrated in ethanol, and embedded in the resin. After staining in 2% (w/v) uranyl acetate and lead citrate, samples were observed under an electron microscope (Bio-TEM; Tecnai G2 Spirit Twin; FEI, USA) (Chae and An, 2016).
For fluorescence confocal microscopy, expression of GFP-labeled bacteria or GFP-tagged Shigella effector proteins in plants was observed under a Nikon laser scanning confocal microscope C2 (Nikon, Tokyo, Japan) using filter sets for GFP (λex, 488 nm; λem, 505–530 nm) or RFP (λex, 561 nm; λem, 570–620 nm). For each microscopic method, three leaves were used per treatment and at least three microscopic fields were observed for each leaf, including the control.
Expression of Shigella virulence genes in Arabidopsis plants
Total RNA was extracted from Shigella-infected leaves (from three plants) using RNAiso plus (Cat No: 9108; TaKaRa, Otsu, Japan), according to the manufacturer’s protocol. RT-PCR was performed using M-MLV reverse transcriptase (Invitrogen), according to the manufacturer’s instructions. Quantitative RT-PCR was carried out in a CFX Connect™ Real Time System (BioRad, Hercules, CA, USA) using iQ™ SYBR® Green Supermix (BioRad) and primers specific for target genes (ipaB, ipaC, icsA, icsB, virB, and virF; Table S1 (Bando et al., 2010)). The qRT-PCR results were normalized to the expression of 16s rRNA.
Immunoblotting
Total protein was extracted from Shigella- or Agrobacterium-infected leaves (from three plants) using denaturing extraction buffer [150 mM NaCl, 50 mM Tris-HCl (pH 7.5), 5 mM EDTA, 0.1% Triton-X, 1× protease inhibitor cocktail (Roche, Basel, Switzerland), 0.4 M DTT, 1 M NaF, and 1 M Na3VO3]. The extracted proteins were separated on 12% SDS-PAGE gels and transferred to a PVDF membrane (Pierce, Rockford, IL, USA). Antibodies specific for phospho-p44/p42 MAPK ERK1/2 (Cat No: 4377; Cell Signaling Technology, Danvers, MA, USA) and ERK1 (Cat No: sc-94; Santa Cruz, Dallas, TX, USA) were used for immunoblot analyses. Target proteins were detected using ECL plus reagent (GE Healthcare, Wauwatosa, WI, USA) and visualized using an Alliance 9.7 chemiluminescence imaging system (UVITEC, Cambridge, UK).
Statistical analysis
All data are expressed as the mean ± SD. The statistical significance of bacterial cell growth in infected plants was examined using Student’s t-test (Microsoft Office Excel) and ANOVA (SPSS v.18; IBM, Armonk, NY, USA) (Moon et al., 2016). Asterisks and letters indicate significant differences between samples (P < 0.05).
Results
Four Shigella spp. strains interact differently with Arabidopsis
To observe the behavior of the human pathogen Shigella in plants, we investigated the interaction of four Shigella spp. strains representing three serogroups (S. b, S. s, S. f 2a, and S. f 5a), with the model plant A. thaliana. Shigella is a water-borne pathogen, with infection spreading via contaminated water (Pandey et al., 2014). Hence, we chose to use a flood-inoculation approach (Ishiga et al., 2011), which is thought to mimic natural inoculation. Infection of the phytopathogen Pseudomonas syringae pv. tomato DC3000 (Pst) was observed as a positive control and Pst ΔhrcC, a mutant lacking the T3SS of Pst, and the non-pathogenic bacterium E. coli DH5α were used as negative controls for infection. When 2-week-old Arabidopsis seedlings were inoculated with S. s or S. f 2a, clear infections were observed, such as yellowing and necrosis of the leaves, while mild symptoms were detected after inoculation with S. b and S. f 5a (Figure 1A) (Liu et al., 2015).
In addition to observing the symptoms, the bacterial growth in planta was also evaluated to detect initial plant adherence and proliferation. Early attachment of all Shigella strains and DH5α was more than 10 times lower than that of Pst and Pst ΔhrcC (Figure 1B; Day 0, grey bars). At 3 days post-inoculation (dpi), the cell number of all strains significantly increased than initial bacterial cells, although the extent of cell amplification differed depending on the Shigella strain (Figure 1B; black bars). Notably, S. s cell numbers increased over 105 times, comparable to the proliferation of Pst. By contrast, the interaction between S. f 5a and Arabidopsis was similar to that between the plant and non-pathogenic DH5α.
Previous studies demonstrate that several human pathogenic bacteria, including E. coli O157:H7 and Enterococcus faecalis, invade the leaves and roots of A. thaliana (Jha et al., 2005; Deering et al., 2012). Therefore, to determine whether Shigella strains invade through A. thaliana roots, we dropped bacterial solutions onto root tips and observed symptoms in Arabidopsis plants for 14 d. Interestingly, all the strains inhibited Arabidopsis root growth (Figure 1C). Interestingly, S. b strain caused severe root growth inhibition unlike the mild disease symptom of the leaves after the infection of this strain (compare Figure 1A, 1C). Inoculation of S. f 5a caused slight inhibition of root growth but caused much less damage to the plant than other strains. Taken together, these results indicated that, although the proliferative capacity of the Shigella strains differs, the cells can invade and colonize the apoplast of plants, thereby causing structural damage to the host.
Penetration and subsequent internalization of Shigella spp. into Arabidopsis
Since we observed that Shigella proliferate and induce disease-like symptoms in Arabidopsis, we examined whether the bacterium multiplies on the leaf surface and in the intercellular space (apoplast) by scanning electron microscopy (SEM) and transmission electron microscopy (TEM), respectively. Pst, known to infects plants via open stomata (Panchal et al., 2016), colonized guard cells at 24 h post-infection (Figure 2A). Similarly, all tested Shigella strains clustered around guard cells and the surface of epidermal cells. S. s and S. b formed relatively broad clusters in the surrounding areas, including guard cells (Figure 2A). In particular, S. b and S. f 2a intensely colonized guard cells (Figure 2A), leading us to speculate that they enter plants via open stomata, similar to Pst.
Most plant pathogenic bacteria infect and colonize the apoplast (Sattelmacher, 2009; Gao et al., 2016). Therefore, we investigated whether Shigella is capable of intercellular colonization and causing damage to plant cells. Indeed, the TEM images revealed colonization of Shigella in the intercellular spaces and attachment to the host cell walls (Figure 2B). The presence of the microbes in the intercellular space resulted in the alteration of the host organelle structure, such as the separation of the plasma membrane from the cell wall, the liberation of cell organelles, and disruption of chloroplasts (Figure 2B (Gao et al., 2016)). This effect was most pronounced in plant cells inoculated with S. b and S. s; further, S. b and S. s were more commonly found in the intercellular spaces than S. f 2a and S. f 5a.
To observe internalization of bacterial cells in the plant via root cells, we attempted to label the four Shigella strains with a green fluorescent protein (GFP) and obtained S. b, S. f 2a, and S. f 5a strains labeled with GFP successfully. GFP labeling did not affect the bacterial growth in plants (Figure S2). Bacterial suspensions were dropped onto the root tips of vertically grown Arabidopsis plants. Five days later, whole plants and root tissues were photographed under ultraviolet (UV) light to observe the distribution of fluorescently labeled bacteria (Figure 3A). In accordance with the disease phenotypes observed, Pst (which is a foliar pathogen) exhibited strong fluorescence in the leaves, despite the fact that it was applied to root tips (Figure 3A). To observe Pst in root tissues, inoculated roots were disinfected with hypochlorite solution and observed under a fluorescence microscope. GFP fluorescence was observed only in epidermal cells after application of Pst to root tips (Figure 3B). Consistantly, this finding indicates that Pst infects throughthe leaf tissues, not by invading root tissues or colonizing the roots. Interestingly, when GFP-labeled S. f 2a was applied to the root tip, strong green fluorescence was observed in both roots and leaves, while other GFP-labeled Shigella inoculated plants showed only weak fluorescence in leaves (Figure 3A). In root tissues, GFP fluorescence was observed in root endodermal cells only in S. f 2a-treated plants (Figure 3B). Taken together, these results indicate that Shigella invades plant leaves and roots in a strain-dependent manner, and then moves along the surface and through the internal vascular tissues of the plant.
Shigella T3S effectors are necessary for attachment and multiplication in Arabidopsis
T3SS is the crutial determinant for the virulence of many Gram-negative bacteria, including animal and plant pathogens (He, 1998; Ogawa et al., 2008; Buttner and He, 2009). Thus, we investigated if pathogenic proteins for animal host infections are required for invasion and multiplication of Shigella in plants. To study the biological role of the T3SS in the interaction between Shigella and plants, we used noninvasive variants of S. f 2a and S. f 5a (strains Δvp and BS176, respectively) (Sansonetti et al., 1982; Wenneras et al., 2000; Shim et al., 2007). To facilitate observation of bacterial invasion in living plants, the noninvasive variants of the Shigella strains were also labeled with GFP, and bacterial growth and expression of effector proteins were verified (Figure S1-S3). Bacterial proliferation in plants after inoculation with Δvp or BS176 strains was 10 times lower than that after inoculation by parental Shigella strains; similar results were observed for GFP-labeled Shigella strains and mutants (Figure 4A). To examine the involvement of the T3SS in bacterial invasion of the plant surface, leaves were flood-inoculated with GFP-labeled bacteria, and leaf surfaces were observed 24 h later (Figure 4B). Examination under UV light revealed that the levels of GFP-labeled S. f 2a and S. f 5a on leaf surfaces were higher than those of Δvp and BS176, especially in open stomata (Figure 4B). We also generated a non-pathogenic mutant of S.s and have tested if the severe infection of S.s strain is T3SS dependent (Figure S4). When plants were inoculated with this T3SS-deficient S. s strain, disease phenotypes and bacterial growth were significantly reduced (Figure S4), as observed with Δvp and BS176. Altogether, these results suggested that the T3SS of Shigella that operate during infection in human is also required for the colonization in plants.
To further investigate involvement of Shigella T3S effectors in plant interactions, we measured expression of critical virulence effectors related to human pathogenesis, including ipaB, ipaC, icsA, icsB, virB, and virF, in S. f 2a-, S. f 5a-, Δvp-, and BS176-infected Arabidopsis plants (Bando et al., 2010). These effectors play a role in mammalian cell lysis (ipaB, ipaC), intracellular spread (icsA, icsB), and regulation of virulence factor (virF, virB) (Ogawa et al., 2008). Total RNAs were isolated from Arabidopsis leaves at 6 and 12 h after Shigella inoculation, and changes in expression of virulence genes were confirmed by quantitative RT-PCR (Figure 4C). Expression of all Shigella virulence genes examined in WT S. f 2a- or S. f 5a-inoculated Arabidopsis leaves increased. Interestingly, induction was higher and faster in S. f 2a-treated plants than in S. f 5a-treated plants, which is in agreement with the earlier results showing that S. f 2a was more pathogenic to plants than S. f 5a. Expression of virulence genes was not detected in plants inoculated with non-pathogenic mutants Δvp or BS176, similar to buffer-treated control plants (Figure 4C), suggesting that common virulence factors regulate interactions with both plants and human intestinal cells.
To investigate the plant innate immune responses to Shigella inoculation, MAPK phosphorylation was evaluated. The activation of MAPK by phosphorylation is a conserved response of the earliest microbe-triggered immune signaling in both plants and animals (Zipfel, 2009). The flg22 peptide is a representative microbe-associated molecular pattern in plants (Bethke et al., 2009). In plants treated with flg22, pronounced MAPK phosphorylation was apparent within 5 min of treatment and this response lasted up to 30 min (Figure 4D). On the other hand, MAPK phosphorylation in plants treated with S. f 2a or S. f 5a was reduced; from 15 min on, it was strongly suppressed and almost completely disappearing after ca. 30 min (Figure 4D). Meanwhile, in plants treated with the virulence plasmid-deficient mutants, Δvp or BS176, MAPK activation was recovered, in contrast to S. f 2a- or S. f 5a-treated plants, although the degree and duration of the phosphorylation were lower than those elicited by the flg22 treatment (Figure 4D). These observations indicated that Shigella suppresses the innate immunity of Arabidopsis via its T3SS.
Suppression of immune signaling in Arabidopsis plants by Shigella T3S effectors OspF and OspG
Shigella OspF has been reported to inhibit MAPK signaling, which is conserved in plants and animals (Arbibe et al., 2007; Li et al., 2007). OspG is an essential immunosuppressive effector protein secreted at the later stages of infection; this protein interferes with activation of the NF-κB pathway, which is absent from plants (Kim et al., 2005). Therefore, we examined whether OspF and OspG have virulence activity in plants by introducing them heterologously into a phytopathogen, Pst, and monitoring its pathogenicity. First, we used an AvrRpt2-derived T3SS reporter system (Figure S5A; (Mudgett, 2005). AvrRpt2101-255 is sufficient to induce cell death but lack in delivery to the plant cells. When Arabidopsis leaves were syringe-infiltrated with various Pst strains which express effectors fused to AvrRpt2101–255, we observed that Pst producing OspF:AvrRpt2101–255 or OspG:AvrRpt2101–255 successfully induced cell death at 1 day after delivery implying that both OspF and OspG are successfully delivered into Arabidopsis cells via Pst T3SS (Figure 5A). Next, we examined whether the virulence of Pst was increased by expression of Shigella OspF or OspG. Production of OspF:HA and OspG:HA by Pst was confirmed by immunoblotting with an anti-HA antibody (Figure S5B, S5C). Plants infected with Pst cells producing OspF: HA or OspG: HA showed more severe symptoms than plants infected with the empty vector control (Figure 5B). In addition, the number of bacterial cells was 10 times higher than in the plants infected with the empty vector control (Figure 5C). Taken together, these results indicate that OspF and OspG preserve their function as virulence proteins in plant hosts.
On the other hand, in vitro studies demonstrated that OspF exerts a phosphothreonine lyase activity and irreversibly removes phosphate groups from MAPK (Arbibe et al., 2007; Li et al., 2007). To investigate whether the virulence associated with Shigella OspF in plants was linked to the same mechanism of action as in animals, we used a phenotypic screening system involving a MEK2 (a tobacco MAP kinase kinase2) mutant, MEK2DD (Yang et al., 2001; Kim and Zhang, 2004). MEK2DD, a constitutively active mutant of MEK2, induces cell death when overproduced in plant leaves In this experiment, an HA:MEK2DD clone fused to an HA epitope tag at the N-terminus was used to monitor expression of MEK2DD. As expected, co-production of HA:MEK2DD and GFP (control) resulted in pronounced cell death (Figure S6A). Co-production of OspF:GFP, but not OspG:GFP, and MEK2DD completely inhibited the MEK2DD-induced cell death (Figure S6A). The production of the two effectors fused with GFP was then evaluated in the MEK2DD-producing plant leaves; the production of OspF:GFP was apparent, but that of OspG:GFP was not (Figure S6B), even though both proteins were stably produced in the absence of MEK2DD (Figure S6B). The degradation of the OspG protein might be associated with the activation of MAPK. Indeed, as assessed by immunoblotting with specific anti-phosphorylated MAPK antibodies, MAPK phosphorylation was very weak in the OspF:GFP-producing plant samples in comparison with GFP- or OspG:GFP-producing plant samples (Figure S6B). The production of MEK2DD protein was confirmed in all samples using anti-HA antibodies (Figure S6B). These observations strongly suggest that the Shigella effector OspF inhibits plant immune responses by inhibiting activation of MAPK (as in humans), and that OspG induces immunosuppression in plants by targeting distinct MAPK pathways.
OspF or OspG affects Shigella proliferation in plants
To confirm the role of the OspF or OspG proteins in the interaction between Shigella and plants (as in human cells), we inoculated plants with S. f 5a mutants lacking the OspF or OspG proteins and examined their growth. The growth of mutants lacking ospF or ospG was as deficient as that of the virulence plasmid-deficient mutant BS176 (Figure 6A). Reduced growth of S. f 5a ΔospF or S. f 5a ΔospG mutants was entirely restored by complementation of the mutation, indicating that these two effector proteins are indeed important for bacterial growth in plants (Figure 6A). Next, we monitored the activation of MAPKs to determine whether plant immune suppression was affected by the deletion of ospF or ospG. As shown in Figure 6B, S. f. 5a ΔospF induced stronger phosphorylation of MAPKs than wild-type S. f 5a; this was offset by complementation with OspF. By contrast, phosphorylation of MAPK in S. f 5a ΔospG-inoculated plants was no different from that in plants inoculated with wild-type S. f 5a (Figure 6B). These results are consistent with the inhibition of MAPK-induced immune responses by OspF, but not by OspG, using the plant expression binary vector described above (Figure S6). Taken together, these results indicate that Shigella effectors OspF or OspG play an important role in increasing bacterial proliferation in both plant and animal hosts.
Delivery of OspF and OspG into plant cells during Shigella inoculation
To visualize direct delivery of type III effectors to plant cells by Shigella, we took advantage of a newly developed split superfolder green fluorescent protein (sfGFPOPT) system (Park et al., 2017). In this system, the non-fluorescent GFP protein (GFP1-10) with β –strand 1 to 10 is expressed in the host cell and the bacterial effector is linked to the 11th β-strand (GFP11). Only when the effector fused with GFP11 is introduced into plant cells via bacterial T3SS, fluorescent signals are induced through GFP reconstitution (Park et al., 2017; Figure 7A). First, in order to select a host plant expressing GFP1-10 capable of exploring the delivery of OspF and OspG, we observed the subcellular localization of OspF:GFP and OspG:GFP by expressing them in Nicotiana benthamiana leaves using the Agrobacterium system (Figure S7). The fluorescence signal for both proteins was strong in the plant cell nucleus; OspF:GFP fluorescence was also observed along the cytoplasmic membrane, and punctate OspG:GFP fluorescence was observed in the cytosol (Figure S7C). However, immunoblot of OspG:GFP expression in Nicotiana benthamiana detected OspG:GFP as well as free GFP (Figure S7B), suggesting that overexpression of free GFP might be responsible for the cytosolic GFP expression. Based on these results, we choose Arabidopsis plants expressing sfGFP1-10 (sfGFP1-10OPT) in nucleus or cytosol and the OspF and OspG were fused with 11th β-strand of sfGFP, respectively (Figure S8A). GFP signals were observed at nucleus in both S. f 5a::OspF-sfGFP11 and S. f 5a::OspG-sfGFP11 flood-inoculated Arabidopsis leaves at 3 hpi, respectively, while control S. f 5a WT did not (Figure 7B). In contrast, when Arabidopsis seedlings expressing cytosolic sfGFP1-10OPT were flood inoculated with S. f 5a producing OspF-sfGFP11 or OspG-sfGFP11, we detected no reconstituted GFP signals in all tested bacteria (Figure S8B). This suggests that Shigella effectors OspF and OspG were successfully secreted into Arabidopsis cells through the Shigella T3SS. In the host human cell, OspF localization is nuclear (Zurawski et al., 2006; Arbibe et al., 2007) and that of OspG is nuclear and cytoplasmic (Kim et al., 2005; Zhou et al., 2013; de Jong and Alto, 2014), confirming that the subcellular localization in plant cells is similar to that of animal cells. Taken together, the Shigella effectors OspF and OspG can be delivered into plant cells through T3SS.
Discussion
In the current study, we investigated the interaction of the human pathogenic bacterium, Shigella, with an alternative host, the Arabidopsis plant. We demonstrated that four Shigella strains, S. b, S. s, S. f 2a, and S. f 5a, invade and colonize Arabidopsis to different extents. Different symptoms and bacterial proliferation rates for different Shigella strain suggested that the variability of the plant interaction mechanisms among the strains, e.g., the adherence and multiplication, might contribute to the differences. Differences in the nutritional requirements of bacterial strains may constitute another reason for the differences in the growth rates within the Arabidopsis host.
It has been demonstrated that, in human, Shigella initially enters the epithelial layer via the M cells through transcytosis, leading to the invasion of the basolateral surfaces of the intestinal epithelial cells. A subsequent gut inflammation leads to the activation of the innate immune response (Phalipon and Sansonetti, 2007). We demonstrated, in addition to the human host, Shigella invades Arabidopsis through vascular tissues and leaf stomata pores (Figure 2-4). In particular, S. s and S. b formed relatively wide clusters in the surrounding areas, including the guard cells (Figure 2), similar to Pseudomonas syringae which has built a host-pathogen relationship with Arabidopsis (Panchal et al., 2016). Interestingly, we found that the four studied-strains of Shigella associate with the plant cells and induce different plant responses. The bacterial loads of S. s in planta were relatively higher than those of the other strains. By contrast, inoculation of S. f 5a was associated with lower bacterial proliferation and less severe symptoms than observed for other strains (Figure 1-3). S. f 2a and S. f 5a, which belong to the same serogroup (Lindberg et al., 1991), elicited distinctly different plant responses with respect to disease symptoms, suggesting that the virulent effectors may play a relatively more important role in Shigella-plant interactions than PAMPs. These observations also indicate that specific plant immune systems may be useful in the search for novel virulence factors expressed by different Shigella strains.
Many Gram-negative bacterial pathogens utilize common infection strategies to colonize and invade plant and animal cells, and pathogenicity appears to depend on highly conserved T3SSs, which deliver the effector proteins to host cells (Buttner and Bonas, 2003). By using avirulent mutant strains, we were able to show that effectors that regulate the pathogenesis of shigellosis in humans also play a central role in regulating interactions with Arabidopsis. We showed that secretion of T3S effectors is required for the Shigella proliferation and attachment in plants (Figure 4, Figure S4). Furthermore, the effector proteins impacted MAPK-dependent/independent immune responses in Arabidopsis (Figure 4, 6). These observations further support the suggestion that the function of the main effector proteins of Shigella appears to be conserved in plant and animal hosts, and that this contributes to bacterial intracellular survival or suppression of the host defense against the pathogens. Reduced colonization of T3SS-deficient pathogenic E. coli in plants was previously reported, they suggested that E. coli uses the T3SS apparatus for attachment to leaves, rather than for bacterial growth inside plants (Shaw et al., 2008). The relevance of T3SS for multiplication of Salmonella in plants remains unclear due to the different effects of T3SS function on Salmonella-plant interactions (Iniguez et al., 2005; Schikora et al., 2011; Shirron and Yaron, 2011).
In general, animal bacterial pathogens with short needle lengths have been thought to be difficult to secrete T3S effectors on plant cells due to their cell walls and have been regarded as a major cause of not infecting plants. Recently, however, several studies have reported that the length and width of bacterial T3SS depend on the type of host cell, the environment, and the infection cycle of the bacteria (Deane et al., 2010). In this paper, we showed direct evidence that the effector proteins of Shigella were delivered into plant cells through T3SS utilizing recently developed the split GFP technology (Figure 7). Further study, identifying plant factors that trigger and control Shigella’s T3SS assembly will be a great help in developing a strategy to prevent the spread of Shigella through plants.
Expression of the T3SS of Shigella is regulated at the transcriptional level and is activated at a permissive temperature (≥ 32°C) (Tobe et al., 1991; Campbell-Valois and Pontier, 2016). We were able to observe the expression of the T3SS genes of Shigella under temperatures at which plants grow (22 ± 3°C) (Figure 5A). A recent study showed that elevation of the temperature increases T3SS-mediated virulence of the phytopathogen Pst in plants, which is in contrast with the negative effect of high temperature on expression of the T3SS of Pst in vitro (Huot et al., 2017). Regardless of the temperature of host cells, it will be interesting to determine whether Shigella regulates T3SS gene expression in vivo and to identify factors that influence T3SS gene expression other than plant temperature.
Salmonella strains capable of proliferating on plant leaves and actively entering plant tissues, root hairs, or trichomes were recently shown to exhibit virulence in animals (Barak et al., 2011; Golberg et al., 2011; Schikora et al., 2011). We demonstrated that Shigella strains actively colonize the surface of and inside Arabidopsis leaves and root tissues (Figure 1–3) and that bacteria recovered from plants maintain expression of pathogenic proteins (Figure S3). Collectively, these findings suggest that, similar to Salmonella, Shigella-inoculated plants are a serious risk to food safety and that contamination of plants is another route underlying infection of Shigella, an important human pathogenic bacterium. In this study, we only observed plants artificially inoculated with Shigella in a laboratory environment. Therefore, to confirm the food safety concern surrounding Shigella-inoculated plants, the ability of Shigella to infect a variety of crops grown in the field should be tested. The pathogenicity of plant-isolated Shigella in animals should also be investigated.
The current study provides new insights into host invasion mechanisms utilized by Shigella to interact with an alternative host, the plant Arabidopsis. Studying trans-kingdom pathogenesis involving human-adapted pathogens, such as Shigella, may uncover novel pathogenic mechanisms uniquely activated in response to specific hosts. When we isolated the two Shigella effectors OspF and OspG, and produced them in plant cells, their localization coincided with that in the animal cells (Figure 7 and S6C), and it was apparent that the production of both proteins increased the virulence of plant pathogens (Figure 6). In addition, we confirmed that OspF inhibits the innate immune response of plants via the same enzymatic activity as in animals (Figure 5 and S6). In animals, OspG inhibits the host immune response by inhibiting the activity of NF-κB by blocking degradation of IκB (Ashida et al., 2015). Plants possess an IκB-like protein called NIM1 (Ryals et al., 1997); however, no other published studies have investigated whether an NF-κB–induced immune response exists in plants. In the current study, we demonstrated the ability of OspG to increase the phytopathogenicity of non-Shigella bacteria (Figure 5), and also observed degradation of OspG:GFP upon constitutive activation of MAPK signaling (Figure S6B). The existence of a plant immune signaling pathway similar to that of animal NF-κB, which would also be the target of OspG, may hence be assumed. Characterization of the previously unrecognized stress-activated mediators of the innate immunity in plants upon infection with food-borne pathogens, would help us to define the defensive functions of plants. Finally, the characterization of plants as an alternative host for food-borne pathogens will be critical in developing effective means to prevent their transmission and disease.
Author contributions
JMP, SHJ, and JL conceived and designed the study. SHJ, JL, and DHL carried out the experiments. JMP, SHJ, JL, DWK, and EP analyzed the data. JMP, SHJ, JL, and EP wrote the manuscript. JMP, SHJ, EP, CMR and DC edited and discussed the manuscript and all authors agreed with the final version.
Supporting Information
Additional Supporting Information may be found online in the Supporting Information tab for this article:
Fig. S1 The growth of Shigella strains producing GFP or virulence-deficient mutants is not affected in vitro growth.
Fig. S2 Comparison of Bacterial proliferation in Arabidopsis inoculated with Shigella and GFP-labeled Shigella by flooding
Fig. S3 Expression of effector proteins of Shigella recovered from inoculated plants.
Fig. S4 Spontaneous type III secretion system-deficient mutant of S. sonnei is reduced growth and symptoms in Arabidopsis.
Fig. S5 Expression of OspF or OspG in Pst for virulence assay.
Fig. S6 Co-expression of OspF:GFP suppresses tobacco MEK2DD-triggered cell death.
Fig. S7 Expression GFP fused OspF or OspG in plant.
Figure S8. Delivery of OspF or OspG fused with GFP11 through T3SS
Table S1 Bacterial strains and plasmids used in the study.
Table S2 Sequences of PCR primers used for Gateway cloning.
Table. S3. PCR primer list and sequences used for qRT-PCR.
Acknowledgments
We thank Dr. Myung Hee Kim for initial support with the preparation of pathogenic Shigella strains, Dr. Jean T. Greenberg for providing the pBAV178, pBAV179, and pME6012 plasmids, and Dr. Cha Young Kim for providing the HA:MEK2DD clone. This work was supported by the KRIBB Initiative Program and the Basic Research Program of National Research Foundation of Korea grant (NRF-2017R1A2B4012820 to JMP and NRF-2018R1A5A1023599 to DC) funded by the Korea government (MSIT).