Expression of Francisella pathogenicity island protein intracellular growth locus E (IglE) in mammalian cells is involved in intracellular trafficking, possibly through microtubule organizing center

Abstract Francisella tularensis is the causative agent of the infectious disease tularemia and is designated a category A bioterrorism agent. The type VI secretion system encoded by the Francisella pathogenicity island (FPI) is necessary for intracellular growth; however, the functions of FPI proteins are largely unknown. In this study, we found that the FPI protein intracellular growth locus E (IglE) showed a unique localization pattern compared to other FPI proteins. Deleting iglE from Francisella tularensis subsp. novicida (F. novicida) decreased intracellular growth. Immunoprecipitation and pull‐down assays revealed that IglE was associated with β‐tubulin. Additionally, GFP‐fused IglE colocalized with microtubule organizing centers (MTOCs) in 293T cells. The iglE deletion mutant was transferred with dynein toward MTOCs and packed into lysosome‐localizing areas. Conversely, the wild‐type F. novicida exhibited intracellular growth distant from MTOCs. In addition, IglE expressed in 293T cells colocalized with dynein. These results suggest that IglE helps to prevent dynein‐ and MTOC‐mediated intracellular trafficking in host cells to inhibit the transport of F. novicida toward lysosomes.

2012; Eshraghi et al., 2016), the molecular mechanisms underlying the functions of these proteins are poorly understood.
In this study, we carried out an expression analysis of FPI proteins and found that the F. novicida intracellular growth locus E (IglE) shows unique localization and is associated with microtubule-organizing centers (MTOCs) to modulate membrane trafficking for the intracellular growth of the bacterium.

| Bacterial strains and culture conditions
F. novicida U112 was obtained from the Pathogenic Microorganism Genetic Resource Stock Center (Gifu University). F. novicida was cultured aerobically at 37°C in a chemically defined medium (CDM) (Nagle, Anderson, & Gary, 1960) or in a brain-heart infusion broth (Becton, Dickinson and Company, Franklin Lakes, NJ) supplemented with cysteine (BHIc) (Mc Gann et al., 2010) containing 1.5% agar (Wako Laboratory Chemicals, Osaka, Japan).  Table S1 shows the primer sets and templates used to construct plasmids used in this study. PCR was carried out using KOD-Plus-Neo polymerase (Toyobo, Osaka, Japan), and ligation was performed with the Ligation High Ver. 2 kit (Toyobo) or the In-Fusion HD Cloning Kit (Takara Bio, Otsu, Japan). Plasmids were transformed into F. novicida by cryotransformation (Pavlov, Mokrievich, & Volkovoy, 1996).

| Plasmid construction, transformation, and transfection
Briefly, bacterial cells were suspended in transfer buffer (0.2 M MgSO4, 0.1 M Tris acetate [pH 7.5]) with 1 μg of plasmid DNA. The bacterial cells were frozen in liquid nitrogen, thawed at room temperature, and then cultured in CDM. Then, bacterial cells were collected and cultured on BHIc plates containing 50 μg/ml kanamycin or 2.5 μg/ml chloramphenicol. Plasmids were transferred into cell lines with FuGENE HD (Promega, Madison, WI) according to the instruction manual.

| Construction of F. tularensis iglE and dotU mutants
The ΔdotU mutants of F. novicida were generated by group II intron insertion using the TargeTron Gene Knockout System (Sigma-Aldrich) modified for Francisella species (Rodriguez, Yu, Davis, Arulanandam, & Klose, 2008), as described previously (Uda et al., 2014). Briefly, 2 μg of each pKEK-DotU was transformed, and bacterial cells were precultured in CDM at 30°C for 6 hr. Then, the cells were collected and cultured on BHIc plates containing 50 μg/ml kanamycin at 30°C. Mutagenesis was confirmed using PCR to detect the 915-bp insertion. To remove the plasmids, mutants cells were further cultured on BHIc plates without antibiotics at 37°C.
The ΔiglE mutant was constructed by homologous recombination. To make the suicide vector pFRSU, the promoter region of sacB and antibiotic-resistance marker of pSR47s (Merriam, Mathur, Maxfield-Boumil, & Isberg, 1997) were replaced with the bfr promoter of pNVU1 (Uda et al., 2014) and the kanamycin-resistance gene kanR from pKEK1170 (Rodriguez et al., 2008), respectively. The upstream and downstream regions of iglE (1.5 kb each) were cloned into the SalI site of pFRSU to make pFRSU-IglE. One microgram of pFRSU-IglE was transformed into F. novicida, and the cells were cultured on BHIc plates containing 50 μg/ml kanamycin. Isolated bacteria were cultured in CDM without antibiotics for 6 hr and then plated on BHIc plates containing 5% sucrose. The deletion of the iglE gene was confirmed by PCR.

| Immunoblotting
To generate an antiserum against IglE, rabbits were immunized with the C+TGKNEFPLDKDIKD peptide. The peptide and antiserum were prepared by Eurofine Genetics (Tokyo, Japan). F. novicida was cultured in CDM containing 5% KCl to an OD 595 of 0.25. The culture supernatants were desalted to remove KCl with Amicon Ultra filters (Merck Millipore, Billerica, MA) and concentrated fivefold. Samples were mixed with SDS sample buffer (Thermo Fisher Scientific, Waltham, MA). Fifteen microliters of sample was loaded onto a NuPAGE Novex 4%-12% Bis-Tris Gel (Thermo Fisher Scientific) and separated by SDS-PAGE. Separated proteins were transferred onto a polyvinyl difluoride (PVDF) membrane (Merck Millipore). The membrane was treated with anti-IglE antiserum (1:100) or anti-PdpC antibody (Chong et al., 2008;Uda et al., 2014), a generous gift from Dr. J. Celli, followed by the treatment with HRP-conjugated anti-rabbit IgG (ab6717, 1:20,000; Abcam, Cambridge, UK). Proteins were detected with the ECL Prime Western Blotting System (GE Healthcare, Buckinghamshire, UK) and the LAS-4000 mini Imaging System (Fujifilm Life Science, Tokyo, Japan).

| Intracellular growth assay
THP-1 cells (4 × 10 5 cells/well) were preincubated in a 48-well tissue culture plate with 100 nM of phorbol myristate acetate (PMA) for 48 hr. F. novicida strains were added at a multiplicity of infection of 1. These plates were centrifuged for 10 min at 300× g and incubated for 30 min at 37°C. Then, THP-1 cells were washed twice with RPMI1640 medium, and extracellular bacteria were killed with a 60-min gentamicin (50 μg/ml) treatment. To measure the intracellular growth of F. novicida, the THP-1 cells were incubated in fresh medium at 37°C for the indicated time, washed three times with phosphate-buffered saline (PBS), and then lysed with 0.1% Triton X-100 in CDM. Colony-forming units were determined by serial dilution on BHIc plates.
Samples were suspended in SDS sample buffer and heated at 70°C for 15 min. Fifteen microliters of sample was loaded onto a NuPAGE Novex 4%-12% Bis-Tris Gel, and the proteins that co-precipitated with IglE-GFP were separated by SDS-PAGE followed by staining with Quick-CBB PLUS (Wako Laboratory Chemicals). Peptide mass fingerprinting was performed according to the method of Yoshino et al. (Yoshino, Oshiro, Tokunaga, & Yonezawa, 2004). Briefly, the CBB-stained bands obtained from SDS-PAGE were excised and sliced into small strips. To remove the CBB, the strips were incubated in 50% methanol and 5% acetic acid for 1 hr and washed twice with water. The strips were dehydrated by incubation with 100% acetonitrile. To alkylate the proteins, the strips were incubated at 60°C for 1 hr with 10 mM dithiothreitol in 100 mM ammonium hydrogen carbonate followed by treatment at room temperature for 30 min with 55 mM iodoacetamide (Nacalai Tesque, Kyoto, Japan) in 100 mM ammonium hydrogen carbonate. In-gel trypsin digestion was performed by incubating a gel with 10 μg/ml trypsin (Promega). The digested peptides were eluted using 5% formic acid (Wako Laboratory Chemicals). The peptides were desalted with ZipTip C18 Pipette Tips (Merck Millipore), spotted onto sample plates, and mounted with saturated α-cyano-4-hydroxycinnamic acid (Nacalai Tesque) in 50% acetonitrile and 0.1% trifluoroacetic acid. An Autoflex mass spectrometer (Bruker Daltonics, Billerica, MA, USA) was used to measure the molecular weights of peptides. The reference database was searched by MASCOT software (Science Matrix, London, UK).

| Pull-down assay
The iglE gene from F. tularensis subsp. tularensis SCHU P9 was cloned

| Statistical analysis
One-way analysis of variance was used to compare the results expressed as the means and standard deviations. Differences between the groups were determined by multiple comparisons using the Bonferroni/Dunnett method. The differences were considered significant at p values < 0.01.

| IglE shows unique localization
Among the FPI proteins, 8 of them (PdpA, IglE, VgrG, IglF, IglI, IglJ, PdpE, and IglC) are considered to be secreted into the host cytosol (Bröms et al., 2012). To elucidate the function of FPI proteins secreted by F. novicida, we performed a comprehensive expression analysis of these eight FPI proteins. We assayed the localization of these proteins in host cells by expressing GFP fused to FPI proteins in 293T cells ( Figure S1). Among the 8 FPI proteins, only IglE showed unique localization-large foci near the nucleus and dot foci ( Figure 1a). The fractions of cells containing large or dot foci were 46.7 ± 4.2% and 86.0 ± 5.3%, respectively (Figure 1b).

| Intracellular replication of F. novicida depends on IglE secretion
We focused on IglE as an effector protein and assessed its characteristics because of its unique localization. First, we constructed an iglE deletion (ΔiglE) mutant of F. novicida by homologous recombination.
This mutation decreased the intracellular growth of F. novicida in THP-1 cells, but complementation with wild-type iglE restored the intracellular growth ( Figure 2a). These results indicated that IglE was important for intracellular growth. To examine whether IglE protein was secreted into the culture medium, we cultured F. novicida in medium containing high concentrations of potassium chloride to mimic the host intracellular environment. Although the secretion of PdpC, an FPI protein, was not observed ( Figure S2), we observed IglE secretion in the wild-type and IglE overexpressing strains. However, the amount of IglE secreted from the wild-type bacteria was limited. Importantly, the secretion was not observed in the dotU deletion mutant (ΔdotU, a gene encoding part of the T6SS apparatus), and the secretion decreased in the ΔdotU mutant that overexpressed IglE (Figures 2b and S2a).

| IglE is associated with β-tubulin and MTOCs
To identify IglE-binding proteins, GFP-fused IglE was expressed in 293T cells, and GFP protein was precipitated with GFP-binding protein-conjugated agarose beads. The co-precipitated proteins were separated by SDS-PAGE. A protein of approximately 50 kDa was co-precipitated with GFP-fused IglE (Figure 3a). We identified this 50 kDa protein by matrix-assisted laser-desorption ionization/time-of-flight mass spectrometry as β-tubulin. To confirm the interaction between IglE and β-tubulin, a pull-down assay was conducted. HA-tagged β-tubulin and Myc-tagged IglE were expressed in 293T cells and precipitated with anti-HA antibody using protein G agarose beads. Co-precipitated Myc-tagged IglE was detected by immunoblotting with an anti-Myc antibody.

| IglE disturbs membrane trafficking through MTOCs
In general, phagosomes are transported toward MTOCs on microtubules. Lysosomes are also present around MTOC and fuse with phagosomes (Blocker, Griffiths, Olivo, Hyman, & Severin, 1998) (Figure 4d, white arrow). In cells with large F I G U R E 3 Intracellular growth locus E (IglE) is associated with β-tubulin and microtubule organizing centers (MTOCs). (a) 293T cells were transfected with pAcGFP-C1-IglE and incubated for 48 hr. The cells were disrupted, and IglE-GFP protein was precipitated with the GFP-Trap. Co-precipitated proteins were separated by SDS-PAGE and extracted from the gel. The extracted protein was examined with matrixassisted laser-desorption ionization/time-of-flight mass spectrometry. (b) Binding of IglE and β-tubulin was confirmed with a pull-down assay. 293T cells were transfected with pCMV-HA-Nβ-tubulin and pCMV-Myc-N-IglE. β-tubulin was precipitated by anti-HA antibody-conjugated agarose beads and separated by SDS-PAGE. Co-precipitated IglE-Myc protein was detected by immunoblotting for anti-Myc antibody (IB: Myc). (c) 293T cells were transfected with pAcGFP-C1-IglE and incubated for 48 hr. β-tubulin was stained using Alexa555-conjugated antiβ-tubulin antibody. Scale bar: 20 μm. (d) 293T cells were transfected with pAcGFP-C1-IglE and incubated for 48 hr. MTOCs were stained using anti-pericentrin antibody and Alexa555-conjugated anti-rabbit antibody. To observe the detailed localization of GFP-fused IglE, the sensitivity of detection for GFP was decreased compared to the experiment in Figure 1a. Scale bar: 20 μm. (e) THP-1 cells were infected with Francisella novicida harboring pOM5-GFP at multiplicity of infection = 1 and treated with 50 μg/ml of gentamicin. At 24 hr postinfection, the cells were treated with anti-pericentrin antibody and stained with Alexa555-conjugated anti-rabbit IgG. Scale bar: 40 μm. (f) The number of cells with MTOCs surrounded by F. novicida was calculated for F. novicida-infected cells. *p < 0.01 foci near the nucleus, dextran particles were aggregated or not ingested in 65.6 ± 8.6% of cells (Figure 4e). These results suggest that IglE disturbs membrane trafficking in host cells by interacting with MTOCs, allowing F. novicida to escape from fusion with lysosomes.

| IglE inhibits dynein-based membrane trafficking
Because the minus-end of microtubules is located near the MTOC, IglE may inhibit membrane trafficking toward the minus-end of microtubules. To confirm this hypothesis, we assayed the localization of the motor protein dynein, which moves toward the minus-end of microtubules. In 293T cells-expressing mCherry-fused IglE, dynein was localized to the tips of cells and colocalized with the dot foci of IglE (Figure 5a). Among IglE-expressing cells, 78.3 ± 8.8% of cells contained dynein colocalized with IglE (Figure 5b). In 293T cellscontaining mCherry-fused IglE, IglE also colocalized with pericentrin

| D ISCUSS I ON
The molecular mechanisms underlying the actions of effector proteins from Francisella species are poorly understood. In this study, we revealed the function of the T6SS effector protein IglE, which associated with MTOCs and modulated the membrane trafficking of host cells. To identify how IglE affects the intracellular environment, we analyzed IglE-binding proteins and found that β-tubulin and pericentrin were associated with IglE. Pericentrin is a component of the γ-tubulin ring complex (γ-TuRC). The γ-TuRC is the functional core of the MTOC and acts as a scaffold or a template for α/β-tubulin dimers (Conduit et al., 2015). In THP-1 cells, the ΔiglE mutant of F. novicida was transported to MTOCs, where lysosomes are located. Some bacterial effectors were reported to interact with microtubules or MTOCs and control the intracellular trafficking of bacteria. In Salmonella enterica, some T3SS effectors such as SseF or SseG interact with microtubules to form Salmonella-induced filaments (Müller, Chikkaballi, & Hensel, 2012). In Pseudomonas aeruginosa, the T6SS effector VgrG2b associates with the γ-TuRC, facilitating internalization of the bacterium (Sana et al., 2015). The Chlamydia trachomatis T3SS effector Francisella species are ingested through phagocytosis and grow into the cytosol or in autophagosomes after they escape from phagosomes (Checroun et al., 2006;Chong et al., 2012;Clemens et al., 2004Clemens et al., , 2005Golovliov et al., 2003). For the maturation of phagosomes, endosomes and autophagosomes, motor-based migration on microtubules toward the cell center, where lysosomes are located, is necessary (Blocker et al., 1998;Harrison, Bucci, Vieira, Schroer, & Grinstein, 2003;Kimura, Noda, & Yoshimori, 2008). In THP-1 cells infected with the F. novicida ΔiglE mutant, bacteria accumulated with dynein around MTOCs where the lysosome marker LAMP-1 was located, whereas the wild-type bacteria were located far away from MTOCs. Together, these results imply that IglE associates with MTOCs to inhibit the trafficking of F. novicida-containing phagosomes on microtubules and their subsequent fusion with lysosomes. This may allow F. novicida to escape from phagosomes and grow in the cytosol or in autophagosomes.
With a microscopic observation, IglE seemed to colocalize with MTOCs or dynein. Although IglE co-precipitated with β-tubulin, MTOC proteins such as pericentrin or dynein were not detected by the co-precipitation assay with IglE. This may due to the abundance of β-tubulin in cell cytosol. In addition, IglE expression in 293T cells failed to inhibit the depolymerization or repolymerization in the presence or absence of colchicine, an inhibitor of tubulin polymerization (data not shown). Therefore, the direct target of IglE is still unclear.
However, IglE is expected to associate with β-tubulin through MTOC or dynein because the colocalization of IglE and β-tubulin was not observed with microscopy.
Several reports indicate that IglE is a bacterial lipoprotein with a signal peptide and is located at the bacterial membrane where it forms part of the T6SS apparatus (Bröms, Meyer, & Sjöstedt, 2017;Nguyen, Gilley, Zogaj, Rodriguez, & Klose, 2014;Robertson, Child, Ingle, Celli, & Norgard, 2013). In addition, we observed limited secretion of IglE into the culture medium. Thus, our results suggest F I G U R E 5 Intracellular growth locus E (IglE) colocalizes with dynein and disturbs dynein-based membrane trafficking. (a) 293T cells were transfected with pmCherry-C1-IglE and incubated for 48 hr. Dynein was stained using antidynein antibody and FITC-conjugated anti-mouse antibody. Fluorescent images were merged with differential interference contrast microscopy images. Microtubule organizing center (MTOC) was stained with anti-pericentrin antibody and FITC-conjugated anti-rabbit IgG. Arrowheads indicate colocalization of IglE and MTOC. Scale bar: 40 μm. (b) The number of cells containing dynein or pericentrin colocalized with IglE was calculated for IglE-expressing cells. (c) THP-1 cells were infected with Francisella novicida harboring pOM5-mCherry at multiplicity of infection = 1 and treated with 50 μg/ml of gentamicin. At 24 hr after infection, the cells were treated with anti-dynein antibody and stained with FITC-conjugated anti-mouse IgG. Scale bar: 40 μm. (d) The number of cells containing F. novicida colocalized to the dynein-positive area was calculated for F. novicida-infected cells. *p < 0.01 that IglE may not be the so-called effector protein. In fact, the secretion of other effectors, such as IglC, is inhibited in the ΔiglE mutant (Bröms et al., 2017). Therefore, in the case of the ΔiglE mutant, we could not rule out the possibility that its transportation to MTOCs and the inhibition of its intracellular growth were due to other effectors or a combination of IglE and other effectors. However, under our condition of IglE overexpression in host cells, IglE was associated with MTOCs and disturbed intracellular trafficking. These results at least suggest that IglE may have effector-like functions when bacteria escape into the cytosol and IglE is exposed, or if bacteria are lysed and IglE is released into the cytosol. Because IglE had a signal peptide and was still secreted in the ΔdotU mutant that overexpressed IglE, IglE may be secreted by an unknown Sec protein-related secretion system. Indeed, IglE is detected in the cytosol during Francisella infection (Bröms et al., 2012).
IglE could be a therapeutic target for treating Francisella infections or a biological tool for inhibiting intracellular trafficking because our results suggest that IglE affects MTOCs and modulates intracellular trafficking.

ACK N OWLED G EM ENTS
We thank Dr. Jean Celli for kindly supplying anti-PdpC polyclonal antibody. This study was supported by JSPS KAKENHI Grant Number 15K08463. We acknowledge help with mass spectrometry measurements, which were supported by the general support team at Osaka City University funded by the Grantin-Aid for Scientific Research on Innovative Area "Harmonized Supramolecular Motility Machinery and Its Diversity" (25117501) directed by Makoto Miyata.

CO N FLI C T O F I NTE R E S T
All contributing authors declare no conflicts of interest.