Intracellular fate of Ureaplasma parvum entrapped by host cellular autophagy

Abstract Genital mycoplasmas, including Ureaplasma spp., are among the smallest human pathogenic bacteria and are associated with preterm birth. Electron microscopic observation of U. parvum showed that these prokaryotes have a regular, spherical shape with a mean diameter of 146 nm. U. parvum was internalized into HeLa cells by clathrin‐mediated endocytosis and survived for at least 14 days around the perinuclear region. Intracellular U. parvum reached endosomes in HeLa cells labeled with EEA1, Rab7, and LAMP‐1 within 1 to 3 hr. After 3 hr of infection, U. parvum induced the cytosolic accumulation of galectin‐3 and was subsequently entrapped by the autophagy marker LC3. However, when using atg7 −/− MEF cells, autophagy was inadequate for the complete elimination of U. parvum in HeLa cells. U. parvum also colocalized with the recycling endosome marker Rab11. Furthermore, the exosomes purified from infected HeLa cell culture medium included U. parvum. In these purified exosomes ureaplasma lipoprotein multiple banded antigen, host cellular annexin A2, CD9, and CD63 were detected. This research has successfully shown that Ureaplasma spp. utilize the host cellular membrane compartments possibly to evade the host immune system.


| INTRODUCTION
Phylogenetically, mycoplasmas are probably the latest product of evolution among bacteria resulting from genome reduction (also known as regressive evolution). These organisms have limited metabolic pathways for replication and depend on the host for the supply of exogenous membrane components including fatty acids, cholesterol, and complex lipids for survival (Kornspan & Rottem, 2012). These organisms are bound by a plasma membrane and lack rigid cell walls. Mycoplasmas have mainly been isolated from the mucosal surfaces (Volgmann, Ohlinger, & Panzig, 2005;Waites, Schelonka, Xiao, Grigsby, & Novy, 2009;Yi, Yoon, & Kim, 2005) and some have been reported to be internalized into the host cells (Marques et al., 2010;Winner, Rosengarten, & Citti, 2000;Yavlovich, Katzenell, Tarshis, Higazi, & Rottem, 2004). Ureaplasma spp. belong to the family Mycoplasmataceae. These are the smallest self-replicating organisms in terms of genome size and cellular dimensions. Ureaplasma parvum (U. parvum) and U. urealyticum, which are common inhabitants of the human lower genital tract, have been isolated from 40% to 80% of women among child-bearing age (Taylor-Robinson & McCormack, 1980). However, when Ureaplasma spp. spread in the genital tract during gestation, they have the potential to be pathogens and cause chorioamnionitis, resulting in spontaneous abortion or preterm birth (Namba et al., 2010).
The mechanisms by which viruses and bacteria are internalized into host cells are known mainly to involve two pathways, fusion and endocytosis. The endocytotic pathways exploited by animal viruses to gain entry into host cells include macropinocytosis, clathrin-dependent endocytosis, and caveolae-dependent endocytosis. Rab proteins are involved in various aspects of endocytic and exocytic protein transport through their specific association with membrane vesicles or organelles. Early endosome antigen 1 (EEA1) is a major marker of the early endosome stage (Christoforidis, McBride, Burgoyne, & Zerial, 1999;Simonsen et al., 1998). After this stage, the phagosome loses the marker associated with early endosome, Rab5, and acquires Rab7 and markers associated with the late endosome stage such as a transmembrane protein enriched in late endosomes and lysosome-associated membrane protein 1 (LAMP-1) (Desjardins, 1995;Desjardins, Huber, Parton, & Griffiths, 1994;Pitt, Mayorga, Schwartz, & Stahl, 1992). The eukaryotic cytoskeleton is targeted by a variety of bacterial pathogens during the course of infection and dynamic changes of the cytoskeleton influence the interaction of microbial pathogens with the host cells. Microbial pathogens deliver a number of effector proteins to the host cells to rearrange the cytoskeleton in a way that promotes infection. Many bacterial pathogens modulate microtubule dynamics by employing virulence proteins to promote infection (Radhakrishnan & Splitter, 2012). Intact microtubules are also essential for other polyomaviruses, including SV40 (Pelkmans, Kartenbeck, & Helenius, 2001).
Galectins are beta-galactoside-binding lectins that accumulate in the cytosol before being secreted via a leader peptide-independent pathway (Houzelstein et al., 2004;Rabinovich & Toscano, 2009). During an infection, galectin-3 was suggested to be a potential receptor for pathogen recognition based on its ability to bind certain bacterial, parasitic, and fungal products (Sato & Nieminen, 2004). Galectin-3 was also proposed to be a potential immunological danger signal based on its passive release from cells at the site of infection and its active release from inflammatory macrophages (Liu & Hsu, 2007;McClung et al., 2007;Sato & Nieminen, 2004). The accumulation of galectin-3 in host cells is known to induce autophagy (Chen, Weng, Hong, & Liu, 2014).
Atg7 mutant mice should be useful for examining the role of autophagy in the cell death pathway or in a cellular defense mechanism in the pathogenesis of these diseases (Komatsu et al., 2005). LC3 is the first mammalian protein localized in the autophagosome membrane.
However, the autophagic machinery may play a role in establishing resident bacteria. These processes are collectively called xenophagy and the molecular mechanisms that govern these processes are only now beginning to be analyzed (Deretic & Levine, 2009;Huang & Brumell, 2009;Levine, 2005;.
It was reported that the genital mycoplasmas Ureaplasma spp.
and Mycoplasma hominis were detected in the cord blood of 23% of preterm birth babies (Goldenberg et al., 2008). This report also indicated that fetal U. parvum infection is caused not only by ascending infection from the lower genital tract but also by hematogenous vertical transmission; nevertheless, the mechanism underlying feto-maternal transmission is still unknown. In this report, we reveal the mechanisms of internalization and intracellular survival of U. parvum in HeLa cells and present the possibility of host cellular exosome-mediated transmission of bacteria. Such exosome-mediated transmission may facilitate escape from the human immune system and may contribute to the feto-maternal transmission of U. parvum.

| The clinical specimen and U. parvum strains
All clinical specimens were obtained after receiving informed consent and approval from the Ethics Committee of Osaka Medical Center and Research Institute for Maternal and Child Health. For the pathological examinations, placenta from preterm delivery at 29 weeks of gestation (Namba et al., 2010) was used. This placenta was culturepositive for U. parvum, which was determined by analysis of the DNA sequence of the 16S rRNA gene (data not shown). The sequence primers used were as follows: (27f: 5′-AGAGTTTGATCCTGGCTCAG-3′, 1525r: 5′-AAAGGAGGTGATCCAGCC-3′). For in vitro infection studies, we used clinical isolates of the U. parvum serovar 3 strain derived from human placenta of a preterm delivery at 26 weeks of gestation [U. parvum OMC-P162 (Uchida et al., 2013)].

| Immunohistochemistry
Paraffin-embedded sections of human placenta with Ureaplasma spp.
Immunoreactivity with horseradish peroxidase-conjugated labeled polymer was detected using the Envision TM + Dual Link System-HRP (Dako, Carpinteria, CA, USA). Mayer's hematoxylin staining (Muto Pure Chemicals, Tokyo, Japan) was also performed for light microscopy (BX51; Olympus, Tokyo, Japan) evaluation. The study was approved by the Institutional Human Ethical Committee and recommended guidelines were followed during sample collection.  (Hanaichi et al., 1986), and observed by transmission electron microscopy (TEM) (HT-7700;

| Scanning electron microscopy
Ureaplasma parvum was grown on a UMCH agar plate (Namba et al., 2010) and fixed with 2% glutaraldehyde (Wako) overnight before being cut into small pieces. The samples were then washed, resuspended in PBS, postfixed with 1% osmium tetroxide, and dehydrated with a graded ethanol series. We then conducted a t-butanol drying process and coated the samples with platinum/palladium by ion sputter E-1030 (Hitachi). Microscopy was performed with an SU3500 SEM (Hitachi).
Cotransfection with the expression clones, the pEXPR series described in the above section, and the ϕC31 integrase expression clone, pJTI TM ϕC31 Int (Invitrogen), was performed at a mass ratio of 1:1. After culturing in a 6well plate for 24 hr, the approximate numbers of transiently transformed cells were determined by fluorescence microscopy and the cells were split into 10-cm plates and cultured. After 48 hr of transfection, the cells were selected in medium containing 2-4 μg/ml of blasticidin S HCl (Sigma).
Selection continued for 10-14 days, when the colonies became visible.
The individual colonies were picked using a pipette tip and transferred to individual wells of a 24-well plate. Surviving colonies were expanded for stock. EGFP-galectin-3 transient transfection was performed with FuGENE ® HD in accordance with the manufacturer's protocols.

| Labeling of U. parvum cells
The U. parvum cells were cultured in 2 ml of UMCHs medium (Namba et al., 2010)

| U. parvum infection of cultured cells
The HeLa, atg7 −/− and WT MEF cells, and several stable transformant cells were grown on poly-L-lysine-coated glass coverslips (13 mm; Matsunami Glass Ind. Ltd., Osaka, Japan) to approximately 70% confluence (5 × 10 4 / ml) before they were infected with U. parvum. These cells were initially washed with PBS and then infected with DiI-labeled U. parvum contained in 1 ml of DMEM with 2% FBS. The sets of inoculated cells were incubated at 37°C in a 5% CO 2 atmosphere for 0, 0.5, 3, 6, or 24 hr.

| Immunofluorescence and microscopy
After each period of infection, the bacterial suspension was gently re- and a HeNe laser with a wavelength of 543 nm was used.

| Inhibition of U. parvum entry into HeLa cells by selective inhibitors
HeLa cells were plated on coverslips and incubated with 10 mM chlorpromazine (CPZ) dissolved in distilled water to make 1 mg/ml, 0.

| Small interfering RNA (siRNA) transfection
The siRNA duplex was synthesized as a 21-mer with UU overhangs (Ambion by Life Technologies Co.). The clathrin heavy chain target sequence was GGUUGCUCUUGUUACGGAU. Negative control siRNA sequences that did not target any gene product (Ambion) were also used. The siRNA duplex was resuspended in 50 μM nuclease-free water before transfection. HeLa cells were seeded on coverslips with 4 μM of siRNA duplex (clathrin and negative control) and 5 μl of siPORT NeoFX (Ambion) in 300 μl of Opti-MEM medium. Three days later, the cells were infected. Both infected and uninfected cell lysates were harvested for western blotting or immunofluorescence.
Lysates were clarified by pipetting and rotated for 30 min at 4°C. antibodies, the blots were visualized using an enhanced chemiluminescence detection system (GE). The blots were also probed for GAPDH, which was used as a loading control.

| Gentamicin invasion assay
The gentamicin invasion assay was performed to determine the rate of invasion of viable U. parvum into eukaryotic cells

| Isolation of exosomes
Exosomes were extracted from peripheral plasma using miRCURY TM Exosome Isolation Kit (EXQON A/S, Vedbaek, Denmark), in accordance with the manufacturer's instructions. The precipitat from miR-CURY TM exosome pellets was lysed in 100 μl of resuspension buffer for western blot analysis.

| Statistical analysis
To quantify the intracellular survival of U. parvum DNA and the suppression of U. parvum internalization using inhibitors, data were analyzed using one-way analysis of variance with a Tukey-Kramer post hoc test. Student's t test was used to examine the number of infected/uninfected viable cells, colocalized signaling data, and the gentamicin invasion assay results. Differences were considered significant when p < .05. Error bars represent mean ± standard error of the mean (SEM).

| Intracellular U. parvum is detected in fetal cells from chorioamnionitis placenta and mammalian cells
To

| U. parvum invasion of mammalian cells requires clathrin-dependent endocytosis
To provide further information on U. parvum entry, the colocalization of U. parvum with endocytosis-related molecules, clathrin heavy chain and caveolin-1, was observed. After 30 min of infection, colocalization of U. parvum and clathrin was observed as a spotted pattern in the perinuclear region (Figure 2a). In contrast, the caveolin-1 protein did not colocalized with U. parvum (Figure 2a). We observed the colocalization of clathrin and U. parvum 5 min after infection (Figure 2b). The colocalization of clathrin and U. parvum showed rapid changes. It was detected at a very early stage of at least 5 min, peaked at 30 min, and then decreased after 1 hr of infection. These results demonstrated that U. parvum was internalized into the host cell through a clathrin-dependent pathway.
We next investigated the role of clathrin in U. parvum infection.
CPZ is a commonly used inhibitor of clathrin-coated pit formation by the reverse translocation of clathrin and its accessory proteins from the plasma membrane to intracellular vesicles. PAO is another inhibitor of clathrin-dependent endocytosis. In a control experiment, as shown in

| U. parvum is trafficked along microtubules and enters endosomes
We prepared a stable transformant of HeLa cells expressing tau protein-fused EYFP to follow the direct involvement of   for 3 hr, galectin-3 signals colocalized with LAMP-1 (Figure 5c,d).

| Intracellular interactions of galectins with U. parvum
Therefore, U. parvum-induced galectin-3-positive endosomal damage occurred at the late endosome and the lysosome.

| Involvement of U. parvum-induced autophagosomes
LC3 is a canonical autophagosome marker. To evaluate U. parvuminduced autophagosomes, we used mouse embryo fibroblasts (MEFs) These findings also suggested that colocalization of part of the incorporated U. parvum and the lysosome had occurred and that the intracellular U. parvum degradation process was mediated by autophagy.
We also investigated whether U. parvum induced lipidation of LC3, which is essential for the translocation of LC3 from the cytosol to autophagosomes. The lipidated form of LC3 (LC3-II) has increased mobility on SDS-PAGE relative to its unlipidated form.  Rab11 *** *** 1' 12" 54" 38" 20" 2" shows a western blot using an antibody against LC3. The ratio between LC3-II and β-tubulin was increased in WT MEF after U. parvum infection for 3 hr (Figure 6e). These results suggested that LC3 is an important component for in U. parvum degradation within the host cell. The results from the gentamicin invasion assay showed that U.
parvum was increased more in atg7 −/− MEF cells compared with MEF cells (Figure 6f). These results suggested that the autophagic machinery is involved in U. parvum degradation.
parvum into HeLa cells involves early to late endosomes and recycling endosomes. As expected, the intracellular transport of U. parvum toward the perinuclear region is associated with the microtubules.
Several previous studies identified a correlation between galectin-3 and the endocytic pathway via the former's interaction with two lysosomal/late endosomal proteins, LAMP-1 and LAMP-2 (Dong & Hughes, 1997;Sarafian et al., 1998). LC3-positive autophagosomes colocalized with galectin-3, ubiquitin, and p62/SQSTM1 . In this study, we showed galectin-3 accumulation proximal to the invading bacteria. Moreover, the partial targeting of galectin-3 to the LAMP-1-positive endosomes may explain the minor overlap of U. parvum-infected cells. Galectin-8 can target the vacuole containing damaged bacteria for autophagosome degradation; therefore, it is considered a danger receptor that restricts intracellular bacterial proliferation (Thurston, Wandel, von Muhlinen, Foeglein, & Randow, 2012). Furthermore, we found that galectin-8 accumulated in cells infected with U. parvum. These observations indicated the possibility that U. parvum is a vesicle-damaging pathogen. However, future studies will be needed to clarify this issue.
Although some bacteria are killed by autophagy, others can evade or even exploit autophagy to cause diseases. For example, Shigella flexneri can evade autophagic capture in the cytosol (Ogawa et al., 2005).
This bacterium enters host cells and escapes from the phagosome into the cytosol (Cemma & Brumell, 2012). Autophagy is a dynamic process consisting of the formation and fusion of membrane compartments.
Here, we report that the induction of autophagy occurred in the initial stage after infection (3 hr) with U. parvum. To understand how autophagy plays a role during U. parvum infection, we analyzed the localization of autophagosome markers. We observed the locations of LC3 and U. parvum by fluorescence microscopy, which demonstrated that they colocalized with each other (Figure 8). The degradation of U.
parvum in the phagosome is, at least in part, mediated by autophagy in MEF cells.
It is known that EGFP-tagged galectin-3 is recruited to the bacterial entry site within seconds after vacuolar rupture and targets the disassembling membranes surrounding the bacterium (Paz et al., 2010).
Furthermore, the role of galectin-8 in targeting bacterially damaged vesicles for autophagy has been convincingly described. Therefore, galectins may function in a manner similar to antimicrobial peptides and play an important role in innate immunity through this mechanism.
Although a role of exosomes in the shuttling of infectious agents between cells has been postulated, this has still not been extensively demonstrated. We showed that U. parvum infection can be transmitted by exosomes between HeLa cells and can establish a productive infection ( Figure 7). In general, the mechanism of uptake of exosomes by cells is not fully understood. Further research is required to determine F I G U R E 7 Ureaplasma parvum localized to the recycling endosome and exocytosis with Annexin A2. (a) Schematic representation of infection experiments. (b) HeLa cells stably expressing EGFP-annexin A2 were infected with DiI-labeled U. parvum for 24 hr. Exosome the pathway involved in the uptake of exosomes and U. parvum. Our data indicated that urease-producing Ureaplasma may survive in low-pH environments such as in the lysosome. Although the precise escape mechanism remains unknown, U. parvum appeared to induce significant endosomal-lysosomal damage. Cytoplasmic U. parvum was recognized by galectin-3, a component of the innate immune system; this may form the basis for its interaction with autophagosomes. The precise molecular basis behind the mechanism of ureaplasmal host cell membrane damage should be elucidated in the future.
It is widely accepted that bacteria ascending from the vagina after initial colonization are the main cause of preterm birth (Goldenberg, Hauth, & Andrews, 2000). Ureaplasma spp. is one of the pathogenic organisms most commonly detected in the amniotic fluid (Kacerovsky et al., 2013;Ueno et al., 2015), but its mechanism of tissue invasion remains unknown. Thus, in this study, we focused on clarifying the mechanisms underlying the cellular invasiveness of Ureaplasma spp. and found that it has the potential to survive intracellularly by escaping lysosome degradation and autophagic elimination. We further revealed the in vivo intracellular localization of Ureaplasma spp. in fetal-derived cells of infected placenta. We thus clarified the mechanisms of Ureaplasma spp. invasion; however, because Ureaplasma spp.
are commonly found in the vagina of many women in reproductive age, the host factor responsible for preterm birth remains to be determined.