Co-culture models illustrate the digestion of Gemmata spp. by phagocytes

Gemmata spp. bacteria thrive in the same aquatic environments as free-living amoebae. DNA-based detection of Gemmata spp. sequences in the microbiota of the human digestive tract and blood further questioned the susceptibility of Gemmata spp. to phagocytes. Here, Gemmata obscuriglobus and Gemmata massiliana were co-cultured with the amoebae Acanthamoeba polyphaga, Acanthamoeba castellanii, Acanthamoeba griffini and THP-1 macrophage-like phagocytes. All experiments were performed in five independant replicates. The ratio amoeba/bacteria was 1:20 and the ratio THP-1/bacteria was 1:10. After a 2-hour co-culture, extracellular bacteria were killed by kanamycin or amikacin and eliminated. The intracellular location of Gemmata bacteria was specified by confocal microscopy. Microscopic enumerations and culture-based enumerations of colony-forming units were performed at T = 0, 1, 2, 3, 4, 8, 16, 24, 48 and 72 hours post-infection. Then, Gemmata bacteria were engulfed into the phagocytes’ cytoplasmic vacuoles, more than (98 ± 2)% of Gemmata bacteria, compared to controls, were destroyed by phagocytic cells after a 48-h co-culture according to microscopy and culture results, and no positive culture was observed at T = 72-hours. Under our co-culture conditions, Gemmata bacteria were therefore susceptible to the environmental and host phagocytes here investigated. These data suggest that these Acanthamoeba species and THP-1 cells cannot be used to isolate G. massiliana and G. obscuriglobus under the co-culture conditions applied in this study. Although the THP-1 response can point towards potential responses that might occur in vivo, these responses should first bevalidated by in vivo studies to draw definite conclusions.

of Legionella 4 and waterborne mycobacteria including Mycobacterium massiliense 12 . In addition, amoebae may be used as model organisms to investigate the relationship between bacteria and phagocytes, including macrophages 29 . Some studies reported the relationship between planctomycetes and aquatic microbial communities such as sponges 23,30 and macroalgae [31][32][33] , but the interactions between amoebae and Gemmata bacteria have not been specifically explored.
Here, we investigated the relationship between phagocytic amoebae and THP-1 cells with the two cultured Gemmata species. After preliminary experiments ensuring the viability of both Gemmata species in the medium used to culture amoebae and in the medium used to culture macrophages, both Gemmata species were exposed to amoebae and macrophages and the intracellular uptake and survival of bacteria were observed by optic and confocal microscopy, and colony-forming unit measurement.
Compared to controls not submitted to thermal shock, the amoebae did not lead phagocyted Gemmata bacteria out of their cytosol, since no growth was observed. More than (98 ± 2%) of the Gemmata spp. organisms were internalized and destroyed by amoeba after a 48-h co-culture according to microscopic results, and no positive culture was observed at T = 72 -hours. Overall, no replication occured under the experimental conditions here reported. Compared to non-inoculated amoebae controls, almost all amoebae possessed many vacuoles without bacteria at T = 72 -hours. Gemmata spp. bacteria became indetectable by microscopic examination and by subculturing on agar plates at day 3. No discrepancy was observed between microscopy and culture results.  In a first step, we observed that Gemmata spp. bacteria survived for 72 hours in RPMI. Then co-culture experiments and microscopic examination of THP-1 cells after trypan blue staining indicated no significant difference in the viability of THP-1 cells over time, whether they be Gemmata spp.

Discussion
Amoebae and THP-1 cells are phagocytic cells which were previously used to probe microorganism-phagocyte interactions [34][35][36][37][38][39] . In the present study, we investigated such interactions between only two culturable representatives of the bacterial genus Gemmata, i.e. G. obscuriglobus and G. massiliana. In these series of experiments, culture of the washing product and controls not submitted to thermal shock action of amoebae and THP-1  cells inoculated with G. obscuriglobus or G. masiliana remained sterile, indicating that no extracellular bacteria remained alive or escaped from the cytosol of phagocytic cells, and that all further observations in culture of lysate obtained after thermal shock derived from intracellular bacteria. Also, the data presented in this study were interpreted as authentic because negative controls remained negative in each experimental step. At last, the observations were reproduced five times.
In every co-culture experiment, amoebae and THP-1 cells were found to be bactericidal against the two investigated Gemmata spp. The fact that results obtained with amoebae mirror results obtained with THP-1 cells is not surprising, as demonstrated previously with the similar actions of these phagocytic cells in numerous bacteria such as Legionnela, many mycobacteria and Chlamydia, which can resist both amoebae and macrophages 4,6,13 .
Results obtained with the amoebae here investigated suggest that these unicellular eukaryotes are unlikely hosting Gemmata species in the aquatic environments where they could thrive together. Gemmata are mainly found in soil and aquatic environments along with FLA but have also been found in the human gut microbiota, in the blood of immune-compromised aplastic leukemic patients 17,18,24,26 and hospital water networks in close proximity to patients 18  This high susceptibility of Gemmata spp. to bactericidal phagocytosis contrasted with the fact that Gemmata spp. are multidrug-resistant bacteria 28 and have a high ability to adapt to harsh environments. Moreover, these bacteria may thrive in the human body 24,27 . However, under the co-culture conditions of our experiments, Gemmata bacteria did not resist to amoebae and THP-1 bactericidal action despite their panoply of attack and defense mechanisms 40 . The rapid and significant phagocytosis of Gemmata bacteria could be explained by the presence of holdfasts of glycoproteic nature on their outer membrane 16,18 which could facilitate adhesion to phagocytic cell receptors and increase cell-to-cell contact and rapid internalization. Consequently, Gemmata spp. life cycle requires an attached state. Their proliferation starts when they attach through their holdfast to a bracket. Then, flagellated budding new cells are formed which move freely in the water until developing their holdfast and returning to the attached form 41 . FLA and THP-1 cells do not offer these survival and proliferation conditions compared to macroalgae, marine sponges and crustaceans, which are closely related to planctomycetes in aquatic environments and biofilms 31 .
In conclusion, the data here reported show that coculturing the two Gemmata species under investigation with Acanthamoeba amoebae results in Gemmata death. Acanthamoeba ameobae are unlikely hosting Gemmata spp. in the environment. Moreover, these amoebae could not be used for the tentative isolation of Gemmata bacteria in the laboratory under the co-culture conditions tested in this study. Moreover, the interactions between THP-1 cells and Gemmata spp. could provide insight into the action of the monocytes-macrophages against Gemmata spp. cells during the colonization and infection. Though unlikely, patients with compromised macrophage activities, such as aplastic patients, could be infected by opportunistic Gemmata spp. Although the THP-1 response can hint to potential responses that might occur in vivo, these reponses should first be validated by in vivo studies before drawing more definite conclusions.

Materials and Methods
Preliminary experiments. The methodological procedures here presented have been designed and standardized after a series of preliminary studies. Preliminary experiments were used to define the phagocyte/bacteria ratio; the duration of co-culture of 6 h, 4 h, 3 h and 2 h before T0; antibiotic treatment conditions (concentration and duration of treatment) to kill rapidly extracellular bacteria, and the effectiveness of the wash in removing extra-cellular bacteria (later controlled by microscopy and culture). Also, different types of staining (Giemsa, Gimenez, Hemacolor, acridin orange) were tested to choose the most appropriate staining, and the thermal shock procedure was evaluated and standardized before the experiments. Then, the protocol was successively applied with each amoeba species and to THP-1 cells. Bacterial strains and culture conditions. G. obscuriglobus DSM 5831 T and G. massiliana DSM 26013 T (CSUR P189 T ) were obtained from the Collection de Souches de l'Unité des Rickettsies, Marseille, France, and the German Collection of Microorganisms and Cell Cultures (Braunschweig, Germany). Bacteria of both species were sub-cultured on Caulobacter medium DSMZ 595 supplemented by 5% Escherichia coli filtrate or Staley's maintenance medium DSMZ 629 prepared as described on the website (http: //www.dsmz.de). Bacteria were grown on these solid media incubated aerobically at 30 °C for 7 to 14 days. Identification of colonies was ensured by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS) analysis as previously described 42 . Prior to co-culture, colonies were harvested in a 15-mL tube containing 5 mL of sterile phosphate-buffered saline (PBS), the tube was rigorously vortexed and the suspension was passed three times through a 29-Gauge needle in order to separate aggregates. The inoculum was adjusted at 2 × 10 7 cells colony-forming units (CFU)/mL after calibration by Kovas slide 10 (Hycor Biomedical, Indianapolis, IN, USA) for co-culture. In parallel, this inoculum was maintained in RPMI 1640 and PAS for three days to check Gemmata spp viability after one to three days and compared with Staley's liquid medium as reference. Culture-based microbial enumeration on Staley's solid agar has been performed in order to assess bacterial survival in these co-culture liquid media.
Amoebae and culture conditions. Acanthamoeba polyphaga (strain Linc AP1), Acanthamoeba castellanii (strain ATCC 30234) and Acanthamoeba griffin (strain ATCC 50702) were cultured independently in axenically Peptone Yeast-extract Glucose (PYG) medium placed in 75-cm 2 culture flasks. The flasks were then incubated for 48 hours at 30 °C for A. polyphaga and 28 °C for A. griffini and A. castellanii. Trophozoites were suspended by tapping the flasks, centrifuged in 50-mL tubes and the pellet was suspended in Page's amoebal saline (PAS). Then, the amoebal cells were adjusted at 10 5 cells/mL using Kovas slide 10 and we checked their viability using trypan blue staining before co-cuture. Amoeba viability and growth were assessed at the end of co-culture (day 3) using trypan blue staining.
Human monocytic THP-1 cell lines and culture conditions. THP-1 cells were retreived from the blood of a patient with acute monocytic leukemia. Since their establishment in 1980 43 , THP-1 cells have become one of most widely used cell lines to investigate the function and regulation of monocytes and macrophages in the cardiovascular system. After exposure to phorbol-12-myristate-13-acetate (PMA, also known as TPA, 12-O-tetradecanoylphorbol-13-acetate), nearly all the THP-1 cells started to transform into macrophages 44 . In this study, unstimulated THP-1 cells were preferred over macrophages (present in the tissues) in order to mimic the behavior of peripheral blood monocytes cells. Indeed, Gemmata DNA has been previously detected in the blood collected from two aplastic leukemic febrile patients but not in tissues. THP-1 cells were kindly provided by Dr. E. GHIGO, IHU Méditerranée Infection, Marseille, France. Cells were grown in RPMI 1640 (Gibco ™ , Eggenstein, Germany) tissue culture medium supplemented with 2% glutamine and 10% heat-inactivated fetal bovine serum in 75-cm 2 tissue culture flasks, incubated at 37 °C in 5% CO 2 . The culture medium was refreshed every 3 days. Prior to co-culture, cells were harvested and washed thoroughly twice with PBS and adjusted at 2 × 10 6 cells/mL using Kovas slide 10. We checked their viability using trypan blue staining before co-cuture. The viability and growth were assessed at the end of co-culture (day 3).
Amoeba-Gemmata species co-culture conditions. All experiments were performed in five independant replicates during 72 hours in 12-well tissue culture plates (Becton Dickinson, Le Pont-de-Claix, France) and repeated five times for ensuring reproducibility. Each species of amoeba was co-cultured with each Gemmata species. Each plate contained six wells with amoebae co-cultured with each Gemmata species and six wells with amoebae as negative control wells (bacteria-free). More precisely, 1.8 mL of the amoebae-containing suspension was pipetted in each well of a 12 well-plate, 200 μL of Gemmata suspension at 2 × 10 7 CFU/mL in PBS (ratio amoebae/bacteria was 1:20) was added in challenged wells and 200 μL of PBS in control wells. Plates were incubated at 30 °C for two hours. After a 2 -hour co-culture, in order to eliminate extracellular bacteria, the supernatant was removed, the amoeba monolayer was washed twice and two milliliters of modified PAS containing 150 g/L kanamycin was added in each well, including negative control wells. After a 30-min incubation period with the antibiotic, the amoeba monolayer was rinsed twice with PBS to eliminate the extracellular bacteria, this operation was repeated once with a 30-min incubation period and then rinsed twice to obtain a third washing. Antibiotic treatment protocol in 2 × 30 min was chosen to combine inactivation and physical removal of extracellular bacteria to have low extracellular bacteria in the last rinse product. Antibiotic inactivation tests of Gemmata spp were performed in PAS in the presence or absence of phagocytic cells for the interpretation of culture results. Finally, after a series of antibiotic rinsing, the amoeba monolayer was covered with two mL of PAS in each well and the plates were incubated at 30 °C for A. polyphaga and 28 °C for A. griffini and A.castellanii for the rest of the experiment. Negative controls (uninfected) of each amoeba were cultured separately in PAS medium as described above.
THP-1 cells-Gemmata spp.co-culture conditions. Co-culture was performed in RPMI 1640 medium without foetal bovine serum and glutamine in order to deplete the stored nutrients and slow down the growth of THP-1 cells. The experiments were performed in five independent replicates during 72 hours in 12-well tissue culture plates, each containing independently Gemmata species, as previously described for amoebae. Each plate contained six wells with THP-1 cells co-cultured with each Gemmata species and six wells with THP-1 cells as negative control wells (bacteria-free). More specifically, 1.8 mL of 2.10 6 cells/mL THP-1 cells (calibrated as described above) suspension was pipetted in each well of a 12-well plate, 200 μL of Gemmata suspension at 2 × 10 7 CFU/mL (ratio THP-1 cells/bacteria was 1:10) was added in challenged wells and 200 μL of PBS in control wells. Plates were incubated at 37 °C with 5% CO 2 for two hours. After a 2 -hour co-culture, in order to eliminate ScIEntIFIc REPORTS | (2018) 8:13311 | DOI:10.1038/s41598-018-31667-0 extracellular bacteria, the supernatant was removed, THP-1 cells were washed with PBS and two mL of RPMI containing 150 g/L of amikacin was added in all wells, including negative control wells for 30 min. After a 30-min incubation period at 37 °C under 5% CO 2 with the antibiotic, the cells were rinsed to eliminate the extracellular bacteria, then antibiotic treatment was repeated once with a 30-min incubation period and then rinsed twice to obtain the last rinse called washing product. 100 μL of this washing product was cultured in Staley's medium to ensure the absence of viable extracellular bacteria. Microscopic controls (using Kova slide) at the fresh state of the washing product were performed, as described above, before being plated on solid agar. Finally, THP-1cells were covered with 2-mL of RPMI in each well and the preparations were incubated at 37 °C in the presence of 5% CO 2 for 72 hours . Uninfected THP-1 cells were cultured in the same conditions as negative controls.
Microscopy and culture conditions of lysate obtained after thermal shock. After a 2 -hour co-culture duration, antibiotic treatment and a series of rinsing, microscopic enumerations and culture-based enumerations of colony-forming units were performed at T = 0 (time beginning after the last rinse of antibiotic), 1, 2, 3, 4, (close kinetic counting to see more intracellular bacteria phagocyted at the beginning) and T 8, 16, 24, 48, 72 -hour post-inoculation (procedure standardized after preliminary sudies).
Microscopy. The phagocyte cells and the supernatant were removed from each well, 200-µL volume was used to prepare a Cytospin (smears) centrifuged at 44 g during 5 mins for microscopy analysis. Smears were prefixed with 90° ethanol. Amoeba-Gemmata spp coculture smears were examined by fluorescence microscopy after acridine orange staining as described on the website http://www.memobio.fr/html/bact/ba_te_acr.html. In order to precise the intracellular location of bacteria, smears were observed under a Zeiss LSM 800 confocal microscope using a 488 nm excitation laser (Carl Zeiss S.A.S., Marly-le-Roi, France). The 63X/1.4NA oil immersion objective was used for image acquisition. 3D reconstruction was performed using Zen software (Zeiss) from a z-stack acquisition of 10 images with a z-spacing of 1.2 µm. Images were post-processed using ImageJ software by adjusting contrast and brightness. For THP-1-Gemmata co-culture, slides were stained by the Gimenez method and obseved by light microscopy. Microscopic results were expressed by counting the number of intracellular bacteria per 100-amoeba or THP-1 counted. Means and standard errors of five independent experiments have been calculated using Excel 2013 software. The phagocytic cells and the supernatant were removed from each well, a 200-µL volume was used to prepare a Cytospin (smears) centrifuged at 44 g during 5 mins for microscopy analysis. Smears were prefixed with 90° Ethanol. Amoeba-Gemmata spp co-culture smears were examined by fluorescence microscopy after acridine orange staining as described on the website http://www.memobio.fr/ html/bact/ba_te_acr.html. In order to observe intrabacteria localization in vacuoles or in the cytoplasm, smears were observed under a Zeiss LSM 800 confocal microscope using a 488 nm excitation laser (Carl Zeiss S.A.S., Marly-le-Roi, France). The 63X/1.4NA oil immersion objective was used for image acquisition. 3D reconstruction was performed using Zen software (Zeiss) from a z-stack acquisition of 10 images with a z-spacing of 1.2 µm. Images were post-processed using ImageJ software by adjusting contrast and brightness. For THP-1-Gemmata co-culture, slides were stained by the Gimenez method and obseved by light microscopy. Microscopic results were expressed by counting the number of intracellular bacteria and extracellular bacteria (remaining bacteria not removed by last rinsing) per 100-amoebae or THP-1 counted. Means and standard errors of five independant replicates have been calculated using Excel 2013 software.
Culture conditions of lysate obtained after thermal shock. To determine the viability of intra-phagocyte bacteria, 200 μL of the phagocytic monolayer and the supernatant were removed from the bottom of wells and transferred in 1.5-mL polypropylene tubes. The preparation was submitted to thermal shock. It was frozen in liquid nitrogen at −196 °C, submitted to direct defrosting in a water bath at 40 °C for two minutes and then vortexed for 30 seconds. This operation was repeated once to ensure more lysis of host cells (Phagocytic lysis was checked using Trypan blue staining). A 100-μL volume of the lysate and 100 μL of a one-fold dilution in PBS were cultured for two weeks in Staley's medium and incubated at 30 °C to quantify CFU. A 100-μL volume of culture product not subjected to the thermal shock action was cultured in parallel for control. Colonies were counted after 2 weeks of growth at 30 °C. Futhermore, in order to assess the impact of thermal shock on the viability of Gemmata spp. bacteria, the Gemmata inoculum at 2.10 7 bacteria/mL was subjected to the thermal shock procedure and then plated on staley's agar medium to check its viability. Data are expressed as CFU per milliliter. Means and standard errors have been calculated using Excel 2013 software.