Microbes and Infection 16 (2014) 746e754 www.elsevier.com/locate/micinf
Original article
Antibiotic susceptibility of Estrella lausannensis, a potential emerging pathogen Marie de Barsy, Lavinia Bottinelli, Gilbert Greub* Institute of Microbiology, University Hospital Center and University of Lausanne, Lausanne, Switzerland Received 27 June 2014; accepted 5 August 2014 Available online 20 August 2014
Abstract Estrella lausannensis is a new Chlamydia-related bacterium, belonging to the Criblamydiaceae family. As suggested by its species name, this bacterium harbors a peculiar star shape. E. lausannensis is able to infect a wide range of amoebal, fish and mammalian cell lines. Moreover, seroprevalence of 2.9% was reported in children and in women with tubal pathology, showing that humans are commonly exposed to this recently discovered strict intracellular bacteria considered as a potential pathogen. Antibiotic susceptibility was determined using two approaches: qPCR and cellular mortality assay. Antibiotics classically used against intracellular bacteria were tested, including b-lactams, fluoroquinolones, cyclines and macrolides. We showed that E. lausannensis is resistant to b-lactams and fluoroquinolones, and sensitive to cyclines. Interestingly, E. lausannensis is slightly resistant to azithromycin with a MIC of 2 mg/ml, which is 10 fold higher compared to Waddlia chondrophila and Parachlamydia acanthamoebae MIC's. A single A2059C mutation in 23S rRNA gene could be responsible for this unexpected resistance. © 2014 Institut Pasteur. Published by Elsevier Masson SAS. All rights reserved.
Keywords: Antibiotic susceptibility; Cell culture; Chlamydiae; Chlamydia-like organism
1. Introduction Estrella lausannensis is a newly discovered Chlamydiarelated bacterium, isolated from a water sample taken upstream of a water treatment plant fed by the river Llobregat in Spain [1]. E. lausannensis belongs to the Criblamydiaceae family in the Chlamydiales order [1,2]. Like other members of the Chlamydiales order [3,4], E. lausannensis harbors a biphasic developmental cycle, with infectious elementary bodies (EBs), which infect cells and differentiate into reticulate bodies (RBs), the proliferative developmental forms [1]. * Corresponding author. Center for Research on Intracellular Bacteria (CRIB), Institute of Microbiology, University Hospital Center and University of Lausanne, Bugnon 48, 1011 Lausanne, Switzerland. Tel.: þ41 21 314 49 79; fax: þ41 21 314 40 60. E-mail address: [email protected] (G. Greub).
Then, RBs redifferentiate into EBs, which are released after cell lysis [1]. A new infection cycle can then be initiated. E. lausannensis EBs exhibit a star shape, similar to that observed for Criblamydia sequanensis [1,2]. Although this star shape corresponds to a fixative artefact, this highlights differences in structure and composition of the bacterial cell wall [5]. E. lausannensis is able to infect a wide range of amoebal hosts [1] as well as fish cell lines [6] and macrophages (Rusconi and Greub, unpublished). The ability to resist to human macrophage microbicidal effectors suggests a potential pathogenic role of E. lausannensis. Moreover, different recent studies suggest that human are commonly exposed. Indeed, a 2.9% seroprevalence of E. lausannensis was documented in Swiss children without pneumonia (Lienard et al., unpublished) and in Dutch women suffering from tubal pathology (Verweij et al., unpublished) and a 12.7% seroprevalence in Vietnamese women was also observed (Hornung et al., unpublished). In
http://dx.doi.org/10.1016/j.micinf.2014.08.003 1286-4579/© 2014 Institut Pasteur. Published by Elsevier Masson SAS. All rights reserved.
M. de Barsy et al. / Microbes and Infection 16 (2014) 746e754
this context, it is important to determine the antibiotic susceptibility of E. lausannensis and we particularly wondered whether this bacterium is susceptible to cyclines and macrolides as already reported for other Chlamydia-related bacteria such as Waddlia chondrophila and Parachlamydia acanthamoebae. We also investigated the resistance of Estrella to quinolones, which are active against Chlamydiaceae but not against Waddlia and Parachlamydia [7e10]. To this end, we tested the antibiotic susceptibility of E. lausannensis on infected Vero cells using two different read-outs, a specific real time PCR (qPCR) [1] and a propidium iodide test. 2. Materials and methods 2.1. Cell culture and bacterial strains Vero cells (ATCC CCL-81) were cultivated, at 37 C in the presence of 5% of CO2, in Dulbecco's modified minimal essential medium (DMEM; GE Healthcare, Pasching, Austria) supplemented with 10% fetal bovine serum (FBS, GE Healthcare). E. lausannensis strain CRIB30 was co-cultivated at 32 C with Acanthamoeba castellani strain ATCC 30010 in 75 cm2 flasks containing 30 ml of peptone-yeast-extract-glucose broth [7]. After 4 days of culture, bacteria were recovered by filtrating the suspension on a 5 mm filter to eliminate trophozoites and cysts. The filtrate containing the bacteria was then diluted to the appropriate dilution in DMEM to proceed to Vero cells infection. 2.2. Infection procedure For the growth kinetics and the determination of MIC by quantitative PCR (qPCR), 2 105 Vero cells were seeded in 24-well plates. For the propidium iodide assay used to determine the cellular mortality, 4 104 Vero cells were seeded in 96-well microplates. Vero cells were infected with a 1/100 dilution of E. lausannensis, corresponding to a multiplicity of infection (MOI) of 10. With this infectious dose, 25% of cells were infected with 1e3 bacteria. The bacterial suspension was centrifuged onto Vero cells at 1790 g for 10 min at room temperature. Infected cells were then incubated for 15 min at 37 C with 5% CO2 atmosphere and cells were washed with the medium to remove non-internalized bacteria and to obtain a synchronous infection that was incubated for different time periods. Antibiotics were added at 2 h post-infection (p.i.), to avoid interference on the internalization process of the bacteria. For the propidium iodide assay, the propidium iodide was added at the same time as the antibiotics. 2.3. Antibiotics In this study, several different antibiotics were used: penicillin G (SigmaeAldrich, Buchs, Switzerland), ceftriaxone (Sigma-Aldrich), ciprofloxacin (Axon Lab, Le-Mont-surLausanne, Switzerland), ofloxacin (SigmaeAldrich), doxycycline (Clontech, Mountain View, CA), tetracycline (Axon Lab) and azithromycin (SigmaeAldrich).
747
2.4. Determination of MIC by qPCR At 32 h p.i., infected cells were recovered by scraping the wells. Genomic DNA was extracted from 50 ml of cells suspension using the Wizard SV Genomic DNA purification kit (Promega, Madison, WI). The qPCR was performed as described [1]. Briefly, the primers and probe, EstF (50 ACACGTGCTACAATGGCCGGT-30 ), EstR (50 -CCGGGAA CGTATTCACGGCGTT-30 ) and EstS (50 -FAM-CAGCCAAC CCGTGAGGG-BHQ1-30 ) were used at a concentration of 200 nM for the primers and 100 nM for the probe (Eurogentec, Seraing, Belgium). In addition to the primers and the probe, the reaction mixture contained 10 ml of iTaq supermix (BioRad, Reinach, Switzerland) and 5 ml of sample for a total volume of 20 ml. The cycling conditions were 95 C during 3 min, followed by 45 cycles of 15 s at 95 C and 1 min at 60 C. The qPCRs were performed on the Step One PCR system (Applied Biosystems, Zug, Switzerland). As negative and positive controls, we used water and a plasmid previously constructed, respectively [1,11]. This plasmid was also used to obtain a standard curve for the quantification of the genomic copy number per well of the 24-well plate used for the infection. Negative controls, standard curve and samples were all analyzed in duplicate. Experiments were repeated three times and a representative experiment was shown. 2.5. Cellular mortality assay by propidium iodide Propidium iodide (SigmaeAldrich) was added at 2 h p.i., at a concentration of 6.6 mg/ml. The 96-well plates were read for fluorescence at different time p.i., using a microplate reader (FLUOstar Omega, BMG LABTECH, Ortenberg, Germany) (Emission: 530 nm, Excitation: 620 nm). As a positive control of cellular mortality, we used Vero cells treated with Triton 0.1% (Axon lab), which was added at the same time p.i. as the antibiotics, i.e. 2 h p.i.. Untreated Vero cells were used as negative control. Each value was compared to the positive control, which was considered as 100% of cellular mortality. Samples and controls were analyzed in triplicate. 2.6. Confocal microscopy At different times p.i., infected Vero cells cultivated on glass coverslips were fixed with cold-ice methanol for 5 min and then washed three times with PBS. The cells were permeabilized in blocking solution (0.1% saponin (SigmaeAldrich), 0.2% NaN3 (Acros, Geel, Belgium) and 10% FBS (GE Healthcare)) at 37 C for at least 30 min. Coverslips were then incubated 1 h with a home-made polyclonal mouse antiE. lausannensis antibody diluted 1/1000 in the blocking solution. Coverslips were washed three times in PBS and incubated with the secondary antibody (Alexa-Fluor 488conjugated goat anti-mouse IgG) diluted 1/1000 in the blocking solution, containing also DAPI (Molecular Probe, Eugene, Oregon, USA) 1 mg/ml and a 1/50 dilution of concanavalin A-Texas Red conjugate (Molecular Probe). After two washes with PBS and once with deionized water,
748
M. de Barsy et al. / Microbes and Infection 16 (2014) 746e754
Fig. 1. E. lausannensis growth within Vero cells and cytopathic effect of the infection on Vero cells. (A) Growth curve of E. lausannensis determined by qPCR. Results of one representative experiment are shown. The qPCR was done in duplicate. The squares represent the mean ± SD. (B) Growth of E. lausannensis assessed by immunofluorescence and confocal microscopy. Bacteria (green) were detected using a polyclonal mouse anti-Estrella antibody followed by a secondary antibody Alexa-Fluor 488-conjugated goat anti-mouse IgG. Vero cells (red) were stained with concanavalin A-Texas Red conjugate and DNA (blue) was stained with the DAPI. Scale bars ¼ 10 mm. (C) Percentage of dead cells determined by the propidium iodide assay. The percentage of dead cells was compared to Vero cells treated with 0.1% of Triton, which was considered as 100% of dead cells. The percentage of dead cells was determined for Vero cells and Vero cells infected with a 1/100 dilution of E. lausannensis.
M. de Barsy et al. / Microbes and Infection 16 (2014) 746e754
coverslips were mounted with mowiol (Calbiochem, CA). Cells were observed under a confocal microscope (Zeiss LSM710). 2.7. Electron microscopy Two 25 cm2 flasks containing 2.5 106 Vero cells were infected with E. lausannensis as described previously. One infected flask was treated with Penicillin G (500 mg/ml). At 32 h p.i., infected cells were harvested after cell scraping and centrifuged at 1000 g for 10 min. Pellets were fixed overnight at 4 C with 1 ml of a “PB” phosphate buffer (19 ml of NaH2PO4 0.2 M þ 81 ml Na2HPO4 0.2 M, pH 7.4) containing 4% of paraformaldehyde (Electron Microscopy Sciences, Hatfield, USA) and 0.2% of glutaraldehyde (Fluka Biochemika, Buchs, Switzerland). Cells were then washed once with the PB buffer. Thin sections on grids were observed using a transmission electron microscope Philips CM100 (Philips, Eindhoven, Netherlands). 2.8. Statistical analysis Statistical analyses were performed using Prism 6.00 for Windows (GraphPad software, San Diego, CA). Multiple comparisons were performed using one-way ANOVA analysis of variance followed by a Dunnet test. 3. Results 3.1. Growth kinetic of E. lausannensis in Vero cells and cytopathic effect Vero cells were infected with E. lausannensis and the growth kinetics were assessed by qPCR to determine the number of genomic copies per well (Fig. 1A). In parallel, immunostaining was performed on fixed infected Vero cells using mouse polyclonal antibody against E. lausannensis (Fig. 1B). We observed a lag phase during the first 8 h, corresponding to the entry of the EBs and the differentiation of EBs into RBs. This lag phase is followed by an exponential phase during which, RBs extensively proliferate. The number of bacteria increased by about 1.5 log (Fig. 1A). Between 24 h and 32 h p.i., we observed a plateau during which EBs are released after cell lysis. A second cycle of infection is then
749
initiated. The same observations can be made by confocal microscopy (Fig. 1B). To better characterize the cytopathic effect of E. lausannensis on Vero cells, the percentage of dead cells was determined using propidium iodide at 8, 24, 32 and 48 h p.i. (Fig. 1C). The cellular mortality of infected Vero cells constantly increased from 24 h p.i. to 48 h p.i.. This is in agreement with the confocal microscopy observations. This cellular mortality induced by E. lausannensis can be used to follow the bacterial growth in the presence of antibiotics. An inhibition of the bacterial growth will lead to a decrease of the cellular mortality induced by E. lausannensis. 3.2. Antibiotic susceptibility of E. lausannensis in infected Vero cells In this study, four classes of antibiotics were tested: blactams (penicillin G, ceftriaxone), fluoroquinolones (ciprofloxacin, ofloxacin), cyclines (doxycycline, tetracycline) and macrolides (azithromycin). We determined the minimal inhibitory concentration (MIC) of each antibiotic, which is the minimal dose of antibiotic that prevents bacterial growth, using qPCR and the propidium iodide assay. The concentrations of antibiotics tested were chosen based on the antibiotics MICs previously determined for W. chondrophila [8] and Parachlamydia acanthamoebae [7,12] (Table 1). None antibiotics induced the mortality of the Vero cells (data not shown). E. lausannensis was resistant to penicillin G (MIC>32 mg/ ml), ceftriaxone (MIC>32 mg/ml), ciprofloxacin (MIC ¼ 32 mg/ ml) and ofloxacin (MIC ¼ 16 mg/ml), as demonstrated by the significant bacterial growth measured by qPCR (Fig. 2A). We considered that E. lausannensis is resistant to ciprofloxacin and ofloxacin because usually the antibiotic seric concentration do not exceed 4 mg/ml. These results were confirmed using the propidium iodide assay. Indeed, we observed no decrease (or a minor decrease) of the cellular mortality of infected cells treated with penicillin G (32 mg/ml) and ceftriaxone (32 mg/ml) (Fig. 2B). Infected cells treated with ciprofloxacin and ofloxacin exhibited a significant decrease of the cellular mortality when using 32 mg/ml of ciprofloxacin and 16 mg/ml of ofloxacin (Fig. 2B). This is in agreement with the MIC determined by qPCR. Interestingly, treatment of infected Vero cells with a high concentration of penicillin or ceftriaxone (500 mg/ml) induced
Table 1 Minimal inhibitory concentration (MIC) of E. lausannensis and other members of the Chlamydiales order.
Ceftriaxone Penicillin derivatives Ciprofloxacin Ofloxacin Doxycycline Tetracycline Azithromycin a
E. lausannensis [this study]
W. chondrophila [8]
P. acanthamoebaea [7,12]
C. trachomatis [10,20e26]
C. pneumoniae [9,25e28]
>32 >32 ¼32 ¼16 0.25 0.25e0.5 2
>32 >32 >16 >16 0.25 ND 0.25
>32 >32 >16 >16 0.5e1 ND ND
>32 ND >0.5e2 0.5e1 0.03e0.025 ND 0.125e0.25
ND >100 1e2 0.5e1 0.03e0.06 ND 0.06e0.125
MIC determined in amoebae; ND ¼ not determined.
750
M. de Barsy et al. / Microbes and Infection 16 (2014) 746e754
Fig. 2. Antibiotic susceptibility of E. lausannensis to penicillin G (A and B), ceftriaxone (A and B), ciprofloxacin (A and B), ofloxacin (A and B), doxycycline (C and D), tetracycline (C and D) and azithromycin (E and F), determined by qPCR (A, C and E) and a propidium iodideebased cellular mortality assay (B, D and F). Graphs of one representative experiment are shown. The values correspond to the means ± SD. Significant differences, compared to the untreated infected Vero cells at 32 h p.i., were indicated by *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.
M. de Barsy et al. / Microbes and Infection 16 (2014) 746e754
751
Fig. 3. Antibiotic treatment effect on E. lausannensis growth and morphology assessed by confocal microscopy and electron microscopy. (A) Infected Vero cells were treated with azithromycin 0.5 mg/ml or 4 mg/ml, penicillin G 500 mg/ml and ceftriaxone 500 mg/ml. Cells were fixed at 32 h p.i. and an immunofluorescence was performed using the polyclonal mouse anti-Estrella (green). Vero cells were stained with concanavalin A-Texas Red conjugate (red) and DNA was stained with the DAPI (blue). Scale bars ¼ 10 mm. (B) Electron micrographs of E. lausannensis aberrant bodies in Vero cells at 32 h p.i. treated with penicillin (500 mg/ml). AB ¼ aberrant bodies, N ¼ nucleus. EBs are indicated by arrowheads and mitochondria by arrows. Scale bars ¼ 2 mm.
the formation of aberrant bodies observed by confocal microscopy and electron microscopy (Fig. 3A and B). We also showed that E. lausannensis is susceptible to doxycycline and tetracycline with a MIC of 0.25 mg/ml and 0.25e0.5 mg/ml, respectively (Fig. 2C and D). For azithromycin, we determined a MIC of 2 mg/ml (Fig. 2E). These results were confirmed by the propidium iodide with a strong decreased cellular mortality of infected Vero cells at a
concentration of 2 mg/ml for azithromycin (Fig. 2F). The MIC of the tetracycline was confirmed with the propidium iodide assay (Fig. 2D). However, we were not able to observe an increased cellular mortality when the doxycycline concentrations were decreased (Fig. 2D), suggesting that doxycycline may interfere with this assay. Using confocal microscopy (Fig. 3A), we observed inclusions with few bacteria when adding 4 mg/ml of
752
M. de Barsy et al. / Microbes and Infection 16 (2014) 746e754
Fig. 4. QRDR alignment of DNA gyrase A (GyrA) (A) and topoisomerase IV (ParC) (B) of Estrella lausannensis strain CRIB30, E. coli strain K-12 substr. MG1655, C. trachomatis strain D/UW-3/CX, C. pneumoniae strain CWL029, C. caviae strain GPIC, S. negevensis strain Z, Waddlia chondrophila strain WSU 861044, Parachlamydia acanthamoebae strain UV-7 and strain Hall's coccus, Protochlamydia neagleriophila strain KNIC, Protochlamydia amoebophila strain UWE25, Criblamydia sequanensis strain CRIB18. Multiple alignments were performed using MUSCLE program. Genbank accession numbers are indicated in brackets. QRDR are delimited by the two red bar. Amino acids that could confer resistance to quinolone are indicated in blue (E. coli numbering).
azithromycin, showing the bacterial growth inhibition with this concentration. In contrast, 0.5 mg/ml of azithromycin had no effect on the bacterial growth since we observed large inclusions filled with many bacteria. 3.3. Sequences of the QRDR and of the 23S encoding genes To identify possible mutations of the quinolone resistancedetermining region (QRDR) of GyrA and ParC that may explain the observed resistance phenotype to fluoroquinolones, we compared the sequences of E. lausannensis with that of other Chlamydiales. Mutations at position 83 and 70 in the GyrA QRDR and at position 80 and 84 in the ParC QRDR were observed (Fig. 4). A mutation at position 2059 was also observed in the 23S encoding gene of E. lausannensis, likely explaining the observed resistance to azithromycin (Fig. 5). 4. Discussion E. lausannensis is resistant to b-lactams and fluoroquinolones antibiotics and sensitive to cyclines, demonstrating that E. lausannensis have the same antibiotics susceptibility pattern as W. chondrophila and P. acanthamoebae, two other Chlamydia-related bacteria [7,8,12]. Quinolone resistance could be associated to amino-acids substitutions in the quinolone resistance-determining region (QRDR) of ParC and GyrA. Indeed, we observed two amino acids substitutions in the GyrA QRDR of E. lausannensis, Ser83 / Gln and Val70 / Ser, which are also present in the GyrA QRDR of W.
chondrophila and P. acanthamoebae likely participating to the quinolones resistance phenotype [7,8]. Ser83 substitution of GyrA is frequently observed and may induce a high level of resistance to quinolones [13]. In the ParC QRDR, we identified two amino acid substitutions, Ser80 / Ala and Glu84 / Asp. These two substitutions are also observed in the ParC QRDR of W. chondrophila [8]. Substitutions of Ser80 and Glu84 are frequently observed in fluoroquinolone resistant E. coli and Streptococcus pneumoniae [13]. Substitutions in the GyrA QRDR and ParC QRDR are thus likely responsible for fluoroquinolone resistance phenotype of E. lausannensis. Resistance of E. lausannensis to the b-lactam antibiotics is probably due to the presence of several b-lactamase genes in the genome (data not shown). Nevertheless, a high concentration of Penicillin G inhibits division leading to aberrant bodies. These aberrant bodies were initially reported for Chlamydia trachomatis and were considered as a persistence form [14,15]. Such aberrant bodies were also reported following treatment of W. chondrophila and Chlamydia pneumoniae with penicillin derivatives demonstrating the major role of peptidoglycan in chlamydial division [16]. Concerning azithromycin, we determined a MIC of 2 mg/ml for E. lausannensis. This is more than 10 fold higher than the MIC for C. trachomatis, C. pneumoniae and W. chondrophila (Table 1). Thus, E. lausannensis should be considered as resistant to azithromycin. The resistance to macrolides is often associated with a mutation in the peptdyl transferase domain of the 23S ribosomal RNA for pathogens that possess a low number of rRNA operons. Several mutations were already reported as conferring resistance to macrolides, such as
M. de Barsy et al. / Microbes and Infection 16 (2014) 746e754
753
Fig. 5. Alignment of the 23S rRNA gene. In blue, nucleotide mutations that could confer resistance to macrolides in the beginning (A) of the peptidyl transferase domain and in the end of this domain (B). Genbank accession numbers are indicated in brackets and the alignment was performed using MUSCLE program. The mutation that may explain the slight resistance of E. lausannensis to macrolide is highlighted in red (E. coli numbering). In green, the 2 base-paired nucleotides involved in the hairpin.
mutations at positions 2057, 2058, 2059, 2611 [17]. Mutations at position 2057 or 2611 only confer a low-level of resistance to macrolides compared to mutations at position 2058 or 2059 [17]. It is thought that these mutations perturb the structure of the binding pocket of the drug leading to the inability to interact and to inhibit the ribosome activity. Chlamydiaceae members have one or two copies of the rRNA operon, depending on the species (Table 2). This could give rise to the emergence of macrolide resistant bacteria with mutations in the 23S rRNA gene. Recently, several spontaneous resistant mutants of Chlamydia psittaci strain 6BC, C. trachomatis strain L2/434/Bu and several Chlamydia caviae isolates, were isolated in vitro after treatment with azithromycin [18,19]. Interestingly, C. psittaci and C. caviae, which possess only one 23S rRNA gene copy, harbored a single mutation in the 23S rRNA gene, i.e. A2058C or A2059C or A2059G. For C. trachomatis strain L2/434/Bu, which possesses two copies of the 23S rRNA gene, no mutation was found in any of these two gene copies. Nevertheless, a Gln to Lys substitution at position 66, was identified in the L4 ribosomal protein, which probably confers resistance to azithromycin. E. lausannensis possesses three copies of the 23S rRNA gene. Multiple alignments allowed us to identify A2059C mutation in all the three copies of E. lausannensis 23S rRNA gene, explaining the observed resistance to azithromycin. Interestingly, E. lausannensis is the only bacterium among all Chlamydia-related bacteria sequenced so far, which possesses this mutation. It could be interesting to test the azitromycin susceptibility of Criblamydia sequanensis, the other member of the Criblamydiaceae family, which do not exhibit this mutation (Fig. 5).
Based on this in vitro study, doxycycline should be considered as the treatment for E. lausannensis infection. It will be also important to assess the activity of this antibiotic against E. lausannensis in vivo. Our work also showed that the propidium iodide assay is useful to indirectly assess the growth of bacteria that generally induce host cell lysis, such as most Chlamydia-related bacteria. The main advantages of the propidium iodide assay over qPCR are that (1) it can be performed in 96-well plates, (2) it does not need to extract genomic DNA and (3) we can follow the mortality in real time. Thus, propidium iodide is less time consuming and represents a suitable screening approach that may be applied to a large amount of compounds and that may help identifying new antibiotics efficient on E. lausannensis. Table 2 Number of rRNA operon in different Chlamydiales bacterial strains compared to E. coli. Number of rRNA operon Escherichia coli strain K-12 Chlamydia trachomatis strain L2/434/Bu Chlamydia pneumoniae strain CWL029 Chlamydia caviae strain GPIC Simkania negevensis strain Z Waddlia chondrophila strain WSU 86-1044 Parachlamydia acanthamoebae strain UV-7 Parachlamydia acanthamoebae strain Hall's coccus Protochlaymdia neagleriophila strain KNIC Protochlamydia amoebophila strain UWE25 Criblamydia sequanensis strain CRIB18 Estrella lausannensis strain CRIB30
7 2 1 1 1 2 3 1 4 3 1 3
754
M. de Barsy et al. / Microbes and Infection 16 (2014) 746e754
Conflict of interest The authors have no conflict of interest. We thank S. Aeby for technical help. We also thank the Electron Microscopy Facility platform of the University of Lausanne as well as the PFMU at the Medical Faculty of Geneva, for assisting with the electron microscopy analyses. We also thank the Cellular Imaging Facility platform of the University of Lausanne for confocal microscopy. Acknowledgment The group of Prof. G. Greub is supported by the Swiss National Science Foundation grants n 310030-141050 and the SNSF Sinergia grant n CRSII3-141837. References [1] Lienard J, Croxatto A, Prod'hom G, Greub G. Estrella lausannensis, a new star in the Chlamydiales order. Microbes Infect 2011;13:1232e41. [2] Thomas V, Casson N, Greub G. Criblamydia sequanensis, a new intracellular Chlamydiales isolated from Seine river water using amoebal coculture. Environ Microbiol 2006;8:2125e35. [3] Greub G, Raoult D. Crescent bodies of Parachlamydia acanthamoeba and its life cycle within Acanthamoeba polyphaga: an electron micrograph study. Appl Environ Microbiol 2002;68:3076e84. [4] Moulder JW. Interaction of chlamydiae and host cells in vitro. Microbiol Rev 1991;55:143e90. [5] Rusconi B, Lienard J, Aeby S, Croxatto A, Bertelli C, Greub G. Crescent and star shapes of members of the Chlamydiales order: impact of fixative methods. Ant Van Leeuwenhoek 2013;104:521e32. [6] Kebbi-Beghdadi C, Batista C, Greub G. Permissivity of fish cell lines to three Chlamydia-related bacteria: Waddlia chondrophila, Estrella lausannensis and Parachlamydia acanthamoebae. FEMS Immunol Med Microbiol 2011;63:339e45. [7] Casson N, Greub G. Resistance of different Chlamydia-like organisms to quinolones and mutations in the quinoline resistance-determining region of the DNA gyrase A- and topoisomerase-encoding genes. Int J Antimicrob Agents 2006;27:541e4. [8] Goy G, Greub G. Antibiotic susceptibility of Waddlia chondrophila in Acanthamoeba castellanii amoebae. Antimicrob Agents Chemother 2009;53:2663e6. [9] Hammerschlag MR, Qumei KK, Roblin PM. In vitro activities of azithromycin, clarithromycin, L-ofloxacin, and other antibiotics against Chlamydia pneumoniae. Antimicrob Agents Chemother 1992;36:1573e4. [10] Smelov V, Perekalina T, Gorelov A, Smelova N, Artemenko N, Norman L. In vitro activity of fluoroquinolones, azithromycin and doxycycline against Chlamydia trachomatis cultured from men with chronic lower urinary tract symptoms. Eur Urol 2004;46:647e50.
[11] Casson N, Posfay-Barbe KM, Gervaix A, Greub G. New diagnostic realtime PCR for specific detection of Parachlamydia acanthamoebae DNA in clinical samples. J Clin Microbiol 2008;46:1491e3. [12] Maurin M, Bryskier A, Raoult D. Antibiotic susceptibilities of Parachlamydia acanthamoeba in amoebae. Antimicrob Agents Chemother 2002;46:3065e7. [13] Nakamura Shinichi. Mechanisms of quinolone resistance. J Infect Chemother 1997;3:128e38. [14] Coles AM, Reynolds DJ, Harper A, Devitt A, Pearce JH. Lownutrient induction of abnormal chlamydial development: a novel component of chlamydial pathogenesis? FEMS Microbiol Lett 1993;106:193e200. [15] Raulston JE. Response of Chlamydia trachomatis serovar E to iron restriction in vitro and evidence for iron-regulated chlamydial proteins. Infect Immun 1997;65:4539e47. [16] Jacquier N, Frandi A, Pillonel T, Viollier P, Greub G. Cell wall precursors are required to organize the chlamydial division septum. Nat Commun 2014;5:3578. [17] Vester B, Douthwaite S. Macrolide resistance conferred by base substitutions in 23S rRNA. Antimicrob Agents Chemother 2001;45:1e12. [18] Binet R, Bowlin AK, Maurelli AT, Rank RG. Impact of azithromycin resistance mutations on the virulence and fitness of Chlamydia caviae in guinea pigs. Antimicrob Agents Chemother 2010;54:1094e101. [19] Binet R, Maurelli AT. Frequency of development and associated physiological cost of azithromycin resistance in Chlamydia psittaci 6BC and C. trachomatis L2. Antimicrob Agents Chemother 2007;51:4267e75. [20] Dreses-Werringloer U, Padubrin I, Zeidler H, Kohler L. Effects of azithromycin and rifampin on Chlamydia trachomatis infection in vitro. Antimicrob Agents Chemother 2001;45:3001e8. [21] Hammerschlag MR, Gleyzer A. In vitro activity of a group of broadspectrum cephalosporins and other beta-lactam antibiotics against Chlamydia trachomatis. Antimicrob Agents Chemother 1983;23:493e4. [22] Samra Z, Rosenberg S, Soffer Y, Dan M. In vitro susceptibility of recent clinical isolates of Chlamydia trachomatis to macrolides and tetracyclines. Diagn Microbiol Infect Dis 2001;39:177e9. [23] Smelov V, Krylova T, Smelova N, Norman L. Azithromycin treatment follow-up: antibacterial susceptibility of Chlamydia trachomatis in patients with chronic prostatitis. Int J Antimicrob Agents 2004;23(Suppl. 1):S79e82. [24] Nilius AM, Shen LL, Hensey-Rudloff D, Almer LS, Beyer JM, Balli DJ, et al. In vitro antibacterial potency and spectrum of ABT-492, a new fluoroquinolone. Antimicrob Agents Chemother 2003;47:3260e9. [25] Miyashita N, Niki Y, Kishimoto T, Nakajima M, Matsushima T. In vitro and in vivo activities of AM-1155, a new fluoroquinolone, against Chlamydia spp. Antimicrob Agents Chemother 1997;41:1331e4. [26] Senn L, Hammerschlag MR, Greub G. Therapeutic approaches to Chlamydia infections. Expert Opin Pharmacother 2005;6:2281e90. [27] Chirgwin K, Roblin PM, Hammerschlag MR. In vitro susceptibilities of Chlamydia pneumoniae (Chlamydia sp. strain TWAR). Antimicrob Agents Chemother 1989;33:1634e5. [28] Kuo CC, Grayston JT. In vitro drug susceptibility of Chlamydia sp. strain TWAR. Antimicrob Agents Chemother 1988;32:257e8.
Original article
Antibiotic susceptibility of Estrella lausannensis, a potential emerging pathogen Marie de Barsy, Lavinia Bottinelli, Gilbert Greub* Institute of Microbiology, University Hospital Center and University of Lausanne, Lausanne, Switzerland Received 27 June 2014; accepted 5 August 2014 Available online 20 August 2014
Abstract Estrella lausannensis is a new Chlamydia-related bacterium, belonging to the Criblamydiaceae family. As suggested by its species name, this bacterium harbors a peculiar star shape. E. lausannensis is able to infect a wide range of amoebal, fish and mammalian cell lines. Moreover, seroprevalence of 2.9% was reported in children and in women with tubal pathology, showing that humans are commonly exposed to this recently discovered strict intracellular bacteria considered as a potential pathogen. Antibiotic susceptibility was determined using two approaches: qPCR and cellular mortality assay. Antibiotics classically used against intracellular bacteria were tested, including b-lactams, fluoroquinolones, cyclines and macrolides. We showed that E. lausannensis is resistant to b-lactams and fluoroquinolones, and sensitive to cyclines. Interestingly, E. lausannensis is slightly resistant to azithromycin with a MIC of 2 mg/ml, which is 10 fold higher compared to Waddlia chondrophila and Parachlamydia acanthamoebae MIC's. A single A2059C mutation in 23S rRNA gene could be responsible for this unexpected resistance. © 2014 Institut Pasteur. Published by Elsevier Masson SAS. All rights reserved.
Keywords: Antibiotic susceptibility; Cell culture; Chlamydiae; Chlamydia-like organism
1. Introduction Estrella lausannensis is a newly discovered Chlamydiarelated bacterium, isolated from a water sample taken upstream of a water treatment plant fed by the river Llobregat in Spain [1]. E. lausannensis belongs to the Criblamydiaceae family in the Chlamydiales order [1,2]. Like other members of the Chlamydiales order [3,4], E. lausannensis harbors a biphasic developmental cycle, with infectious elementary bodies (EBs), which infect cells and differentiate into reticulate bodies (RBs), the proliferative developmental forms [1]. * Corresponding author. Center for Research on Intracellular Bacteria (CRIB), Institute of Microbiology, University Hospital Center and University of Lausanne, Bugnon 48, 1011 Lausanne, Switzerland. Tel.: þ41 21 314 49 79; fax: þ41 21 314 40 60. E-mail address: [email protected] (G. Greub).
Then, RBs redifferentiate into EBs, which are released after cell lysis [1]. A new infection cycle can then be initiated. E. lausannensis EBs exhibit a star shape, similar to that observed for Criblamydia sequanensis [1,2]. Although this star shape corresponds to a fixative artefact, this highlights differences in structure and composition of the bacterial cell wall [5]. E. lausannensis is able to infect a wide range of amoebal hosts [1] as well as fish cell lines [6] and macrophages (Rusconi and Greub, unpublished). The ability to resist to human macrophage microbicidal effectors suggests a potential pathogenic role of E. lausannensis. Moreover, different recent studies suggest that human are commonly exposed. Indeed, a 2.9% seroprevalence of E. lausannensis was documented in Swiss children without pneumonia (Lienard et al., unpublished) and in Dutch women suffering from tubal pathology (Verweij et al., unpublished) and a 12.7% seroprevalence in Vietnamese women was also observed (Hornung et al., unpublished). In
http://dx.doi.org/10.1016/j.micinf.2014.08.003 1286-4579/© 2014 Institut Pasteur. Published by Elsevier Masson SAS. All rights reserved.
M. de Barsy et al. / Microbes and Infection 16 (2014) 746e754
this context, it is important to determine the antibiotic susceptibility of E. lausannensis and we particularly wondered whether this bacterium is susceptible to cyclines and macrolides as already reported for other Chlamydia-related bacteria such as Waddlia chondrophila and Parachlamydia acanthamoebae. We also investigated the resistance of Estrella to quinolones, which are active against Chlamydiaceae but not against Waddlia and Parachlamydia [7e10]. To this end, we tested the antibiotic susceptibility of E. lausannensis on infected Vero cells using two different read-outs, a specific real time PCR (qPCR) [1] and a propidium iodide test. 2. Materials and methods 2.1. Cell culture and bacterial strains Vero cells (ATCC CCL-81) were cultivated, at 37 C in the presence of 5% of CO2, in Dulbecco's modified minimal essential medium (DMEM; GE Healthcare, Pasching, Austria) supplemented with 10% fetal bovine serum (FBS, GE Healthcare). E. lausannensis strain CRIB30 was co-cultivated at 32 C with Acanthamoeba castellani strain ATCC 30010 in 75 cm2 flasks containing 30 ml of peptone-yeast-extract-glucose broth [7]. After 4 days of culture, bacteria were recovered by filtrating the suspension on a 5 mm filter to eliminate trophozoites and cysts. The filtrate containing the bacteria was then diluted to the appropriate dilution in DMEM to proceed to Vero cells infection. 2.2. Infection procedure For the growth kinetics and the determination of MIC by quantitative PCR (qPCR), 2 105 Vero cells were seeded in 24-well plates. For the propidium iodide assay used to determine the cellular mortality, 4 104 Vero cells were seeded in 96-well microplates. Vero cells were infected with a 1/100 dilution of E. lausannensis, corresponding to a multiplicity of infection (MOI) of 10. With this infectious dose, 25% of cells were infected with 1e3 bacteria. The bacterial suspension was centrifuged onto Vero cells at 1790 g for 10 min at room temperature. Infected cells were then incubated for 15 min at 37 C with 5% CO2 atmosphere and cells were washed with the medium to remove non-internalized bacteria and to obtain a synchronous infection that was incubated for different time periods. Antibiotics were added at 2 h post-infection (p.i.), to avoid interference on the internalization process of the bacteria. For the propidium iodide assay, the propidium iodide was added at the same time as the antibiotics. 2.3. Antibiotics In this study, several different antibiotics were used: penicillin G (SigmaeAldrich, Buchs, Switzerland), ceftriaxone (Sigma-Aldrich), ciprofloxacin (Axon Lab, Le-Mont-surLausanne, Switzerland), ofloxacin (SigmaeAldrich), doxycycline (Clontech, Mountain View, CA), tetracycline (Axon Lab) and azithromycin (SigmaeAldrich).
747
2.4. Determination of MIC by qPCR At 32 h p.i., infected cells were recovered by scraping the wells. Genomic DNA was extracted from 50 ml of cells suspension using the Wizard SV Genomic DNA purification kit (Promega, Madison, WI). The qPCR was performed as described [1]. Briefly, the primers and probe, EstF (50 ACACGTGCTACAATGGCCGGT-30 ), EstR (50 -CCGGGAA CGTATTCACGGCGTT-30 ) and EstS (50 -FAM-CAGCCAAC CCGTGAGGG-BHQ1-30 ) were used at a concentration of 200 nM for the primers and 100 nM for the probe (Eurogentec, Seraing, Belgium). In addition to the primers and the probe, the reaction mixture contained 10 ml of iTaq supermix (BioRad, Reinach, Switzerland) and 5 ml of sample for a total volume of 20 ml. The cycling conditions were 95 C during 3 min, followed by 45 cycles of 15 s at 95 C and 1 min at 60 C. The qPCRs were performed on the Step One PCR system (Applied Biosystems, Zug, Switzerland). As negative and positive controls, we used water and a plasmid previously constructed, respectively [1,11]. This plasmid was also used to obtain a standard curve for the quantification of the genomic copy number per well of the 24-well plate used for the infection. Negative controls, standard curve and samples were all analyzed in duplicate. Experiments were repeated three times and a representative experiment was shown. 2.5. Cellular mortality assay by propidium iodide Propidium iodide (SigmaeAldrich) was added at 2 h p.i., at a concentration of 6.6 mg/ml. The 96-well plates were read for fluorescence at different time p.i., using a microplate reader (FLUOstar Omega, BMG LABTECH, Ortenberg, Germany) (Emission: 530 nm, Excitation: 620 nm). As a positive control of cellular mortality, we used Vero cells treated with Triton 0.1% (Axon lab), which was added at the same time p.i. as the antibiotics, i.e. 2 h p.i.. Untreated Vero cells were used as negative control. Each value was compared to the positive control, which was considered as 100% of cellular mortality. Samples and controls were analyzed in triplicate. 2.6. Confocal microscopy At different times p.i., infected Vero cells cultivated on glass coverslips were fixed with cold-ice methanol for 5 min and then washed three times with PBS. The cells were permeabilized in blocking solution (0.1% saponin (SigmaeAldrich), 0.2% NaN3 (Acros, Geel, Belgium) and 10% FBS (GE Healthcare)) at 37 C for at least 30 min. Coverslips were then incubated 1 h with a home-made polyclonal mouse antiE. lausannensis antibody diluted 1/1000 in the blocking solution. Coverslips were washed three times in PBS and incubated with the secondary antibody (Alexa-Fluor 488conjugated goat anti-mouse IgG) diluted 1/1000 in the blocking solution, containing also DAPI (Molecular Probe, Eugene, Oregon, USA) 1 mg/ml and a 1/50 dilution of concanavalin A-Texas Red conjugate (Molecular Probe). After two washes with PBS and once with deionized water,
748
M. de Barsy et al. / Microbes and Infection 16 (2014) 746e754
Fig. 1. E. lausannensis growth within Vero cells and cytopathic effect of the infection on Vero cells. (A) Growth curve of E. lausannensis determined by qPCR. Results of one representative experiment are shown. The qPCR was done in duplicate. The squares represent the mean ± SD. (B) Growth of E. lausannensis assessed by immunofluorescence and confocal microscopy. Bacteria (green) were detected using a polyclonal mouse anti-Estrella antibody followed by a secondary antibody Alexa-Fluor 488-conjugated goat anti-mouse IgG. Vero cells (red) were stained with concanavalin A-Texas Red conjugate and DNA (blue) was stained with the DAPI. Scale bars ¼ 10 mm. (C) Percentage of dead cells determined by the propidium iodide assay. The percentage of dead cells was compared to Vero cells treated with 0.1% of Triton, which was considered as 100% of dead cells. The percentage of dead cells was determined for Vero cells and Vero cells infected with a 1/100 dilution of E. lausannensis.
M. de Barsy et al. / Microbes and Infection 16 (2014) 746e754
coverslips were mounted with mowiol (Calbiochem, CA). Cells were observed under a confocal microscope (Zeiss LSM710). 2.7. Electron microscopy Two 25 cm2 flasks containing 2.5 106 Vero cells were infected with E. lausannensis as described previously. One infected flask was treated with Penicillin G (500 mg/ml). At 32 h p.i., infected cells were harvested after cell scraping and centrifuged at 1000 g for 10 min. Pellets were fixed overnight at 4 C with 1 ml of a “PB” phosphate buffer (19 ml of NaH2PO4 0.2 M þ 81 ml Na2HPO4 0.2 M, pH 7.4) containing 4% of paraformaldehyde (Electron Microscopy Sciences, Hatfield, USA) and 0.2% of glutaraldehyde (Fluka Biochemika, Buchs, Switzerland). Cells were then washed once with the PB buffer. Thin sections on grids were observed using a transmission electron microscope Philips CM100 (Philips, Eindhoven, Netherlands). 2.8. Statistical analysis Statistical analyses were performed using Prism 6.00 for Windows (GraphPad software, San Diego, CA). Multiple comparisons were performed using one-way ANOVA analysis of variance followed by a Dunnet test. 3. Results 3.1. Growth kinetic of E. lausannensis in Vero cells and cytopathic effect Vero cells were infected with E. lausannensis and the growth kinetics were assessed by qPCR to determine the number of genomic copies per well (Fig. 1A). In parallel, immunostaining was performed on fixed infected Vero cells using mouse polyclonal antibody against E. lausannensis (Fig. 1B). We observed a lag phase during the first 8 h, corresponding to the entry of the EBs and the differentiation of EBs into RBs. This lag phase is followed by an exponential phase during which, RBs extensively proliferate. The number of bacteria increased by about 1.5 log (Fig. 1A). Between 24 h and 32 h p.i., we observed a plateau during which EBs are released after cell lysis. A second cycle of infection is then
749
initiated. The same observations can be made by confocal microscopy (Fig. 1B). To better characterize the cytopathic effect of E. lausannensis on Vero cells, the percentage of dead cells was determined using propidium iodide at 8, 24, 32 and 48 h p.i. (Fig. 1C). The cellular mortality of infected Vero cells constantly increased from 24 h p.i. to 48 h p.i.. This is in agreement with the confocal microscopy observations. This cellular mortality induced by E. lausannensis can be used to follow the bacterial growth in the presence of antibiotics. An inhibition of the bacterial growth will lead to a decrease of the cellular mortality induced by E. lausannensis. 3.2. Antibiotic susceptibility of E. lausannensis in infected Vero cells In this study, four classes of antibiotics were tested: blactams (penicillin G, ceftriaxone), fluoroquinolones (ciprofloxacin, ofloxacin), cyclines (doxycycline, tetracycline) and macrolides (azithromycin). We determined the minimal inhibitory concentration (MIC) of each antibiotic, which is the minimal dose of antibiotic that prevents bacterial growth, using qPCR and the propidium iodide assay. The concentrations of antibiotics tested were chosen based on the antibiotics MICs previously determined for W. chondrophila [8] and Parachlamydia acanthamoebae [7,12] (Table 1). None antibiotics induced the mortality of the Vero cells (data not shown). E. lausannensis was resistant to penicillin G (MIC>32 mg/ ml), ceftriaxone (MIC>32 mg/ml), ciprofloxacin (MIC ¼ 32 mg/ ml) and ofloxacin (MIC ¼ 16 mg/ml), as demonstrated by the significant bacterial growth measured by qPCR (Fig. 2A). We considered that E. lausannensis is resistant to ciprofloxacin and ofloxacin because usually the antibiotic seric concentration do not exceed 4 mg/ml. These results were confirmed using the propidium iodide assay. Indeed, we observed no decrease (or a minor decrease) of the cellular mortality of infected cells treated with penicillin G (32 mg/ml) and ceftriaxone (32 mg/ml) (Fig. 2B). Infected cells treated with ciprofloxacin and ofloxacin exhibited a significant decrease of the cellular mortality when using 32 mg/ml of ciprofloxacin and 16 mg/ml of ofloxacin (Fig. 2B). This is in agreement with the MIC determined by qPCR. Interestingly, treatment of infected Vero cells with a high concentration of penicillin or ceftriaxone (500 mg/ml) induced
Table 1 Minimal inhibitory concentration (MIC) of E. lausannensis and other members of the Chlamydiales order.
Ceftriaxone Penicillin derivatives Ciprofloxacin Ofloxacin Doxycycline Tetracycline Azithromycin a
E. lausannensis [this study]
W. chondrophila [8]
P. acanthamoebaea [7,12]
C. trachomatis [10,20e26]
C. pneumoniae [9,25e28]
>32 >32 ¼32 ¼16 0.25 0.25e0.5 2
>32 >32 >16 >16 0.25 ND 0.25
>32 >32 >16 >16 0.5e1 ND ND
>32 ND >0.5e2 0.5e1 0.03e0.025 ND 0.125e0.25
ND >100 1e2 0.5e1 0.03e0.06 ND 0.06e0.125
MIC determined in amoebae; ND ¼ not determined.
750
M. de Barsy et al. / Microbes and Infection 16 (2014) 746e754
Fig. 2. Antibiotic susceptibility of E. lausannensis to penicillin G (A and B), ceftriaxone (A and B), ciprofloxacin (A and B), ofloxacin (A and B), doxycycline (C and D), tetracycline (C and D) and azithromycin (E and F), determined by qPCR (A, C and E) and a propidium iodideebased cellular mortality assay (B, D and F). Graphs of one representative experiment are shown. The values correspond to the means ± SD. Significant differences, compared to the untreated infected Vero cells at 32 h p.i., were indicated by *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.
M. de Barsy et al. / Microbes and Infection 16 (2014) 746e754
751
Fig. 3. Antibiotic treatment effect on E. lausannensis growth and morphology assessed by confocal microscopy and electron microscopy. (A) Infected Vero cells were treated with azithromycin 0.5 mg/ml or 4 mg/ml, penicillin G 500 mg/ml and ceftriaxone 500 mg/ml. Cells were fixed at 32 h p.i. and an immunofluorescence was performed using the polyclonal mouse anti-Estrella (green). Vero cells were stained with concanavalin A-Texas Red conjugate (red) and DNA was stained with the DAPI (blue). Scale bars ¼ 10 mm. (B) Electron micrographs of E. lausannensis aberrant bodies in Vero cells at 32 h p.i. treated with penicillin (500 mg/ml). AB ¼ aberrant bodies, N ¼ nucleus. EBs are indicated by arrowheads and mitochondria by arrows. Scale bars ¼ 2 mm.
the formation of aberrant bodies observed by confocal microscopy and electron microscopy (Fig. 3A and B). We also showed that E. lausannensis is susceptible to doxycycline and tetracycline with a MIC of 0.25 mg/ml and 0.25e0.5 mg/ml, respectively (Fig. 2C and D). For azithromycin, we determined a MIC of 2 mg/ml (Fig. 2E). These results were confirmed by the propidium iodide with a strong decreased cellular mortality of infected Vero cells at a
concentration of 2 mg/ml for azithromycin (Fig. 2F). The MIC of the tetracycline was confirmed with the propidium iodide assay (Fig. 2D). However, we were not able to observe an increased cellular mortality when the doxycycline concentrations were decreased (Fig. 2D), suggesting that doxycycline may interfere with this assay. Using confocal microscopy (Fig. 3A), we observed inclusions with few bacteria when adding 4 mg/ml of
752
M. de Barsy et al. / Microbes and Infection 16 (2014) 746e754
Fig. 4. QRDR alignment of DNA gyrase A (GyrA) (A) and topoisomerase IV (ParC) (B) of Estrella lausannensis strain CRIB30, E. coli strain K-12 substr. MG1655, C. trachomatis strain D/UW-3/CX, C. pneumoniae strain CWL029, C. caviae strain GPIC, S. negevensis strain Z, Waddlia chondrophila strain WSU 861044, Parachlamydia acanthamoebae strain UV-7 and strain Hall's coccus, Protochlamydia neagleriophila strain KNIC, Protochlamydia amoebophila strain UWE25, Criblamydia sequanensis strain CRIB18. Multiple alignments were performed using MUSCLE program. Genbank accession numbers are indicated in brackets. QRDR are delimited by the two red bar. Amino acids that could confer resistance to quinolone are indicated in blue (E. coli numbering).
azithromycin, showing the bacterial growth inhibition with this concentration. In contrast, 0.5 mg/ml of azithromycin had no effect on the bacterial growth since we observed large inclusions filled with many bacteria. 3.3. Sequences of the QRDR and of the 23S encoding genes To identify possible mutations of the quinolone resistancedetermining region (QRDR) of GyrA and ParC that may explain the observed resistance phenotype to fluoroquinolones, we compared the sequences of E. lausannensis with that of other Chlamydiales. Mutations at position 83 and 70 in the GyrA QRDR and at position 80 and 84 in the ParC QRDR were observed (Fig. 4). A mutation at position 2059 was also observed in the 23S encoding gene of E. lausannensis, likely explaining the observed resistance to azithromycin (Fig. 5). 4. Discussion E. lausannensis is resistant to b-lactams and fluoroquinolones antibiotics and sensitive to cyclines, demonstrating that E. lausannensis have the same antibiotics susceptibility pattern as W. chondrophila and P. acanthamoebae, two other Chlamydia-related bacteria [7,8,12]. Quinolone resistance could be associated to amino-acids substitutions in the quinolone resistance-determining region (QRDR) of ParC and GyrA. Indeed, we observed two amino acids substitutions in the GyrA QRDR of E. lausannensis, Ser83 / Gln and Val70 / Ser, which are also present in the GyrA QRDR of W.
chondrophila and P. acanthamoebae likely participating to the quinolones resistance phenotype [7,8]. Ser83 substitution of GyrA is frequently observed and may induce a high level of resistance to quinolones [13]. In the ParC QRDR, we identified two amino acid substitutions, Ser80 / Ala and Glu84 / Asp. These two substitutions are also observed in the ParC QRDR of W. chondrophila [8]. Substitutions of Ser80 and Glu84 are frequently observed in fluoroquinolone resistant E. coli and Streptococcus pneumoniae [13]. Substitutions in the GyrA QRDR and ParC QRDR are thus likely responsible for fluoroquinolone resistance phenotype of E. lausannensis. Resistance of E. lausannensis to the b-lactam antibiotics is probably due to the presence of several b-lactamase genes in the genome (data not shown). Nevertheless, a high concentration of Penicillin G inhibits division leading to aberrant bodies. These aberrant bodies were initially reported for Chlamydia trachomatis and were considered as a persistence form [14,15]. Such aberrant bodies were also reported following treatment of W. chondrophila and Chlamydia pneumoniae with penicillin derivatives demonstrating the major role of peptidoglycan in chlamydial division [16]. Concerning azithromycin, we determined a MIC of 2 mg/ml for E. lausannensis. This is more than 10 fold higher than the MIC for C. trachomatis, C. pneumoniae and W. chondrophila (Table 1). Thus, E. lausannensis should be considered as resistant to azithromycin. The resistance to macrolides is often associated with a mutation in the peptdyl transferase domain of the 23S ribosomal RNA for pathogens that possess a low number of rRNA operons. Several mutations were already reported as conferring resistance to macrolides, such as
M. de Barsy et al. / Microbes and Infection 16 (2014) 746e754
753
Fig. 5. Alignment of the 23S rRNA gene. In blue, nucleotide mutations that could confer resistance to macrolides in the beginning (A) of the peptidyl transferase domain and in the end of this domain (B). Genbank accession numbers are indicated in brackets and the alignment was performed using MUSCLE program. The mutation that may explain the slight resistance of E. lausannensis to macrolide is highlighted in red (E. coli numbering). In green, the 2 base-paired nucleotides involved in the hairpin.
mutations at positions 2057, 2058, 2059, 2611 [17]. Mutations at position 2057 or 2611 only confer a low-level of resistance to macrolides compared to mutations at position 2058 or 2059 [17]. It is thought that these mutations perturb the structure of the binding pocket of the drug leading to the inability to interact and to inhibit the ribosome activity. Chlamydiaceae members have one or two copies of the rRNA operon, depending on the species (Table 2). This could give rise to the emergence of macrolide resistant bacteria with mutations in the 23S rRNA gene. Recently, several spontaneous resistant mutants of Chlamydia psittaci strain 6BC, C. trachomatis strain L2/434/Bu and several Chlamydia caviae isolates, were isolated in vitro after treatment with azithromycin [18,19]. Interestingly, C. psittaci and C. caviae, which possess only one 23S rRNA gene copy, harbored a single mutation in the 23S rRNA gene, i.e. A2058C or A2059C or A2059G. For C. trachomatis strain L2/434/Bu, which possesses two copies of the 23S rRNA gene, no mutation was found in any of these two gene copies. Nevertheless, a Gln to Lys substitution at position 66, was identified in the L4 ribosomal protein, which probably confers resistance to azithromycin. E. lausannensis possesses three copies of the 23S rRNA gene. Multiple alignments allowed us to identify A2059C mutation in all the three copies of E. lausannensis 23S rRNA gene, explaining the observed resistance to azithromycin. Interestingly, E. lausannensis is the only bacterium among all Chlamydia-related bacteria sequenced so far, which possesses this mutation. It could be interesting to test the azitromycin susceptibility of Criblamydia sequanensis, the other member of the Criblamydiaceae family, which do not exhibit this mutation (Fig. 5).
Based on this in vitro study, doxycycline should be considered as the treatment for E. lausannensis infection. It will be also important to assess the activity of this antibiotic against E. lausannensis in vivo. Our work also showed that the propidium iodide assay is useful to indirectly assess the growth of bacteria that generally induce host cell lysis, such as most Chlamydia-related bacteria. The main advantages of the propidium iodide assay over qPCR are that (1) it can be performed in 96-well plates, (2) it does not need to extract genomic DNA and (3) we can follow the mortality in real time. Thus, propidium iodide is less time consuming and represents a suitable screening approach that may be applied to a large amount of compounds and that may help identifying new antibiotics efficient on E. lausannensis. Table 2 Number of rRNA operon in different Chlamydiales bacterial strains compared to E. coli. Number of rRNA operon Escherichia coli strain K-12 Chlamydia trachomatis strain L2/434/Bu Chlamydia pneumoniae strain CWL029 Chlamydia caviae strain GPIC Simkania negevensis strain Z Waddlia chondrophila strain WSU 86-1044 Parachlamydia acanthamoebae strain UV-7 Parachlamydia acanthamoebae strain Hall's coccus Protochlaymdia neagleriophila strain KNIC Protochlamydia amoebophila strain UWE25 Criblamydia sequanensis strain CRIB18 Estrella lausannensis strain CRIB30
7 2 1 1 1 2 3 1 4 3 1 3
754
M. de Barsy et al. / Microbes and Infection 16 (2014) 746e754
Conflict of interest The authors have no conflict of interest. We thank S. Aeby for technical help. We also thank the Electron Microscopy Facility platform of the University of Lausanne as well as the PFMU at the Medical Faculty of Geneva, for assisting with the electron microscopy analyses. We also thank the Cellular Imaging Facility platform of the University of Lausanne for confocal microscopy. Acknowledgment The group of Prof. G. Greub is supported by the Swiss National Science Foundation grants n 310030-141050 and the SNSF Sinergia grant n CRSII3-141837. References [1] Lienard J, Croxatto A, Prod'hom G, Greub G. Estrella lausannensis, a new star in the Chlamydiales order. Microbes Infect 2011;13:1232e41. [2] Thomas V, Casson N, Greub G. Criblamydia sequanensis, a new intracellular Chlamydiales isolated from Seine river water using amoebal coculture. Environ Microbiol 2006;8:2125e35. [3] Greub G, Raoult D. Crescent bodies of Parachlamydia acanthamoeba and its life cycle within Acanthamoeba polyphaga: an electron micrograph study. Appl Environ Microbiol 2002;68:3076e84. [4] Moulder JW. Interaction of chlamydiae and host cells in vitro. Microbiol Rev 1991;55:143e90. [5] Rusconi B, Lienard J, Aeby S, Croxatto A, Bertelli C, Greub G. Crescent and star shapes of members of the Chlamydiales order: impact of fixative methods. Ant Van Leeuwenhoek 2013;104:521e32. [6] Kebbi-Beghdadi C, Batista C, Greub G. Permissivity of fish cell lines to three Chlamydia-related bacteria: Waddlia chondrophila, Estrella lausannensis and Parachlamydia acanthamoebae. FEMS Immunol Med Microbiol 2011;63:339e45. [7] Casson N, Greub G. Resistance of different Chlamydia-like organisms to quinolones and mutations in the quinoline resistance-determining region of the DNA gyrase A- and topoisomerase-encoding genes. Int J Antimicrob Agents 2006;27:541e4. [8] Goy G, Greub G. Antibiotic susceptibility of Waddlia chondrophila in Acanthamoeba castellanii amoebae. Antimicrob Agents Chemother 2009;53:2663e6. [9] Hammerschlag MR, Qumei KK, Roblin PM. In vitro activities of azithromycin, clarithromycin, L-ofloxacin, and other antibiotics against Chlamydia pneumoniae. Antimicrob Agents Chemother 1992;36:1573e4. [10] Smelov V, Perekalina T, Gorelov A, Smelova N, Artemenko N, Norman L. In vitro activity of fluoroquinolones, azithromycin and doxycycline against Chlamydia trachomatis cultured from men with chronic lower urinary tract symptoms. Eur Urol 2004;46:647e50.
[11] Casson N, Posfay-Barbe KM, Gervaix A, Greub G. New diagnostic realtime PCR for specific detection of Parachlamydia acanthamoebae DNA in clinical samples. J Clin Microbiol 2008;46:1491e3. [12] Maurin M, Bryskier A, Raoult D. Antibiotic susceptibilities of Parachlamydia acanthamoeba in amoebae. Antimicrob Agents Chemother 2002;46:3065e7. [13] Nakamura Shinichi. Mechanisms of quinolone resistance. J Infect Chemother 1997;3:128e38. [14] Coles AM, Reynolds DJ, Harper A, Devitt A, Pearce JH. Lownutrient induction of abnormal chlamydial development: a novel component of chlamydial pathogenesis? FEMS Microbiol Lett 1993;106:193e200. [15] Raulston JE. Response of Chlamydia trachomatis serovar E to iron restriction in vitro and evidence for iron-regulated chlamydial proteins. Infect Immun 1997;65:4539e47. [16] Jacquier N, Frandi A, Pillonel T, Viollier P, Greub G. Cell wall precursors are required to organize the chlamydial division septum. Nat Commun 2014;5:3578. [17] Vester B, Douthwaite S. Macrolide resistance conferred by base substitutions in 23S rRNA. Antimicrob Agents Chemother 2001;45:1e12. [18] Binet R, Bowlin AK, Maurelli AT, Rank RG. Impact of azithromycin resistance mutations on the virulence and fitness of Chlamydia caviae in guinea pigs. Antimicrob Agents Chemother 2010;54:1094e101. [19] Binet R, Maurelli AT. Frequency of development and associated physiological cost of azithromycin resistance in Chlamydia psittaci 6BC and C. trachomatis L2. Antimicrob Agents Chemother 2007;51:4267e75. [20] Dreses-Werringloer U, Padubrin I, Zeidler H, Kohler L. Effects of azithromycin and rifampin on Chlamydia trachomatis infection in vitro. Antimicrob Agents Chemother 2001;45:3001e8. [21] Hammerschlag MR, Gleyzer A. In vitro activity of a group of broadspectrum cephalosporins and other beta-lactam antibiotics against Chlamydia trachomatis. Antimicrob Agents Chemother 1983;23:493e4. [22] Samra Z, Rosenberg S, Soffer Y, Dan M. In vitro susceptibility of recent clinical isolates of Chlamydia trachomatis to macrolides and tetracyclines. Diagn Microbiol Infect Dis 2001;39:177e9. [23] Smelov V, Krylova T, Smelova N, Norman L. Azithromycin treatment follow-up: antibacterial susceptibility of Chlamydia trachomatis in patients with chronic prostatitis. Int J Antimicrob Agents 2004;23(Suppl. 1):S79e82. [24] Nilius AM, Shen LL, Hensey-Rudloff D, Almer LS, Beyer JM, Balli DJ, et al. In vitro antibacterial potency and spectrum of ABT-492, a new fluoroquinolone. Antimicrob Agents Chemother 2003;47:3260e9. [25] Miyashita N, Niki Y, Kishimoto T, Nakajima M, Matsushima T. In vitro and in vivo activities of AM-1155, a new fluoroquinolone, against Chlamydia spp. Antimicrob Agents Chemother 1997;41:1331e4. [26] Senn L, Hammerschlag MR, Greub G. Therapeutic approaches to Chlamydia infections. Expert Opin Pharmacother 2005;6:2281e90. [27] Chirgwin K, Roblin PM, Hammerschlag MR. In vitro susceptibilities of Chlamydia pneumoniae (Chlamydia sp. strain TWAR). Antimicrob Agents Chemother 1989;33:1634e5. [28] Kuo CC, Grayston JT. In vitro drug susceptibility of Chlamydia sp. strain TWAR. Antimicrob Agents Chemother 1988;32:257e8.
