ABSTRACT
Recently, it was demonstrated using in vitro assays that the avian pathogen Mycoplasma gallisepticum is able to invade nonphagocytic cells. It was also shown that this mycoplasma can survive and multiply intracellularly for at least 48 h and that this cell invasion capacity contributes to the systemic spread of M. gallisepticum from the respiratory tract to the inner organs. Using the gentamicin invasion assay and a differential immunofluorescence technique combined with confocal laser scanning microscopy, we were able to demonstrate in in vitro experiments that M. gallisepticum is also capable of invading sheep and chicken erythrocytes. The frequencies of invasion of three well-defined M. gallisepticum strains were examined over a period of 24 h, and a significant increase in invasiveness occurred after 8 h of infection. In addition, blood samples derived from chickens experimentally infected via the aerosol route with the virulent strain M. gallisepticum Rlow were analyzed. Surprisingly, M. gallisepticum Rlow was detected in the bloodstream of infected chickens by nested PCR, as well as by differential immunofluorescence and interference contrast microscopy that showed that mycoplasmas were not only on the surface but also inside chicken erythrocytes. This finding provides novel insight into the pathomechanism of M. gallisepticum and may have implications for the development of preventive strategies.
Mycoplasmas are cell wall-less prokaryotes that are widespread in nature as parasites or commensals of eukaryotic hosts. Many of them are pathogens of mammals, reptiles, fish, arthropods, and plants (28), causing a wide variety of diseases and having a predilection for the respiratory tract, the genital tract, and joints (30). Among the agents of infection and disease in domestic poultry and wild birds, Mycoplasma gallisepticum is the most important mycoplasma species (18), causing great losses in the poultry industry. Infections often remain asymptomatic, but commercial poultry flocks are required to be M. gallisepticum free, as infected birds become life-long carriers with the means for horizontal and vertical transmission.
The clinical manifestation following infection, which is called chronic respiratory disease in chickens and infectious sinusitis in turkeys, is induced mainly by stress (18). As the infection starts with colonization of the respiratory tract, tracheitis and air sacculitis are the predominant symptoms of a localized infection in chickens. Occasionally, M. gallisepticum infections are also associated with arthritis, salpingitis, conjunctivitis, and fatal encephalopathy (25), indicating that the organism is able to cross the mucosal epithelial barrier and reach distant locations in the chicken. Experimentally, it has been shown that the pathogen is able to spread throughout the body following aerosol infection, as demonstrated by reisolation of M. gallisepticum from the heart, brain, liver, spleen, and kidneys of experimentally infected chickens (25). How this agent manages to convert a local infection into a systemic infection remains unknown.
Until the end of the 1980s, mycoplasmas were considered exclusively extracellular pathogens. This dogma was retracted when in 1989 Lo et al. (20) published the first report of a cell-invasive mycoplasma which was isolated from patients with AIDS and later identified as M. fermentans. In the years after this, the cell-invasive property of M. fermentans was confirmed by other investigators (31, 32), and three other human mycoplasmas, M. genitalium, M. pneumoniae, and M. penetrans, were reported to be similarly invasive for nonphagocytic cells (2, 21, 23, 32, 36). After these first discoveries of the cell-invasive potential of mycoplasmas which are pathogenic for humans, M. gallisepticum, which phylogenetically belongs to the M. pneumoniae cluster, was also described to be cell invasive, as it was shown to invade HeLa cells and chicken embryo fibroblasts in vitro (7, 34). At least for this mycoplasma species it was further shown that the cell-invasive capacity plays an important role in systemic spreading, because the cell-invasive strain Rlow was reisolated from inner organs after aerosol challenge of chickens, whereas the noninvasive strain Rhigh was not (25).
The view that cell invasiveness provides bacterial pathogens with a number of advantages is generally accepted. These advantages include protection from the immune system, reduction of the efficacy of antibiotics during treatment, and nutritional benefits. Moreover, internalization by the eukaryotic host cell may enable the pathogen to pass through cell barriers such as the mucosal epithelium. Of the cell-invasive mycoplasmas, so far only M. fermentans, M. penetrans, and M. genitalium have been found inside cells in vivo. Intracellular M. fermentans and M. penetrans have been visualized in clinical samples or tissue material from AIDS patients using electron microscopy (20, 21), whereas more recently, intracellular M. genitalium has been found in human vaginal samples using confocal immunoanalysis (5). In contrast, no intracellular residence in vivo has been described for M. pneumoniae and M. gallisepticum to date, even though this possibility was implied for M. gallisepticum by the systemic spread of cell culture invasion-positive organisms in the chicken host after experimental infection (25). However, more recently, M. pneumoniae was detected by PCR in the bloodstream of a substantial proportion of patients with mycoplasma pneumonia (10). Since M. gallisepticum and M. pneumoniae are related phylogenetically and have other features in common, including an attachment organelle, homology of adhesins and adhesion-related molecules, gliding motility, and the similarity of the diseases that they cause, this prompted us to investigate the ability of M. gallisepticum to invade red blood cells (RBCs).
In this report, we provide evidence for the first time that M. gallisepticum is able to invade RBCs. Erythrocyte-invasive organisms were detected not only after in vitro infection but also in vivo in blood samples from experimentally infected chickens. These findings indicate that there is an infection strategy that was previously unknown for pathogenic mycoplasmas. By invading the host's RBCs during infection, M. gallisepticum gains access to a perfect transportation system that allows the agent to colonize distant niches while concomitantly being protected from the host's immune system.
MATERIALS AND METHODS
Mycoplasma strains and growth conditions. M. gallisepticum strains Rlow and Rhigh used in this study were kindly provided by S. Levisohn, Kimron Veterinary Institute, Bet Dagan, Israel. Rlow represents the 10th passage of the prototype strain R (19), and Rhigh represents the 164th passage in artificial medium. Vaccine strain 6/85 was kindly provided by Intervet (Intervet Intl., Boxmeer, The Netherlands).
All M. gallisepticum strains were grown in modified Hayflick medium (35) containing 20% (vol/vol) heat-inactivated horse serum (Gibco Products-Invitrogen Ltd., Paisley, United Kingdom). To estimate the numbers of CFU in cultures, serial dilutions were plated on modified Hayflick medium containing 1% (wt/vol) Difco Noble agar (BD, Franklin Lakes, NJ) and incubated at 37°C. CFU were counted 7 to 10 days later using an SMZ-U stereomicroscope (Nikon Corp., Tokyo, Japan).
DNA extraction.DNA extraction from M. gallisepticum cultures was performed by using the phenol extraction method of Bashiruddin (3). The DNA concentration was measured photometrically with a Gene Quant II RNA/DNA calculator (Pharmacia Biotech, Cambridge, United Kingdom). For detection of M. gallisepticum in blood of infected chickens, DNA was extracted from a blood-Alsever's solution mixture with a DNeasy blood and tissue kit (Qiagen, Maryland) by using the manufacturer's protocol. Ten microliters of blood was used for each extraction, and DNA was eluted twice from the column using 100 μl sterile H2O per elution. For subsequent nested PCR analysis 50 μl of eluate was used.
Scanning electron microscopy.Sheep or chicken erythrocytes mixed with M. gallisepticum cells for various periods of time were incubated overnight on poly-l-lysine-coated coverslips to allow binding of the RBCs. The coverslips were washed three times with phosphate-buffered saline (PBS) and then twice with cacodylate buffer (0.1 M sodium cacodylate, pH 7.4) for 10 min. Samples were then fixed in 2.5% glutaraldehyde in cacodylate buffer for 2 h at 4°C and washed three times in cacodylate buffer. After dehydration of the samples with a graded series of ethanol concentrations, the specimens were critical point dried in a Bal-TEC CPD030 (BAL-TEC AG, Balzers, Liechtenstein), and after mounting, they were sputter coated with gold-palladium using a Polaron SC7640 (Quorum Technologies Ltd., Newhaven, United Kingdom). The samples were viewed using a JEOL JSM 5410LV scanning electron microscope (Jeol Ltd., Tokyo, Japan) operated at 10 kV.
Nested PCR.A nested PCR covering the 5′ region of crmA was developed to detect M. gallisepticum in chicken blood. The external primers J3F and J3R were used in a 20-cycle amplification reaction that yielded a 349-bp product, while the internal primers J2F and J2R used in a 30-cycle reaction generated a 288-bp PCR product. The first amplification reaction was performed with a 100-μl (total volume) mixture containing (final concentrations) 3 mM MgCl2, 0.2 mM of each deoxynucleoside triphosphate, 200 nM primer J3F (5′-GCAATTAGTTAATCAAGCAAG-3′), 200 nM primer J3R (5′-ATTACCAATTCTATTTGAGTTAG-3′), and 5 U of GoTaq polymerase (Promega Corp., Madison, WI). The amplification conditions were 94°C for 3 min, followed by 20 cycles of 94°C for 1 min, 55°C for 1 min, and 72°C for 2 min and then a final extension step for 7 min at 72°C. For the nested PCR, 2 μl of the first PCR amplification reaction mixture was used. The reaction was performed with a 25-μl (total volume) mixture containing (final concentrations) 3 mM MgCl2, 0.2 mM of each deoxynucleoside triphosphate, 200 nM primer J2F (5′-GAACGCTAGATGCTAATTCTG-3′), 200 nM primer J2R (5′-GAACGTTAGCTTCATCATTAACC-3′), and 1.5 U of GoTaq polymerase. The amplification conditions were similar to those used for the first PCR but included 30 cycles instead of 20 cycles. Amplification products were detected by electrophoresis of 3 μl of the reaction product in a 1.6% agarose gel containing ethidium bromide and inspection under UV light. Gel pictures were taken with the ChemiDoc XRS gel documentation system (Bio-Rad Laboratories Inc., Hercules, CA).
Gentamicin invasion assay.The number of intracellular M. gallisepticum cells in HeLa-229 cells was analyzed using the gentamicin invasion assay as described elsewhere (34), except for testing the efficacy of gentamicin. Briefly, mycoplasma cultures were centrifuged, washed, and resuspended in Invitrogen minimum essential medium (MEM) to a final density of 3 × 108 to 5 × 108 CFU per ml. Then serial dilutions of gentamicin were added to obtain final concentrations ranging from 25 to 400 μg/ml. After 3 h of incubation at 37°C, aliquots were plated onto agar plates without gentamicin. As a gentamicin concentration of 100 μg/ml was shown to be sufficient to kill all mycoplasmas in the medium, the gentamicin invasion assays were performed with a gentamicin concentration of 400 μg/ml.
For quantification of intraerythrocytic M. gallisepticum, the protocol was adapted as follows. Citrated chicken blood was centrifuged at 1,000 × g for 3 min, and the pellet was washed at least two times in PBS while the buffy coat layer containing white blood cells was removed. The concentration of the remaining RBCs was adjusted to 2 × 108 RBCs/ml with PBS. M. gallisepticum strains Rlow, Rhigh, and 6/85 were grown as described above to mid-exponential phase, as indicated by a color change of the medium due to the metabolic activity of growing mycoplasmas, followed by at least three washes with PBS. During washing, the M. gallisepticum culture was centrifuged at 12,000 × g for 10 min, and after the final centrifugation the pellet was resuspended in Invitrogen MEM containing l-glutamine, Earle's balanced salts, and HEPES supplemented with 5% (vol/vol) Invitrogen tryptose phosphate broth, 0.1 mM nonessential amino acids, and 7.75% (vol/vol) fetal calf serum. Resuspended M. gallisepticum cells were passed through a 23-gauge injection needle at high speed at least 20 times to disperse the cells. For infection of RBCs, M. gallisepticum cultures were diluted to obtain concentrations of approximately 4 × 105 to 8 × 105 CFU/ml and mixed with RBCs to obtain final ratios of erythrocytes to mycoplasmas of 125:1 to 500:1. After 1, 2, 4, 8, and 24 h of infection at 37°C, 1-ml samples were removed and split into two parts. One part was mixed with the same volume of MEM containing gentamicin at a final concentration of 400 μg/ml and incubated for 3 h at 37°C to kill all extracellular mycoplasmas, whereas the other part received the same treatment but without the antibiotic. After incubation, the samples were washed at least three times with PBS and centrifugation at 12,000 × g for 10 min. Finally, appropriate dilutions of both gentamicin-treated and untreated RBC suspensions were plated onto modified Hayflick agar plates to allow intraerythrocytic mycoplasmas to form colonies. The numbers of colonies were determined 7 to 10 days later, and the invasion frequencies were calculated by using the numbers of colonies obtained from dilutions with and without gentamicin treatment.
Differential immunofluorescence assay.The presence of mycoplasmas within RBCs in either the in vitro or in vivo experiments was investigated by a modified version of the double-immunofluorescence (DIF) method described by Heesemann and Laufs (15). An adaptation of this method for use with mycoplasmas and HeLa cells and the method used for generation of polyclonal anti-M. gallisepticum rabbit antibodies have been described elsewhere (34). The DIF method was adapted for use with erythrocytes as follows. Chicken or sheep RBCs were washed and infected with M. gallisepticum cultures by using a procedure similar to the procedure described above for the gentamicin invasion assay. The infected RBCs were gently washed three times with PBS containing 2% (wt/vol) bovine serum albumin, and extracellular mycoplasmas were detected by incubating unpermeabilized cells with rabbit anti-M. gallisepticum hyperimmune serum diluted 1:200 in PBS-bovine serum albumin for 30 min at room temperature and then with 1:2,000-diluted fluorescein isothiocyanate (FITC)-labeled goat anti-rabbit immunoglobulin G (IgG) (Harlan Sera-Lab Ltd., Loughborough, United Kingdom) for 30 min at room temperature. The RBCs were then placed into chambers of an eight-well Lab-Tek II chamber slide (Nalge Nunc International, Rochester, NY) and air dried. The RBCs were fixed and permeabilized with ethanol washes using increasing concentrations (50 to 96% ethanol) to allow intracellular antibody diffusion, followed by two washes with 100% methanol and air drying. In order to stain all extracellular and intracellular mycoplasmas, each chamber was incubated with the same anti-M. gallisepticum hyperimmune serum for 30 min, followed by incubation for 30 min with goat anti-rabbit IgG labeled with Alexa Fluor 405 (Molecular Probes, Invitrogen). In the case of RBCs from experimentally infected chickens, goat anti-rabbit IgG labeled with Alexa Fluor 633 (Molecular Probes, Invitrogen) was used. Finally, the chambers were removed, and the samples were mounted under a glass coverslip in glycerol-PBS (1:1.7, vol/vol) containing 13% (wt/vol) Mowiol (Clariant, Muttenez, Switzerland) and 0.5% (wt/vol) n-propyl gallate (Sigma-Aldrich, St. Louis, MO). Samples were examined with a confocal laser scanning microscope (LSM 510 Meta; Carl Zeiss MicroImaging GmbH, Jena, Germany) using argon (488 nm), diode (405 nm), and helium-neon (633 nm) lasers for specific excitation of the fluorescent dyes FITC, Alexa Fluor 405, and Alexa Fluor 633. The resulting fluorescence images were superimposed on differential interference contrast (DIC) micrographs for visualization of the RBCs.
Animal experiments.One-day-old Ross 308 chickens originating from a commercial flock free of M. gallisepticum and M. synoviae as monitored by monthly serological tests were selected for the infection experiment, which was performed at the Veterinary Medical Research Institute, Budapest, Hungary, in accordance with the guidelines of the Hungarian law for protection of animal rights. The mycoplasma-free status of the chickens was verified by testing sera with MYGA and MYSY enzyme-linked immunosorbent assay kits (Diagnosticum Zrt., Budapest, Hungary) and by cultivation of trachea swab samples for 14 days in modified Hayflick medium. Ten chickens were selected for the animal experiment and raised under isolated conditions. On day 21, they were placed in a 0.22-m3 aerosol chamber. For experimental infection, 10 ml of a freshly grown culture of M. gallisepticum strain Rlow (5.6 × 107 CFU ml−1) was sprayed with a pressure of 1 atm for 100 s into the chamber, and the birds were exposed for another 20 min. Blood was collected the day before challenge and then on days 6, 12, and 20 postinfection (p.i.) from the wing vein. Blood samples from all the chickens, which were obtained on the same day, were pooled and mixed with an equal amount of Alsever's solution. After thorough mixing, the blood was kept at 4°C or frozen at −20°C until further PCR analysis.
Statistical analysis of the gentamicin invasion assay.The data for the gentamicin invasion assays were calculated by using the means of at least five independent experiments ± standard deviations. The normal distribution of the data was tested with the Kolmogorov-Smirnov test. Invasion frequencies for strains Rlow, Rhigh, and 6/85 at different times were compared by one-way analysis of variance using the statistical analysis program SPSS 14.0 (SPSS Inc., Chicago, IL). A P value of <0.05 was considered significant.
RESULTS
Interaction of M. gallisepticum with erythrocytes visualized by scanning electron microscopy.As a first approach to examine the erythrocyte invasion properties of M. gallisepticum, the morphological details of its interaction with erythrocytes was studied by scanning electron microscopy after incubation of sheep and chicken RBCs with mycoplasma cells for various periods of time. As shown in Fig. 1 for an infected sheep erythrocyte (Fig. 1A) and an infected chicken erythrocyte (Fig. 1B), some of the mycoplasmas appeared to adhere with their tip structure to the RBC surface. This observation was not unexpected, as the tip structure of M. gallisepticum is considered a specialized multifunctional organelle that mediates attachment to the respiratory epithelium of the chicken host during infection (6). Some of the RBCs had a misshapen and twisted morphology (Fig. 1B), which has also been reported by Lam (17) and Razin et al. (29). The most intriguing detail, however, was the imprints on the otherwise smooth surface of selected RBCs, as shown in Fig. 1A. The form of these imprints resembled the pearlike shape of M. gallisepticum cells that might have penetrated the RBC membrane at those points. This finding encouraged us to further investigate whether M. gallisepticum is in fact able to invade RBCs in vitro.
Scanning electron micrographs of sheep (A) and chicken (B) erythrocytes after in vitro infection with a clonal derivative of M. gallisepticum strain Rlow. The arrows indicate mycoplasmas or imprints of mycoplasmas appearing to sink into the erythrocyte surface.
M. gallisepticum invades erythrocytes in vitro.To investigate the presence of intracellular mycoplasmas, a DIF technique was used. DIF staining was used with sheep and chicken erythrocytes incubated for different periods of time with the virulent strain M. gallisepticum Rlow. Confocal laser scanning micrographs of sheep RBCs infected for 24 h revealed the presence of intraerythrocytic M. gallisepticum cells (data not shown). The RBCs contained only one organism per cell, and the proportion of RBCs carrying intraerythrocytic mycoplasmas was very low, estimated to be 1 in 2,000. M. gallisepticum Rlow was able to invade chicken erythrocytes after infection for 24 h (Fig. 2). Superimposition of FITC and Alexa Fluor 405 fluorescence images and the DIC micrographs clearly showed both surface and intracellular foci of fluorescence corresponding to extracellular and intracellular mycoplasmas (Fig. 2E).
Confocal z scan of a chicken RBC infected in vitro with M. gallisepticum strain Rlow after DIF staining. The same area of a confocal microscopic image after DIF staining was analyzed for extracellular M. gallisepticum labeled with FITC (green fluorescence) (A) and for extra- and intracellular M. gallisepticum labeled with Alexa Fluor 405 (red fluorescence) (B). (C to F) Superimposition of the green and red fluorescence images (D to F) and a DIC micrograph (C) resulted in yellow fluorescence, indicating extracellular M. gallisepticum, while the red fluorescent focus indicates an intraerythrocytic mycoplasma cell. When the image was scanned from the top to the bottom (D to F) of the erythrocyte, first an extracellular mycoplasma cell located at the surface (yellow) was visible, which slowly faded out as the cross section layer was moved downward (E and F). At the same time the intracellular mycoplasma cell came to the fore (E and F), showing the difference in the localizations of the two mycoplasma cells.
In parallel studies, when scanning along a vertical z axis was performed and micrographs of each 0.5-μm slice were obtained (Fig. 2D to F), an extracellular mycoplasma located at an erythrocyte's surface (yellow) became visible, which faded out as the cross section layer was moved downward. At the same time, an intracellular mycoplasma (red) came into view, showing the relative difference in the localizations of these two mycoplasma cells. Uninfected RBCs treated with anti-M. gallisepticum antiserum and FITC- and Alexa Fluor 405-conjugated antibodies exhibited no fluorescence (not shown). Interestingly, the intracellular mycoplasmas were located in the cytoplasm or perinuclear region and were never located inside the nucleus of the chicken RBCs.
The gentamicin invasion assay, a method first described by Kihlström (16), was used to determine the percentage of intracellular bacteria in the whole population. Since gentamicin is not able to cross intact eukaryotic cell membranes, it kills only extracellular mycoplasmas when it is added to an M. gallisepticum-infected cell culture. A time course experiment (Fig. 3) was used to compare the numbers of CFU in gentamicin-treated and untreated M. gallisepticum-infected RBC suspensions. After 30 min of infection, M. gallisepticum established intracellular residence and therefore survived the gentamicin treatment (data not shown). Significant differences in cell invasiveness were observed between the hemadsorption-positive strain Rlow and the hemadsorption-negative strains Rhigh and 6/85 (Fig. 3), and strain Rlow was the strain with the highest invasiveness. The mean invasion frequencies for Rlow ranged from 0.13% after the first hour to 1.18% after 24 h, whereas the highest invasion rates for strains Rhigh and 6/85 were 0.22% at 8 h and 0.09% at 4 h, respectively. At all times examined, Rlow exhibited the highest invasion rate, while strain 6/85 had the lowest invasion rate. The reduced invasion rate of 6/85 is statistically significant compared to the rates for Rlow and Rhigh after 8 h (P < 0.05) and for Rlow after 24 h (P < 0.02). Statistically significant differences in invasiveness between Rlow and Rhigh were observed only after 24 h of incubation (P < 0.05). Overall, the invasion frequencies of all three M. gallisepticum strains were drastically lower when RBCs were used instead of the HeLa-229 cell line. Consistent with the invasion rates for Rlow and Rhigh reported by Winner et al. (34), we observed invasion frequencies after 2 h of infection of 5.7 and 1.1%, respectively, whereas vaccine strain 6/85 exhibited a lower invasion rate (0.5%). When these frequencies were compared with the invasion frequencies obtained with RBCs, the invasion rates for Rlow, Rhigh, and 6/85 were 18-, 48-, and 11-fold higher, respectively, when HeLa cells were used.
Frequencies of invasion of chicken RBCs by M. gallisepticum strains Rlow, Rhigh, and 6/85 at different times. The values are the means ± standard deviations of a minimum of five independent gentamicin invasion assays. The asterisks indicate statistically significant differences in RBC invasiveness between Rlow (open bars) and Rhigh (gray bars) or 6/85 (black bars), and the plus sign indicates statistically significant differences between Rhigh and 6/85.
Interestingly, in the time course experiment the invasion frequencies of Rlow did not follow a linear course from 1 to 24 h after infection but increased only slightly during the first 8 h. Between 8 and 24 h the invasion frequency of Rlow increased sixfold, whereas the invasion rate stayed relatively constant for Rhigh and 6/85.
M. gallisepticum invades chicken erythrocytes during in vivo infection.Blood samples taken from chickens experimentally infected with M. gallisepticum Rlow were analyzed for the presence of M. gallisepticum by DIF microscopy and by nested PCR. Successful infection and systemic spread of the pathogen were proven by necropsy, including lesion scoring for typical M. gallisepticum-associated lesions and reisolation of the pathogen from inner organs (data not shown).
When the DIF technique that was used in the in vitro experiments was used for these blood samples, M. gallisepticum cells residing not only on the surface but also inside the chicken RBCs could be detected (Fig. 4). The numbers of mycoplasmas found either inside or on the surface of RBCs in blood samples from experimentally infected chickens were rather low. Only 1 of 500 RBCs from the experimentally infected chickens carried an M. gallisepticum cell either intra- or extracellularly, as estimated from an investigation of multiple samples.
RBCs from an experimentally M. gallisepticum-infected chicken after DIF staining: superimposed image after FITC and Alexa Fluor 633 labeling, showing M. gallisepticum attached to an RBC's surface (yellow) and inside an RBC (red) after experimental in vivo infection. The erythrocytes were visualized by differential interference microscopy.
For highly sensitive detection of M. gallisepticum Rlow in chicken blood, a PCR method based on nested amplification of a sequence in the 5′ region of the crmA gene was developed. The sensitivity of the nested PCR approach was determined with both serial dilutions of genomic DNA of Rlow and chicken blood mixed with dilutions of viable M. gallisepticum. An amount as small as 16 fg of genomic DNA of Rlow was positive with this PCR approach and was calculated to correspond to 14.5 M. gallisepticum genome equivalents. This calculation was based on the previously described genome size of M. gallisepticum strain Rlow, 996,422 bp (27). When viable Rlow was mixed with erythrocytes, the detection limit of the nested PCR approach was 1.7 CFU.
Using this highly sensitive PCR approach, blood samples from experimentally infected chickens were analyzed (Fig. 5). Blood from chickens taken before infection (day 0, negative control) did not result in a PCR amplification product, indicating that no mycoplasmas were present in the blood before infection. When the same sample was mixed with 3.7 × 104 CFU per ml of Rlow (positive control), amplification of the 288-bp fragment was observed, showing that no PCR-inhibiting compounds were present in the blood samples. All samples taken 6, 12, and 20 days p.i. were positive, indicating that M. gallisepticum Rlow was present in the bloodstream of chickens as soon as 6 days p.i. Based on the sensitivity of the nested PCR assay with mycoplasma-spiked erythrocytes (see above), the PCR detection signal was calculated to correspond to at least 680 CFU per ml of chicken blood.
Agarose gel electrophoresis of nested PCR products from blood samples of experimentally infected chickens. Blood samples were taken before experimental infection of chickens (lane 0) and on day 6 p.i. (lanes 1 and 2), day 12 p.i. (lanes 3 to 5), and day 20 p.i. (lanes 6 to 8). A positive control containing M. gallisepticum-spiked chicken blood (lane +) and a PCR negative control (lane −) were included. Lane M contained a molecular size marker (1-kb ladder; Invitrogen).
DISCUSSION
When analyzing the erythrocyte adhesion properties of different M. gallisepticum strains, we obtained scanning electron micrographs of M. gallisepticum-infected RBCs, in which we observed invaginations and grooves containing structures that had the typical pearlike shape of surface-attached M. gallisepticum cells. Changes in the surface of mycoplasma-infected chicken erythrocytes, like the appearance of dimples and grooves, were reported previously by Lam (17), and indentations of sheep erythrocyte membranes after exposure to M. gallisepticum were described by Razin et al. (29). Such indentations have also been described for M. penetrans interactions with eukaryotic cells (21) and for the erythrocyte-invading bacterium Bartonella bacilliformis (4). Lam even observed perforations on the surface of M. gallisepticum-exposed erythrocytes and speculated that M. gallisepticum may penetrate the RBCs (17). In this report we provide evidence that M. gallisepticum indeed is able to invade RBCs, not only in an experimental in vitro system but also under in vivo conditions during the course of experimental infection.
The invasion of erythrocytes by M. gallisepticum was examined in vitro using well-established approaches to identify intracellular bacteria in order not to rely on only a single method. The technical problems inherent in each method might lead to an erroneous judgment concerning the cell invasiveness of a given pathogen if only a single method were used. A common problem of the gentamicin invasion assay is that the number of mycoplasma colonies growing on the agar plates represents the number of infected host cells better than it represents the number of invasive mycoplasmas. Several mycoplasmas simultaneously infecting one RBC or intracellularly multiplying mycoplasmas derived from a single infecting organism might lead to the formation of only one mycoplasma colony on Hayflick agar. The same result can be expected if a single mycoplasma infects one RBC without multiplying. The real invasion frequency of a given M. gallisepticum strain therefore might be different. To minimize the possible error, low multiplicities of infection (range of ratios of erythrocytes to mycoplasmas, 125:1 to 500:1) were used to reduce the possibility of multiple infections of any given RBC. To rule out the effect of intracellular multiplication on the results, the numbers of CFU in treated and untreated M. gallisepticum-RBC suspensions were compared. This method is in contrast to the method of Winner et al. (34), who compared CFU values before and after gentamicin treatment, which in our case gave slightly higher invasion rates (data not shown), an effect that could also have been due to extracellular multiplication of M. gallisepticum during incubation in the untreated control group.
The time course experiment used to determine the in vitro invasion capabilities of the three M. gallisepticum strains included in this study was performed for 24 h. Longer infection periods resulted in false-negative results as the erythrocytes apparently started to lyse after 24 h. Interestingly, the invasion rate of the virulent strain Rlow increased only slightly during the first 8 h but then increased by a factor of nearly 6 during the next 16 h. A possible explanation for this might be that after the first contact with the RBCs, M. gallisepticum responded by producing certain gene products which enabled the pathogen to enter the RBCs more efficiently. Another explanation might be that the small percentage of M. gallisepticum cells that successfully invaded the RBCs in the first few hours multiplied intracellularly and, after escaping the originally invaded RBCs, the progeny invaded previously uninfected RBCs. The ability of M. gallisepticum to multiply inside cells has been described previously for HeLa-229 cells (34).
Mycoplasma species other than M. gallisepticum have also been reported to propagate intracellularly; intracellular propagation has been described for M. penetrans and for the closely related species M. pneumoniae and M. genitalium (2, 8, 23). The apparent advantages of entering eukaryotic cells are protection from the host's immune system and easy access to nutrients. In the case of erythrocyte invasion an additional benefit might be the large amounts of iron found inside the erythrocytes in the form of hemin or other trace metals. The requirement for iron for bacterial growth is a common theme in pathogenicity (13), and hemin is known to support the growth of invasive bacteria like Bartonella quintana (26) and Haemophilus influenzae (1). An effect of hemin on the growth of mycoplasmas has, however, not been described to date.
Another main advantage of entering erythrocytes might be that they are the means to reach new sites of infection. M. gallisepticum causes acute and chronic infections in birds (18), and chronicity of mycoplasma infections is speculated to be due to intracellular persistence of the pathogen (2, 8). M. gallisepticum has been reisolated from different chicken organs, including the brain, after experimental in vivo infection (25). Systemic spread via the bloodstream, however, might explain how the pathogen reaches distant niches after the first colonization of the air sacs by causing a transient bacteremia for at least 20 days p.i., as indicated by the nested PCR experiments. After a certain intraerythrocytic residence time, M. gallisepticum may escape by lysing the erythrocyte with the help of the membrane-bound hemolysin activity reported by Minion and Jarvill-Taylor (24).
The process of cell invasion by mycoplasmas is still poorly understood. Fibronectin-binding proteins like those detected in other facultative intracellular bacteria (11, 12) have been identified in M. penetrans (14), in M. pneumoniae (9), and more recently also in M. gallisepticum (22). However, the mechanism of M. gallisepticum invasion of nonphagocytic eukaryotic cells is not clear at present. In M. pneumoniae it has been speculated that adhesion and invasion are independent features, because cytadherence-positive but invasion-negative strains were detected (2). In the strains used in our experiments, we observed that erythrocyte invasion of M. gallisepticum correlates with hemadsorption, indicating that cytadherence and expression of the hemadsorption-mediating genes gapA and crmA (7, 33) at least are prerequisites for cell invasion.
Although host cell invasion has been reported for a few mycoplasma species, including M. gallisepticum (2, 7, 20, 23, 31, 32, 34, 36), to our knowledge this is the first report that provides evidence that M. gallisepticum invades RBCs in vitro and in vivo, which has not been shown for any mycoplasma species previously. A more thorough investigation of the mechanism underlying erythrocyte invasion will be performed in the near future.
ACKNOWLEDGMENTS
We thank M. Skalicky and C. Binter for assistance with the statistical calculations and C. Neubauer for providing chicken blood for the in vitro assays.
This work was supported in part by grant P16538 (to R.R.) from the Austrian Science Fund and by a pilot project grant (to M.P.S.) from the University of Veterinary Medicine of Vienna through its research focus program. G.V. was financially supported by the Austrian Science Fund and by a doctoral fellowship from the University of Veterinary Medicine of Vienna. A.P. was supported by Mycosafe Diagnostics GmbH.
FOOTNOTES
- Received 27 June 2007.
- Returned for modification 7 August 2007.
- Accepted 15 October 2007.
- Copyright © 2008 American Society for Microbiology
