Preparation of reference stocks suitable for evaluation of alternative NAT-based mycoplasma detection methods
Center for Biologics Evaluation and Research, Food and Drug Administration, Rockville, MD, USA
Alena Dabrazhynetskaya, Laboratory of Methods Development, Division of Viral Products, Office of Vaccine Research and Review, Center for Biologics Evaluation and Research, US Food and Drug Administration, HFM-470, 1401 Rockville Pike, Rockville, MD 20895, USA. E-mail: firstname.lastname@example.org
The aim of this study was to optimize conditions for preparation and cryopreservation of mycoplasma reference materials suitable to evaluate alternative nucleic acid testing (NAT)-based assays and to compare their limits of detection (LODs) with those of conventional culture-based methods.
Methods and Results
Acholeplasma laidlawii, Mycoplasma gallisepticum and Mycoplasma arginini stocks with low ratios of genomic copies to colony forming units (12, 8 and 4, respectively) harvested in early stationary phases of growth were preserved with different cryoprotective agents (CPAs) under slow (1°C min−1), moderate (8°C min−1), fast (13°C min−1) and ‘snapshot’ (60°C min−1) cooling rates. Depending on mycoplasma species, increasing the cooling rate from slow to snapshot enhanced cell survival up to 5-fold. The addition of 10% (v/v) dimethyl sulfoxide (DMSO) and 15% (v/v) glycerol significantly improved cell survival of all tested strains. Cryoprotected stocks maintained high and stable titres for at least 1 year during storage at −80°C. Sonication of cell cultures prior to cryopreservation enhanced cell dispersion and reduced of GC/CFU ratios.
It is feasible to prepare stable reference stocks of cryopreserved mycoplasma cells suitable to reliably compare NAT- and culture-based mycoplasma testing methods.
Significance and Impact of the Study
This study describes experimental results demonstrating the preparation and storage of highly viable and dispersed mycoplasma reference stocks suitable for comparing alternative NAT-and conventional culture-based mycoplasma detection methods.
Efficient early detection of mycoplasma contamination remains a challenging problem during the development and manufacture of cell-derived biologics and other pharmaceutical products (Razin et al. 1998; Salman et al. 1998; Drexler and Uphoff 2002; Uphoff et al. 2002). The culture-based methods established several decades ago, and still employed to detect contamination with mycoplasma, are laborious, cumbersome and time-consuming (Eldering et al. 2004; Deutschmann et al. 2010; Lawrence et al. 2010). Consequently, there is a high demand for advanced alternative methods, which are able to rapidly detect low levels of mycoplasma contamination in samples collected during stages of manufacturing that require immediate decisions regarding further processing. Several innovative technologies described recently (McGarrity and Kotani 1986; Mattsson and Johansson 1993; Uphoff and Drexler 2004; Baczynska et al. 2005; Sung et al. 2006; Mariotti et al. 2008) have been developed into alternative mycoplasma testing methods that might replace culture-based methods. Currently, there are several commercially available mycoplasma detection kits that rely on enzyme immunoassay (Roche Diagnostics GmbH, Penzberg, Germany, bioluminescence-based technology (MycoAlert™, Lonza, Walkersville, MD, USA), PCR (PromoCell GmbH, Heidelberg, Germany; Venor™GeM and LookOut®, Sigma-Aldrich, St Louis, MO, USA; Takara/Westburg BV, Leusden, Netherlands; PromoKine GmbH, Heidelberg, Germany; MycoTOOL®, Lonza/Roche, Walkersville, MD, USA), qPCR (PromoCell GmbH, Heidelberg, Germany; LookOut®, Sigma; MycoSEQ®, Life Technologies, Carlsbad, CA, USA), transcription-mediated amplification (MilliProbe®, EMD Millipore Corp., Billerica, MS, USA) or fluorescent probe hybridization (MycoProbe®, R&D Systems, Inc., Minneapolis, MN, USA). Among those methodologies, nucleic acid technique (NAT)-based methods appear to be the most advanced and promising for rapid and sensitive detection of mycoplasmas. However, prior to incorporation into mycoplasma testing protocols and implementation, any useful alternative method must demonstrate a limit of detection (LOD) comparable with those of conventional culture-based methods. It is noteworthy that comparability studies face serious technical challenges when the alternative and conventional methods measure different biological features of mycoplasmas, resulting in dissimilar read-outs difficult to compare directly. Thus, our previous attempts to compare NAT-based methods, which detect the presence of mycoplasma-specific nucleic acids (either genomic DNA or cellular RNA) regardless of cell viability, and culture-based methods, which detect only viable cells, led us to understand that unbiased comparisons require special reference materials with a high percentage of viable cells and a low degree of aggregation (Volokhov et al. 2011). From this standpoint, the ratio between genomic copies (GC) and colony forming units (CFU) represents a valuable parameter to assess both the viability of bacterial cells and their level of aggregation in cultures (Dabrazhynetskaya et al. 2011). In practice, the GC/CFU ratio varies over a wide range, being affected by many factors, for example, innate features of specific mycoplasma strains, culture/incubation conditions, growth phase at which samples are collected, conditions of freezing/thawing and storage, etc. (Razin 1969; Addey et al. 1970; Raccach et al. 1975; Biddle et al. 2004; Cheng et al. 2007; Boonyayatra et al. 2010). For that reason, comparability studies can yield compromised results leading to inadvertent overestimation of the LOD of NAT-based methods when mycoplasma reference materials used in the studies were prepared inappropriately. To avoid this problem, the mycoplasma stocks used for comparability assays should have the lowest possible GC/CFU ratios, reflecting high cell viability and dispersion.
We present here our experimental data from attempts to optimize conditions for preparation and cryopreservation of mycoplasma reference materials suitable to compare LODs of conventional culture-based and alternative NAT-based mycoplasma detection methods.
Materials and methods
All strains used in this study, including Acholeplasma laidlawii PG8 (ATCC 23206), A. laidlawii Laidlaw A (ATCC 14089), Myc. arginini G230 (ATCC 23838) and Myc. gallisepticum PG 31 (ATCC 19610), were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA). Depending on the species, mycoplasma cultures were grown using ATCC 243 medium supplemented with either 0·5% (w/v) d-glucose (Fisher Scientific, Fair Lawn, NJ, USA) or 0·5% (w/v) l-arginine (Sigma-Aldrich, Inc., St Louis, MO, USA). Cells were incubated either aerobically at 37± 1°C in the presence of 5% CO2 or anaerobically using the GazPak EZ pouch system (BD Bioscience, Franklin Lakes, NJ, USA; see Table 1 for details). Culture growth was monitored by measuring light absorption at 600 nm using an UltraSpec3100 pro spectrophotometer (Amersham Biosciences, Cambridge, UK) as described previously (Dabrazhynetskaya et al. 2011). Cell titres were determined by plating 10-fold serial dilutions of culture samples onto solid media of appropriate composition.
Table 1. Incubation conditions and media used to cultivate mycoplasma strains
Acholeplasma laidlawii PG8
ATCC 243 + d-glucose
36 ± 1°C; aerobic
Acholeplasma laidlawii Laidlaw A
ATCC 243 + d-glucose
36 ± 1°C; aerobic
Mycoplasma arginini G230
ATCC 243 + l-arginine
36 ± 1°C; aerobic and anaerobic
Mycoplasma gallisepticum PG 31
ATCC 243 + d-glucose
36 ± 1°C; aerobic
Mycoplasma cells were grown as described above to achieve early stationary phase. Samples were diluted to the appropriate cell density in 1 × PBS and stained with the BacLight Green bacterial dye according to the manufacturer's recommendations (Molecular Probes, Inc., Eugene, OR, USA). Stained cells were washed with 1 × PBS to remove the excess of fluorescent dye, pelleted by centrifugation for 10 min at 8 000 g and finally resuspended in 1 × PBS. Unfixed cell samples were placed onto glass slides and analysed using a Nikon Eclipse 50i fluorescence microscope with Nikon Digital camera DXM 1200F (Nikon InsTech Co., LTD, Kawasaki, Japan) using the filter and magnification recommended by the manufacturer for cells stained with the fluorescent dye. Images were captured and analysed using ACT-1 software (Nikon, Nikon InsTech Co., Ltd, Tokyo, Japan) and represented results of analysis from 10 to 15 fields from three independent experiments.
Sonication of cell cultures
Fifteen-millilitre samples of A. laidlawii PG8 cell cultures in sterile polypropylene tubes (Corning, Inc., Corning, NY, USA) were placed into an Array It ultrasonic water bath (Crest Ultrasonics, Cortland, NY, USA) and sonicated for 10 min at 50 kHz. The viability of cell culture before and after sonication was assessed using a standard agar plating method.
Cryopreservation of cell cultures using different cooling rates
Selected mycoplasma strains (A. laidlawii Laidlaw A, Myc. gallisepticum PG 31 and Myc. arginini G230) were grown as described above and supplemented with either 10% (v/v) DMSO, 15% (v/v) glycerol or 30% (v/v) glycerol. After adding the CPA, suspensions were incubated for 15 min at room temperature to allow protectant to penetrate the mycoplasma cells. One-millilitre aliquots were then dispensed into polypropylene cryogenic vials (Corning Inc., Corning, NY, USA) and used to assess the effect of different cooling rates on survival of mycoplasma cells. All experiments were performed in triplicates to assess the reproducibility of results.
To test different cooling rates, vials containing the cultures were placed into one of the following: (i) −80°C freezer in a special freezing container (‘Mr. Frosty’, Nalgene Nunc Inc., Rochester, NY, USA) that provided a slow cooling rate (c. 1°C min−1); (ii) −80°C freezer in an open rack (intermediate cooling rate; c. 8°C min−1), (iii) fine-pelleted dry ice (fast rate; c. 13°C min−1) or (iv) liquid nitrogen (snapshot rate; c. 60°C min−1). All cooling rates were confirmed using separate vials containing 1 ml of water in three independent experiments and calculated from the average times required to achieve a complete freezing of sample in cryovials.
Stability of frozen stocks
Cultures of A. laidlawii, Myc. gallisepticum and Myc. arginini frozen at the intermediate cooling rate in the presence of different cryoprotectants were stored at −80°C. To assess the stability of frozen stocks, cell titres were measured on the day of freezing and 1, 6 and 12 months later. Stability of frozen stocks at different time points was assessed as a per cent from the initial titre observed at 0 time point. All experiments were conducted for three mycoplasma strains in triplicates taken at each time point. All frozen samples were reconstituted by thawing for 15 min in air at room temperature.
Optimizing recovery of Myc. arginini G230 cells after cryopreservation
Mycoplasma arginini G230 cells were collected during late exponential or early stationary growth phases. After adding 15% glycerol, a stock was equilibrated for 15 min at room temperature, and 1-ml samples were dispensed into cryopreservation vials, frozen under the intermediate cooling rate as described above and stored at −80°C. The frozen samples were reconstituted by thawing for 15 min in air at room temperature, plated onto agar plates and incubated under aerobic and anaerobic conditions. Colonies were counted on the 7th day of incubation. The viability of cell cultures was assessed before and after freezing using a standard agar plating method performed in triplicates to ensure the accuracy of the titre determination.
Isolation of mycoplasma genomic DNA and GC assessment
Genomic DNA of A. laidlawii, Myc. gallisepticum and Myc. arginini was isolated using a standard protocol described previously (Dabrazhynetskaya et al. 2011). GC numbers were assessed using quantitative PCR performed under standard conditions using universal mycoplasma primers and species-specific mycoplasma standard DNAs as described earlier (Dabrazhynetskaya et al. 2011).
Preparation of mycoplasma reference stocks
Acholeplasma laidlawii, Myc. gallisepticum and Myc. arginini cultured as described in materials and methods were used to prepare frozen reference materials. The GC/CFU ratios of A. laidlawii Laidlaw A, Myc. gallisepticum PG 31 and Myc. arginini G230 before preservation were 12, 8 and 4, respectively (Table 2), which reflected high viability and low levels of cell aggregation of prepared stocks.
Table 2. GC/CFU of mycoplasma reference stocks before cryopreservation
Acholeplasma laidlawii Laidlaw A
7·50E + 08
9·00E + 09
Mycoplasma gallisepticum PG 31
4·20E + 08
3·50E + 09
Mycoplasma arginini G230
3·30E + 08
1·30E + 09
Control of cell aggregation in mycoplasma reference materials
While high viability of mycoplasma cells can be achieved by cultures harvested during the exponential–early stationary growth phases (Dabrazhynetskaya et al. 2011), the assessment of cell aggregation in collected reference stocks is more challenging. In this study, we investigated the difference in cell aggregation between two strains of A. laidlawii, PG8 and Laidlaw A, which, when collected during similar growth phase, always revealed significantly different GC/CFU ratios. Thus, the GC/CFU ratios generally observed with A. laidlawii PG8 cultures were always above 50, while those with A. laidlawii Laidlaw A cultures never exceeded 10 (A. Dabrazhynetskaya and V. Chizhikov, unpublished data).
We evaluated cell aggregation of both strains, first by visual examination of the colonies formed on surfaces of agar plates and then by fluorescence microscopy (materials and methods). This analysis showed that Laidlaw A always formed colonies of uniform size (Fig. 1a, right), whereas PG8 formed colonies that varied significantly in size (Fig. 1a, left). This observation suggested that PG8 might form heterogeneous aggregates containing different numbers of cells. To verify this assumption, we stained living cells from cultures of both strains with BacLight Green dye and then evaluated the sizes of cell aggregates by fluorescence microscopy. The fluorescent images confirmed that PG8 cell cultures contained large cell aggregates (Fig. 1b, left), whereas cells from Laidlaw A cultures were significantly less aggregated and more homogeneous (Fig. 1b, right). It is interesting that cell aggregates formed by A. laidlawii PG8 demonstrated the substantial stability to mechanical disruption. We did not observe any effect of vigorous vortexing or pipetting on the resulting titre of PG8 cell culture; however, these aggregates could be efficiently disrupted by sonication. Thus, a 10-min exposure of 15 ml of a PG8 cell suspension to 50 kHz ultrasound (see materials and methods) resulted in a 50-fold increase in cell titre with a proportionate reduction in GC/CFU ratio (Table 3).
Table 3. Acholeplasma laidlawii PG8 cell titres and GC/CFU ratios before and after sonication
Acholeplasma laidlawii PG8
1·6E + 08
9·3E + 09
7·6E + 09
9·3E + 09
Cryopreservation of mycoplasma reference materials
We explored the effects of different CPAs and cooling rates had on survival of mycoplasma cells during cryopreservation. Acholeplasma laidlawii, Myc. gallisepticum and Myc. arginini, grown nearly to stationary phase, were supplemented by adding CPAs to different final concentrations, that is, 10% DMSO, 15% glycerol or 30% glycerol, and then frozen using slow (1°C min−1), moderate (8°C min−1), fast (13°C min−1) or snapshot (60°C min−1) cooling rates. The effect of each CPA was assessed by comparing cell titres before and after freezing with different cooling rates.
The results demonstrated that survival of frozen mycoplasma cells largely depended on the cooling rate (Fig. 2a–c). In general, an increased cooling rate enhanced the survival of both CPA-protected and nonprotected cells. For example, the titre of nonprotected Myc. arginini cells frozen at the snapshot cooling rate almost equalled the initial titre before freezing and was approximately four times higher than that of the same cultures frozen at the slow cooling rate (Fig. 2c). Acholeplasma laidlawii and Myc. gallisepticum cells were less sensitive to cooling rates; nevertheless, they also demonstrated about 1·6-fold higher titres after freezing at the snapshot rate (Fig. 2a–b).
The effects of CPAs on survival of mycoplasma cells were quite different. The addition of DMSO or glycerol in final concentrations of 10 and 15%, respectively, improved the survival of all three strains, regardless of the applied cooling rate (Fig. 2a–c). The most significant protective effect of these CPAs was observed with Myc. gallisepticum and Myc. arginini, while their effect on recovery of A. laidlawii was minor. Surprisingly, we observed a dramatic reduction in cell viability when cultures were frozen with 30% glycerol at any cooling rate; titres of A. laidlawii, Myc. gallisepticum and Myc. arginini cells frozen with 30% glycerol were even lower than those of noncryoprotected cell cultures (Fig. 2a–c).
Some insignificant titre increase observed after freezing of cultures at high cooling rates might be attributed to partial disaggregation of cell clumps (aggregates) during freezing/thawing procedure (Fig. 2a–c)). Thus, Myc. gallisepticum and Myc. arginini stocks frozen at the snapshot cooling rate with 15% glycerol demonstrated 2- and 1·5-fold higher titres, respectively, in comparison with their initial titres (Fig. 2b,c).
Stability of mycoplasma stocks stored frozen at −80°C
To assess the stability of A. laidlawii, Myc. gallisepticum and Myc. arginini reference stocks frozen with and without CPAs, we monitored changes in cell titre during storage at −80°C. Generally, both CPA-protected and nonprotected frozen stocks demonstrated stable titres for at least 1 year in storage (Fig. 3a–c). However, the titre profiles obtained for the mycoplasma stocks preserved with different CPAs were species dependent. Thus, A. laidlawii and Myc. arginini stocks showed some reduction in titre (about 20%) during a year in storage with a sleeper slope during the first month after freezing (Fig. 3a,c). In contrast, Myc. gallisepticum titres remained stable during the entire period of monitoring (Fig. 3b). Interestingly, the relative titres of control samples of both A. laidlawii and Myc. gallisepticum frozen without CPAs were quite stable up to 12 months in storage, comparable or even superior to samples protected with either 10% DMSO or 15% glycerol (Fig. 3a–c).
Optimization of conditions for efficient recovery of Myc. arginini G230 after freezing
Recovery of Myc. arginini G230 cells after freezing depended dramatically on the environment in which the agar plates were incubated. A recovery study was conducted using Myc. arginini G230 stock cryoprotected using 15% glycerol and frozen with the intermediate cooling rate as described in 'Materials and methods'. Cells cultured under anaerobic conditions demonstrated a high rate of recovery, that is, about 99·5% compared with titres determined before freezing (Fig. 4). When cells were cultured under aerobic conditions, however, titres were reduced by at least four logs when compared with initial titres (Fig. 4).
This study was conducted to optimize conditions for the preparation of frozen mycoplasma reference materials suitable for evaluating alternative NAT-based mycoplasma detection methods and their comparison with conventional culture-based methods currently recommended for mycoplasma testing by regulatory agencies. Comparison of such methods, which rely on measuring naturally different biological features (i.e. GC vs CFU) requires specially prepared and well-characterized reference materials. As NAT-based methods detect mycoplasmal DNA in a sample, regardless of cell viability, the results of any comparison depend, to a great extent, on the GC/CFU ratio in the reference material used. The GC/CFU ratio is an important biological parameter that reflects the viability and aggregation of cells in a culture, both of which should be accurately assessed before selecting a reference stock to compare methods, especially, when the stock has been frozen (Dabrazhynetskaya et al. 2011; Volokhov et al. 2011). As many bacteria, including mycoplasmas, have a tendency to aggregate (Maniloff and Morowitz 1972; Rottem 2003; Spiglazov et al. 2004; Voloshin and Kaprel'iants 2005), the monitoring of cell aggregation and application of appropriate approaches to reduce aggregation are important for the proper preparation of reference materials suitable to compare methods.
In this study, we found that aggregation of mycoplasma cells varied significantly from strain to strain, even within a single mycoplasma species. Thus, A. laidlawii strain PG8 formed large and stable cell clumps during growth in a conventional mycoplasma broth, whereas cultures of A. laidlawii strain Laidlaw A were significantly less aggregated and more dispersed when grown under the same conditions (Fig. 1). As result, the GC/CFU ratios of PG8 and Laidlaw A cultures grown for 24 h differed significantly (58 and 12, respectively; Tables 2 and 3). Both visual examination of the morphology of colonies formed on surfaces of agar plates and fluorescence microscopy of stained living cells confirmed the presence of large cell aggregates in PG8 cultures, while Laidlaw A cultures showed much less cell aggregation (Fig. 1). The cell aggregates formed in PG8 cultures were stable after vigorous vortexing; however, the same aggregates could be efficiently disrupted by sonication without any loss of cell viability (Table 3). As result, the GC/CFU ratio in a PG8 culture dropped from 58 to 1 after sonication, indicating that the cells became monodispersed. Therefore, sonicated cultures of those strains having an innate propensity to aggregate may yield highly viable and monodispersed cell cultures well suited to serve as reference materials for comparison of mycoplasma detection methods.
The lack of cell walls and the specific lipid composition of the cell membranes make mycoplasma cells very susceptible to damage by freezing (McElhaney et al. 1973; Rottem et al. 1973; Raccach et al. 1975; Mazur 1984; McElhaney 1984). The use of optimal cooling rates and CPAs helps to establish a proper osmotic balance inside mycoplasma cells and to minimize the detrimental effects of freezing (Mazur et al. 1984; Hubalek 2003). We evaluated the survival of cells in cultures of A. laidlawii, Myc. gallisepticum and Myc. arginini frozen under different cooling rates in the presence or absence of different CPAs. In general, all tested mycoplasmas demonstrated the dependence of cell viability on the cooling rate. Thus, depending on mycoplasma species, an increase in the cooling rate from slow (1°C min−1) to rapid snapshot (60°C min−1) resulted in a 5-fold increase in cell survival (Fig. 2). In contrast to A. laidlawii, both Myc. gallisepticum and Myc. arginini stocks were more susceptible to the damage caused by freezing (Fig. 2). This observation can be explained by a specific lipid composition of the A. laidlawii cell membrane (Raccach et al. 1975; Lazarev et al. 2011) that increases the permeability of the cells and results in better survival (McElhaney et al. 1973; Mazur et al. 1984).
We observed 10% DMSO and 15% glycerol as CPAs that improved the survival of all mycoplasmas tested (Fig. 2a–c). Although the choice of an optimal CPA to preserve mycoplasma cell cultures seems to depend on innate features of individual species (or even strains), both 10% DMSO and 15% glycerol are satisfactory alternatives for efficient cryopreservation. Surprisingly, 30% glycerol had a toxic effect on mycoplasma cells and, thus, cannot be recommended for cryopreservation of mycoplasma cultures. The periodic monitoring of cell viability for 1 year after freezing confirmed the high stability of all mycoplasma stocks protected with either 10% DMSO or 15% glycerol (Fig. 3a–c). Therefore, addition of CPA at an optimum concentration as well as the use of optimal cooling rates may significantly improve the survival of mycoplasma cells during the preparation and cryopreservation of reference stocks suitable to compare NAT-and culture-based methods.
It is noteworthy that the use of proper warming rates to thaw frozen mycoplasma stocks may also increase cell recovery. Previously, it was demonstrated that thawing frozen mycoplasma cell cultures in air at ambient temperature gave an optimal warming rate that ensured efficient cell recovery (Raccach et al. 1975; Biddle et al. 2004; Boonyayatra et al. 2010). Some frozen mycoplasmas require special conditions for efficient recovery. Our results showed that the efficient recovery of frozen Myc. arginini G230 stocks required using anaerobic conditions for incubating agar plates that allowed us to achieve c. 99·5% efficient recovery of cells. Surprisingly, the use of aerobic conditions for agar plate incubation resulted in a four-log titre loss of Myc. arginini G230 (Fig. 4). The molecular mechanisms of this interesting phenomenon remain unknown. However, as do many other pathogens, Myc. arginini G230 utilizes l-arginine via the arginine deiminase pathway supported by anaerobic metabolism (Broman et al. 1978; Zuniga et al. 2002; Surken et al. 2008). The negative effect of aerobic incubation conditions we observed could be caused by inhibition of ornithine carbamoyltransferase, key enzyme of the arginine deiminase pathway, by molecular oxygen (Broman et al. 1978). However, we also cannot exclude the possibility of catabolic repression, as arginine-requiring Myc. arginini G230 can utilize glucose in an alternative nonfermentative way (Maniloff 1992; Pollack et al. 1997).
We thank Dr. Martha Folmsbee from EMD Millipore and Dr. Kurt Brorson from the Center for Drug Evaluation and Research, FDA, for helpful discussions of technical details of mycoplasma sonication and Dr. David Asher from FDA/CBER for valuable comments during the manuscript preparation.
Our comments are an informal communication and represent our own best judgment. These comments do not bind or obligate FDA.