The effect of the air sampling method on the recovery of Mycoplasma gallisepticum from experimentally produced aerosols

Abstract

Background: A reliable air sampling method is a prerequisite to calculate the inhaled aerosol dose by animals exposed to the aerosol as precise as possible.[Comp]: Set abstract according to the journal style.[/Comp]

Objective: To examine if aerosol collection in a fluid medium (buffered peptone water (BPW) in the impinger) improves detection of viable mycoplasmas. Also the effect of adding Mycoplasma Experience (ME) broth and/or BPW to the aerosol fluid on aerosol titres was assessed.

Methods: Aerosols containing a Mycoplasma gallisepticum field or vaccine strain were simultaneously sampled with gelatin filters and by impinger immediately after ending aerosolization and 25 min later.

Results: Sampling of M. gallisepticum aerosols using the impinger did not yield higher aerosol titres compared to sampling with gelatin filters. Initial loss during generation of the field strain aerosol and the half-life time of viable mycoplasmas in the aerosol were 1.1–2.4 log₁₀ and <4–15 min, respectively. The vaccine strain was more vulnerable compared to its field counterpart. In spite of higher aerosolized doses of the vaccine strain (10^8.0 to 10^8.1 versus 10^7.5 cfu per m³ of air of the field strain), mycoplasmas were not recovered from the aerosols neither by gelatin filter nor by impinger. Therefore, half-life times could not be calculated. Addition of BPW to the aerosol fluid did not clearly improve the recovery of the field strain from the aerosol, while addition of ME broth and BPW did.

Conclusion: Gelatin filters likely due to their relative high moisture content (10–14% wt/wt) are at least as useful as the impinger for the recovery of M. gallisepticum from aerosols, provided exsiccation of the filters is prevented.

Keywords:

• aerosol
• M. gallisepticum
• gelatin filter
• buffered peptone water
• Mycoplasma Experience broth

1. Introduction

Under natural conditions, outside the host mycoplasmas are especially vulnerable, which is attributed to their lack of cell wall. Only in egg yolk, avian mycoplasmas can survive for several weeks. Mycoplasma gallisepticum, an important poultry pathogen, maintained its viability in yolk up to 16, 7 and 20 weeks at temperatures of 6°C, 20°C and 37°C, respectively. At −20°C, survival of at least 72 weeks was recorded (Chandiramani et al. 1966). After artificial contamination with M. gallisepticum, M. synoviae or M. iowae of other materials from the poultry house environment, all mycoplasma species showed the longest survival time on feathers with M. gallisepticum, M. synoviae and M. iowae surviving between 2 and 4 days, 2–3 days and 5 to >6 days, respectively (Christensen et al. 1994).

It is generally accepted that organisms of the Mollicutes class to which mycoplasma species belong, are unstable and die rapidly in liquid media. However, a number of animal mycoplasmas including M. gallisepticum can survive between 1 and 12 weeks in liquid media depending on the medium composition and temperature, but are unstable under dry conditions (on paper discs), especially at room temperatures and higher (survival in general ≤7 days) (Nagatomo et al. 2001).

In previous research (Landman et al. 2004) it was shown that initial loss of M. gallisepticum and M. synoviae in experimentally produced aerosols (difference between dose and concentration immediately after aerosol generation) was significantly higher in comparison with the initial loss of Enterococcus faecalis. Aerosols were sampled with gelatin and cellulose nitrate filters coupled to different air sampling devices. Results obtained with cellulose nitrate filters were always poorer. It was speculated that the poor recovery obtained using cellulose nitrate filters is due to the dry conditions in these filters compared to their gelatin counterparts, and furthermore that recovery rates might be improved by collection of the aerosol in liquid versus sampling with gelatin filters.

A reliable air sampling method is a prerequisite to calculate the inhaled aerosol dose by animals exposed to the aerosol as precise as possible. For example, in pathogenesis and protection studies with virulent and vaccine strains this is of importance.

In the present work aerosols of a M. gallisepticum field or vaccine strain were simultaneously sampled with gelatin filters and by impinger. Moreover, the effect of the addition of Mycoplasma Experience (ME) broth and/or buffered peptone water (BPW) to the aerosol fluid on aerosol titres was examined. As positive control, aerosols of E. faecalis were included as it had been demonstrated that this bacterium is quite resistant in aerosolized state (Landman & Van Eck 2001; Landman et al. 2001, 2003, 2004; Zhao et al. 2011). Finally, the moisture content of gelatin and cellulose nitrate filters, both unused and filters through which air had been aspirated (further referred to as exposed filters), was assessed.

2. Materials and methods

2.1. Bacterial suspensions and aerosol fluids

Bacterial suspensions containing a M. gallisepticum field strain, a M. gallisepticum vaccine strain or an E. faecalis strain were made.

The M. gallisepticum field strain (chicken/NL/Deventer/SP1608Vin/99) was obtained in 1999 from broilers with respiratory distress. It was isolated on ME solid agar medium (ME solid medium, Reigate, Surrey, UK) enriched with ampicillin (0.2 mg/ml) and identified by indirect immunofluorescent assay (Kiernan 1990) using rabbit polyclonal antibodies. Per agar plate, three small slices of ME agar containing M. gallisepticum colonies were each submerged in 1 ml ME broth and incubated at 37°C until colour change of the broth was observed. ME broth containing vials with mycoplasma replication were freeze dried and stored at −70°C.

The M. gallisepticum field and live vaccine strain were prepared by dissolving the freeze-dried pellets in such a volume of PBS (phosphate-buffered saline) (Experiments 1, 2, 4 and 5) or ME broth (Experiment 3) in order to obtain approximately 10^7.0 colony forming units (cfu)/ml. Control of the bacterial concentrations of the mycoplasma suspensions before additions were made (see below) was performed by means of bacterial counting following the International Standard Organization (ISO7402 1985) at the start of aerosolizations. Briefly, decimal dilutions of the mycoplasma suspensions were made in ME broth (10⁻¹ to 10⁻⁷) following the International Standard Organization (ISO6887 1983). Of each dilution, 20 μl were plated on ME agar. Incubation was performed at 37°C until mycoplasma colonies were visible, then the initial bacterial concentration was determined.

An E. faecalis strain (isolate 6085.94; GD - Animal Health Service submission number 6085 and year 1994) isolated from a spontaneous case of amyloid arthropathy (Landman et al. 1994) and known to induce amyloid arthropathy (Landman et al. 1997) was used to prepare the bacterial suspensions as described previously (Landman & Van Eck 2001). Briefly, frozen beads containing the isolate (−70°C) were rolled on a sheep blood agar plate. After overnight incubation at 37°C, colonies were scraped off and a suspension aimed at a concentration of approximately 10⁷ cfu/ml was prepared in PBS (ISO6887 1983). Control of the bacterial concentration was performed by means of bacterial counting according to international standards (ISO7402 1985) at the start of aerosolizations. Of each dilution, 1 ml was plated on sheep blood agar. Incubation was performed at 37°C until colonies were visible, then the initial bacterial concentration was determined. The bacterial suspensions were kept at about 6°C until use.

Immediately before aerosol generation aerosol fluids were prepared by adding PBS or BPW to the bacterial suspensions as follows:

• Experiment 1: M. gallisepticum field strain suspension in PBS plus an equal volume of PBS.

• Experiment 2: M. gallisepticum field strain suspension in PBS plus an equal volume of BPW, which contained 10 g peptone, 5 g sodium chloride, 3.7 g disodium phosphate and 1.5 g potassium dihydrogen phosphate per litre of water.

• Experiment 3: M. gallisepticum field strain suspension in ME broth plus an equal volume of BPW.

• Experiment 4: Live M. gallisepticum vaccine strain suspension in PBS plus an equal volume of PBS.

• Experiment 5: Live M. gallisepticum vaccine strain suspension in PBS plus an equal volume of BPW.

• Experiments 6 and 7: E. faecalis strain suspension in PBS plus an equal volume of BPW.

2.2. Aerosol generation and characterization

In each experiment, 20 ml of aerosol fluid was used. The aerosolized doses per m³ of air were calculated using the titres of the bacterial suspensions taking into account the dilution factor due to the additions and the volume of the isolator. From the start of the aerosol generation until 30 min after ending, the isolator ventilation was switched off.

All nebulizations were performed at a temperature and a relative humidity (RH) of approximately 20°C and 60%, respectively, in a single empty isolator (Beyer and Eggelaar, Utrecht, the Netherlands) with a volume of 1.3125 m³. The aerosols were generated using a spray head with an orifice diameter of 0.5 mm (Walther Pilot I spray-head, Walther Spritz- and Lackiersysteme, Wuppertal, Germany) coupled to an air compressor (Mecha Concorde, type 7SAX, 1001, 10 bar/max, SACIM, Verona, Italy) at a pressure of 2 bar and an air yield of 30 l/min, resulting in a flow of 42 ml/min. During aerosolization, the RH of the isolator air raised to approximately 100%.

The droplet spectrum of the aerosol was assessed as described previously (Landman et al. 2004): D(v, 0.1) = 2.7 μm, D(v, 0.5) = 9.9 μm and D(v, 0.9) = 30.7 μm.

Between experiments the isolator was ventilated during 20 min at 40 m³ per hour, resulting in a dilution factor of approximately 10⁻³, which will reduce microorganism titres to irrelevant levels (Landman et al. 2004).

2.3. Assessment of aerosol titres and detection limits

Aerosol samples were taken simultaneously using sterile gelatine filters (3 μm pore size and 80 mm diameter, type 17528-ACD, Sartorius B.V., Nieuwegein, the Netherlands) and an impinger (code 47950, Nater Gas- en Vloeistofsystemen, Krimpen aan de IJssel, the Netherlands) (Figure 1) immediately after aerosol generation and 25 min later. Gelatine filters were coupled to a MD8 Airscan (Satorius B.V.). The aerosol was sampled during 2 min at 2000 l per hour (67 l), after which the filters were stored in a plastic container with a volume of 1000 ml and kept in a refrigerator until further processing. The impinger was filled with 200 ml sterile BPW to which an antifoam emulsion (Dow Corning, 1520, silicone antifoam, Mavon B.V., Alphen aan de Rijn, the Netherlands) (1 ml 1:10 in demineralized water diluted antifoam emulsion per 10 ml BPW) and glass beads with a diameter of 8 mm (Omnilabo International B.V., Breda, the Netherlands) were added.

Figure 1. The image depicts the impiger used in the present study for the recovery of M. gallisepticum and E. faecalis from experimentally produced aerosols. Arrows indicate the direction of the airflow through the impinger.

Glass beads were used to improve air dispersion in the impinger in all experiments, except Experiment 7. The impinger was connected to a membrane vacuum pump (NO22AN.18, KNF, Benelux) with an aspiration capacity of 12 l of air per min. Sampling time was 4 min (48 l). After sampling, the liquid from the impinger was transferred to sterile bottles, which were stored in a refrigerator until processing. The time between storage and processing of all samples was 4 h maximum.

The gelatin filters were dissolved in 50 ml sterile BPW kept at 37°C. Thereafter, three decimal dilutions (10⁻¹ to 10⁻³) of the dissolved gelatine filters and of the BPW from the impinger were made in BPW (ISO6887 1983). For the mycoplasmas, 20 μl of the undiluted suspension and of each dilution were pipetted onto ME agar, while for E. faecalis a volume of 0.1 ml was pipetted onto sheep blood agar plates. Mycoplasma plates were cultured for 1 week at 37°C, and then colonies were counted. E. faecalis was cultured aerobically for 48 h at 37°C. Identification of mycoplasmas and E. faecalis was performed as described previously (Landman et al. 1994; Landman & Feberwee 2001).

The detection limits for the mycoplasmas were 10^4.6 cfu and 10^5.4 cfu per m³ of air for the gelatine filters and the impinger, respectively. For E. faecalis these values were 10^3.9 cfu and 10^4.7 cfu per m³ of air, respectively. Calculation of detection limits was done as described previously (Landman et al. 2004).

Between experiments, the impinger was rinsed thoroughly with hot water (approximately 60°C) and was subsequently tested for bacteriological sterility. Hereto the impinger was filled with 250 ml sterile BPW. The undiluted BPW was then cultured for M. gallisepticum and E. faecalis as described earlier. Results of these bacteriological examinations were always negative.

2.4. Calculation of initial titre loss and half-life time of viable bacteria in the aerosol

The initial titre loss was calculated by subtracting the log₁₀ titre of bacteria per m³ of air found immediately after aerosol generation (t = 0 min) from the log₁₀ dose per m³ of air.

The half-life time of viable bacteria was determined for the period from t = 0 min to t = 25 min by using the following formula: $$t_{1/2}\left(\text{min}\right)=\frac{\left({log}_{10}2\right)\times t}{{log}_{10}\frac{a}{b}},$$ where t is the time interval in minutes, a is the bacterial concentration immediately after aerosol generation (t = 0 min) and b is the bacterial concentration at t = 25 min.

2.5. Assessment of the moisture content of gelatin and cellulose nitrate filters

The experimental design is outlined in Table 2. Moisture content of gelatin (3.0 μm pore size and 80 mm diameter, batch number 17528120130, expiry date 11-2017, type 17528-80-ACD, Sartorius B.V.) and cellulose nitrate filters (0.2 μm pore size and 50 mm diameter, batch number 0911114071101623, expiry date 09-2015, type 11407-50-ACN, Sartorius B.V.) was determined by drying and gravimetrics. Filters placed in separate aluminium cups were dried in a dry oven with forced ventilation at an air temperature of 102–104°C during 2 h. From a preliminary experiment it appeared that a drying period of 4 h did not further reduce weight (data not shown), indicating that the maximal loss of moisture was obtained within 2 h. Based on weights (mg balance) before and after drying, the moisture content was calculated and expressed as percentage of the weight of the non-dried filters (% wt/wt).

Table 1. Recovery of M. gallisepticum (Mg) and E. faecalis (Ef) strain 6085.94 from experimentally produced aerosols. Aerosol samples were simultaneously taken with gelatine filters (GF) coupled to a MD8 Airscan and by impinger (IMP) connected to a membrane vacuum pump. The aerosol doses and the titres of aerosols were assessed by culture and are expressed as log₁₀ colony forming units per m³ of air.

		Aerosol titre
			t = 0 min^1		Initial loss^2		t = 25 min		Half-life time (min)^3
Exp.	Aerosol fluid	Dose	GF	IMP	GF	IMP	GF	IMP	GF	IMP
1	Mg field strain in PBS^4	7.5	5.1	<5.4^5	2.4	>2.1	4.6	<5.4	15	–^6
2	Mg field strain in PBS + BPW^7	7.5	6.3	<5.4	1.2	>2.1	<4.6^5	<5.4	<4	–
3	Mg field strain in ME broth^8 + BPW	7.5	6.4	6.3	1.1	1.2	5.8	<5.4	13	<8
4	Mg vaccine strain in PBS	8.0	<4.6	<5.4	>3.4	>2.6	<4.6	<5.4	–	–
5	Mg vaccine strain in PBS + BPW	8.1	<4.6	<5.4	>3.5	>2.7	<4.6	<5.4	–	–
6	Ef in PBS + BPW	8.1	7.1	7.1	1.0	1.0	6.4	6.8	11	25
7	Ef in PBS + BPW	8.3	7.4	7.1^9	0.9	1.2^9	6.3	6.5^9	7	13

¹t = 0 min is the end of aerosol generation.

²Difference between log₁₀ dose and log₁₀ aerosol titre at t = 0 min.

³Half-life time is calculated based on titres at t = 0 and t = 25 min.

⁴PBS is phosphate-buffered saline.

⁵Value below detection limit.

⁶Calculation of half-life time is not possible.

⁷BPW is buffered peptone water.

⁸ME broth is Mycoplasma Experience broth.

⁹Impinger used without glass beads.

Table 2. Moisture content of gelatin and cellulose nitrate filters assessed by drying¹ and gravimetrics. Both unused filters and filters through which air had been aspirated were examined.

		Air aspirated through filters
Filter type	n	Yes/no	Volume (l)	Duration (min)	T (°C)	RH (%)	Mean moisture content (% wt/wt) ± SD^2 (range)
Gelatin	10	No	–^3	–	–	–	12.2 ± 1.0 (10.4–13.5)^4
	5	Yes	67	2	20–23	52–62	12.2 ± 1.3 (9.8–12.8)^4
	5	Yes	67	2	24–27	20–26	11.9 ± 0.9 (11.4–13.6)^4
	5	Yes	670	20	24–27	20–26	9.9 ± 1.1 (8.8–11.2)^5
Cellulose nitrate	5	No	–	–	–	–	3.0 ± 0.8 (1.9–3.7)^4
	5	Yes	50	2	20–23	52–62	2.7 ± 0.1 (2.7–2.8)^4

¹Drying was performed in air of 102–104°C during 2 h.

²Standard deviation.

³Irrelevant.

^4,5Values of the same filter type with different superscript differ significantly (P < 0.05).

In order to examine the moisture loss of the filters during air filtration (except for unused filters) also the moisture content of exposed filters was assessed. Hereto, various volumes of air with different temperature and RH were aspirated through both filter types (Table 2) with a MD8 Airscan (33.5 l per min) (Sartorius B.V.) for the gelatin filters and a MD2 for the cellulose nitrate filters (25.0 l per min) (Sartorius B.V.). Immediately thereafter, each filter was stored in a plastic bag, which was sealed after voiding air manually. Filters were processed directly after exposure. The temperature and RH of the air was monitored using a temperature and humidity logger (Testo 400 data logger, productnr. 0563 4001 coupled to a comfort sensor (°C/RH/m/s), productnr. 0635 1540, Testo B.V., Almere, the Netherlands).

2.6. Statistical analysis

Statistical analysis of the moisture percentages of the filters was performed with the one-tailed Mann-Whitney U test. P < 0.05 was considered significant. Average values are given ± SD.

3. Results

3.1. Recovery of M. gallisepticum and E. faecalis from aerosols

M. gallisepticum and E. faecalis recovery with gelatin filters and by impinger is presented in Table 1.

3.1.1. M. gallisepticum field strain

The M. gallisepticum field strain was aerosolized in doses of 10^7.5 cfu per m³ air (Experiments 1, 2 and 3). Immediately and 25 min after aerosol generation, aerosol titres per m³ air measured using gelatin filters ranged from 10^5.1 to 10^6.4 cfu and from <10^4.6 (below detection limit) to 10^5.8 cfu, respectively.

M. gallisepticum was recovered from one impinger sample, representing a titre of 10^6.3 cfu per m³ air (Experiment 3; aerosol titre at t = 0 min). In all other impinger samples titres were below the detection limit of 10^5.4 cfu per m³ air.

Highest aerosol titres were measured in case the M. gallisepticum field strain was nebulized suspended in ME broth with BPW (Experiment 3).

3.1.2. Live M. gallisepticum vaccine strain

Aerosol doses were 10^8.0 or 10^8.1 cfu per m³ air (Experiments 4 and 5). Viable mycoplasmas were not detected in the aerosols at any time neither using gelatin filters nor by impinger sampling.

3.1.3. E. faecalis strain 6085.94

E. faecalis doses were 10^8.1 and 10^8.3 cfu per m³ (Experiments 6 and 7). Differences in aerosol titres measured by means of gelatin filters and the impinger were negligible. Titres ranged from 10^7.1 to 10^7.4 and from 10^6.3 to 10^6.8 cfu per m³ air immediately after aerosol generation and 25 min later, respectively.

The presence of glass beads in the impinger seemed not to have any influence on the recovery of E. faecalis.

3.2. Initial loss of viable bacteria

The initial loss of viable microorganisms (loss between start and end of aerosol generation) was lowest in E. faecalis aerosols (0.9–1.2 log₁₀ cfu; Experiments 6 and 7) and highest in aerosols of the live M. gallisepticum vaccine strain (>2.6 to >3.5 log₁₀ cfu; Experiments 4 and 5). Initial loss in the M. gallisepticum field strain aerosol was in between (1.1 to >2.1 and 2.4 log₁₀ cfu; Experiments 1, 2 and 3).

3.3. Half-life time of viable bacteria

The half-life time of viable bacteria in aerosols assessed between the end of aerosol generation and 25 min later was <4–15 min for the M. gallisepticum field strain (Experiments 1, 2 and 3) and 7–25 min for the E. faecalis strain (Experiments 6 and 7). The half-life time of viable mycoplasmas in the live M. gallisepticum vaccine strain aerosols could not be assessed as all aerosol samples had titres below detection limits.

3.4. Moisture content of gelatin and cellulose nitrate filters

Mean weight ± SD (range) of unexposed gelatin (n = 10) and cellulose nitrate filters (n = 5) was 340 ± 45 mg (251–408) and 109 ± 0.4 mg (108–109), respectively. The moisture content of unexposed and exposed filters is presented in Table 2. Moisture content of unexposed gelatin and cellulose nitrate filters ranged from 10.4% to 13.5% (wt/wt) and from 1.9% to 3.7% (wt/wt), respectively. Moisture content of gelatin and cellulose nitrate filters through which 50–67 l of air of various temperatures and RH had been aspirated during a short period of time (2 min) did not differ significantly (P > 0.05) from that of unexposed filters. However, significant loss of moisture (P < 0.05) was obtained in gelatin filters after they had been exposed to 670 l of air with a temperature of 24–27°C and a RH of 20–26% during 20 min. Their mean moisture content dropped from 12.2% (wt/wt) (mean value of unexposed filters) to 9.9% (wt/wt).

4. Discussion

Generally, gelatin filters are considered to be unsatisfactory for collecting airborne bacteria due to the occurrence of desiccation of the filters resulting from (extended) sampling, which inflicts dehydration stress on the collected microorganisms (Crook 1995). In agreement herewith, Li et al. (1999) showed that the average relative recovery values obtained by gelatin filter sampling of airborne Escherichia coli bacteria in proportion to recovery rates measured by AGI 30 all glass impinger were 4–7% at sampling times of 1–5 min decreasing to 0.2% at sampling times of 45–60 min. Airflow rates through the gelatin filters and temperature, and RH of the aerosols were similar to those in the present study. Also, in comparison with the impinger E. coli recovery from gelatin filters decreased rapidly with increasing storage times (Li & Lin 2001). The poor results obtained by gelatin filter sampling were attributed to exsiccation of the filters. In contrast, in the present study sampling of M. gallisepticum aerosols using a fluid medium (BPW in impinger) did not reveal higher aerosol titres than sampling with gelatin filters. Incidentally, titres measured using gelatin filters were even higher compared to those obtained with the impinger. Short lasting sampling (2 min, 67 l of air) did not result in a significant moisture loss even in case the temperature and RH of the sampled air were 24–27ºC and 20–26%, respectively, while a 20 min exposure period (670 l of air with mentioned temperature and RH) induced a significant (P < 0.05) but still limited loss (mean moisture content dropped from 12.2% to 9.9% (wt/wt)). The foregoing illustrates the high moisture restraining ability of gelatin. Exsiccation of the gelatin filters used in the aerosol recovery experiments presented in this study did not occur, as short lasting sampling was performed (2 min). Moreover, RH of the sampled aerosols was almost 100%. We measured a moisture (likely water) content of gelatin filters of approximately 10–14% (wt/wt), which is in agreement with values found by the manufacturer (J. Grigat, Sartorius Stedim Biotech, personal communication, 2013). Apparently, this amount of moisture prevents dehydration of mycoplasmas sufficiently to prevent loss of viability. The low moisture content of the cellulose nitrate filters (approximately 2–4% (wt/wt), which was close to that specified in the product leaflet), likely explains the poor results obtained in the past with these filters (Landman et al. 2004). The latter suggestions are supported by the findings of Nei et al. (1966), who showed that high residual moisture contents after freeze drying rapidly induced lower survival rates of E. coli bacteria stored under vacuum or nitrogen, while the contrary was true for bacteria stored under atmospheric conditions.

In previous investigations, it was demonstrated that differences exist between mycoplasma species in their ability to survive under various conditions (Christensen et al. 1994; Nagatomo et al. 2001). M. synoviae seems to be more vulnerable than M. gallisepticum also in aerosols (Christensen et al. 1994; Landman et al. 2004). Moreover, Christensen et al. (1994) also demonstrated differences in the surviving ability between strains of a single mycoplasma species, namely, between M. gallisepticum strains. The same phenomenon was also observed in the present study. The live M. gallisepticum vaccine strain was clearly more vulnerable compared with the M. gallisepticum field strain. In spite of higher aerosolized doses of the live vaccine strain (10^8.0 to 10^8.1 cfu per m³ of air versus 10^7.5 cfu per m³ of air for the field strain), mycoplasmas were not recovered from the aerosols, neither using gelatin filters nor by impinger sampling. However, it should be noted that based on the former finding it cannot be concluded that viable mycoplasmas of the live vaccine strain were not present in the aerosols at all, and that birds exposed to these aerosols would not develop an immune response. Since the detection limits were relatively high (10^4.6 and 10^5.4 cfu per m³ of air for the gelatin filters and the impinger, respectively) mycoplasma concentration might still be sufficient to provoke an adequate immune response.

Whether the high loss of viability of the vaccine strain occurred during aerosolization or in the samplers can not be concluded from this study, while an explanation for the higher loss of the M. gallisepticum vaccine strain versus the field strain was not found either.

Addition of BPW to the aerosol fluid did not clearly improve the recovery of the M. gallisepticum field strain from the aerosol, while addition of ME broth and BPW did. These conclusions should be interpreted with caution as they are based on one experiment only.

Aerosol recovery rates of E. faecalis with gelatin filters and by impinger were almost equal. Initial losses (0.9–1.2 log₁₀) and half-life times in the aerosols (7–25 min) were in agreement with the results of earlier studies (Landman & Van Eck 2001; Landman et al. 2001, 2003, 2004; Zhao et al. 2011). Filling of the impinger with glass beads did not influence the recovery rate of viable E. faecalis (Experiment 6 versus Experiment 7).

In conclusion, it can be stated that gelatin filters, likely due to the relative high moisture content, are at least as useful as impingers for the recovery of M. gallisepticum from aerosols, provided exsiccation of the filters is prevented. In this case, the use of gelatin filters is to be preferred because of its greater practicability compared to the impinger. However, prevention of dehydration of the filters can only be achieved under experimental conditions.