126311522706J Aerosol SciJ Aerosol SciJournal of aerosol science0021-850226806982471749110.1016/j.jaerosci.2015.08.003NIHMS722569ArticleApplication of ATP-based bioluminescence for bioaerosol quantification: effect of sampling methodHanTaewonWrenMelodyDuBoisKelseyTherkornJenniferMainelisGediminas*Department of Environmental Sciences, Rutgers, The State University of New Jersey, 14 College Farm Road, New Brunswick, NJ 08901, USACorresponding author: Gediminas Mainelis, Phone: 848-932-5712, Fax: 732-932-8644, mainelis@envsci.rutgers.edu810201511220150112201690114123

An adenosine triphosphate (ATP)-based bioluminescence has potential to offer a quick and affordable method for quantifying bioaerosol samples. Here we report on our investigation into how different bioaerosol aerosolization parameters and sampling methods affect bioluminescence output per bacterium, and implications of that effect for bioaerosol research. Bacillus atrophaeus and Pseudomonas fluorescens bacteria were aerosolized by using a Collison nebulizer (BGI Inc., Waltham, MA) with a glass or polycarbonate jar and then collected for 15 and 60 min with: (1) Button Aerosol Sampler (SKC Inc., Eighty Four, PA) with polycarbonate, PTFE, and cellulose nitrate filters, (2) BioSampler (SKC Inc.) with 5 and 20 mL of collection liquid, and (3) our newly developed Electrostatic Precipitator with Superhydrophobic Surface (EPSS). For all aerosolization and sampling parameters we compared the ATP bioluminescence output per bacterium relative to that before aerosolization and sampling. In addition, we also determined the ATP reagent storage and preparation conditions that that do not affect the bioluminescence signal intensity.

Our results show that aerosolization by a Collison nebulizer with a polycarbonate jar yields higher bioluminescence output per bacterium compared to the glass jar. Interestingly enough, the bioluminescence output by P. fluorescens increased substantially after its aerosolization compared to the fresh liquid suspension. For both test microorganisms, the bioluminescence intensity per bacterium after sampling was significantly lower than that before sampling suggesting negative effect of sampling stress on bioluminescence output. The decrease in bioluminescence intensity was more pronounces for longer sampling times and significantly and substantially depended on the sampling method. Among the investigated method, the EPSS was the least injurious for both microorganisms and sampling times.

While the ATP-based bioluminescence offers a quick bioaerosol sample analysis method, this works demonstrates that the method output depends on bioaerosol generation and sampling methods, as well as reagent storage.

BioaerosolsSamplingATP BioluminescenceAerosolizationSampling stressINTRODUCTION

Airborne microorganisms are associated with various adverse human health effects, such as infectious and chronic respiratory diseases (Franklin et al., 2015; Lavoie et al., 2015; Pakpour et al., 2015). Traditionally, bioaerosol samples have been analyzed using agar-based techniques (Shahamat et al., 1997); however, culture-based methods detect only the culturable bioaerosol fraction and require several days before samples can be quantified. Thus, there has been a continuous interest in the development of methods for rapid quantification of airborne biological agents sampled from laboratory and field environments (Han et al., 2010; Marć et al., 2015; Pan, 2015; Ramji et al., 2014; Seshadri et al., 2009). As a result, numerous methods for quickly analyzing bioaerosol samples have been developed and applied, including impedimetry (Choi et al., 2009; Varshney & Li, 2009; Yang et al., 2004), piezoelectricity (Ebersole et al., 1990; Kovář et al., 2014), and nucleic acid based assays (Berchebru et al., 2014). An adenosine triphosphate (ATP) - based bioluminescence method (Lee et al., 2008; Mandal & Brandl, 2011; Park et al., 2014; Seshadri et al., 2009) is another useful tool for simple, rapid, and effective monitoring of microbial contamination. Since ATP is present in all biologically active microorganisms as the basic energy molecule, it could be applied to detect all metabolically active cells in bacteria, fungi, and other cells (Lim et al., 2005). The amount of ATP per cell changes according to various environmental and physiological conditions of bacteria (Mempin et al., 2013) and also varies significantly as a function of growth rate (Schneider & Gourse, 2004). When analyzing bioaerosol samples, the amount of bioluminescence produced during the reaction between ATP and appropriate enzymes is proportional to the collected bioaerosol concentration (Karl, 1980). Thus, the method could be used as an indicator of total bioaerosol load in field or in situ studies (Han et al., 2015b; Lin et al., 2013; Park et al., 2014; Yoon et al., 2010). Some studies are strong correlation between ATP content in indoor air and concentration of culturable bacteria (Yoon et al., 2010). It has also been shown that the exact relationship between the ATP concentration and the intensity of bioluminescence reaction depends on microorganism species (Han et al., 2010). Once a calibration curve relating ATP output of particular species with their cell counts is built, the ATP method can provide a quick and affordable solution for quantifying bioaerosol samples, including in laboratory studies to estimate the efficacy of various bioaerosol samplers (Seshadri et al., 2009).

At the same time, different bioaerosol sampling methods are known to differ in their effect on bioaerosol culturability (Zhao et al., 2011) and cell-wall integrity (Thomas et al., 2011; Zhen et al., 2013). Since ATP-based bioluminescence is based on metabolic activity of microorganisms, the strength of ATP bioluminescence signal and accuracy of bioaerosol analysis may therefore be affected by the chosen sampling method as well. To the best of our knowledge, such information is currently not available.

Thus the main goal of this study was to further investigate the applicability of ATP-based bioluminescence for bioaerosol analysis by analyzing how bioaerosol sampling methods (filtration, impingement, and electrostatic precipitation) and aerosolization parameters affect the strength of ATP bioluminescence signal. In addition, we also investigated how the ATP reagent storage time and the natural degradation of a bioaerosol sample over time affects the ATP bioluminescence signal. The results of this work will aid researchers in adapting ATP-based bioluminescence methods for bioaerosol research.

MATERIALS AND METHODS

The study was structured in three steps to investigate how the following factors affect the strength of the luminescence signal: 1) effect of the ATP reagent age, 2) effect of bioaerosol sample age, and 3) effect of bioaerosol aerosolization and sampling parameters. The experiments were performed with gram-positive Bacillus atrophaeus bacterial cells and gram-negative Pseudomonas fluorescens. Table 1 summarizes all the variables that were investigated in this study.

General procedure to measure ATP bioluminescence

To determine ATP content in bioaerosol samples, we used BacTiter-Glo™ microbial cell viability assay (Promega Corp., Madison, WI) and followed the manufacturer’s instructions as described elsewhere (Seshadri et al., 2009). Briefly, the lyophilized BacTiter-Glo substrate and 10 mL of buffer included in the assay kit were thawed out and reconstituted in equal parts. For measurement of ATP, 100 µL of the resulting reagent (BacTiter-Glo substrate plus buffer) was then mixed with 100 µL of bioaerosol sample. The ensuing reaction lysed the cells and produced a luminescence signal, which was measured by a luminometer (model 20/20n, Turner Biosystems Inc., Sunnyvale, CA). The strength of the signal was measured as Relative Luminescence Units (RLU) and is proportional to the amount of ATP that participated in the reaction.

1) Investigation of the luminescence signal strength as a function of the ATP reagent age

This part had two goals: a) to determine if storing ATP reagents in a freezer (−20 °C) affects intensity of the bioluminescence signal, and b) to determine if storing thawed out ATP reagents in a refrigerator (4 °C) affects intensity of the bioluminescence signal. The first part was designed to determine how long reconstituted reagents could be stored before they affect bioaerosol quantification. The latter part was designed to simulate bioaerosol experiments when ATP reagents are ready (thawed out) but cannot be applied immediately because bioaerosol samples are acquired over several hours. Both of these steps were necessary to develop a reliable protocol for bioaerosol quantification by ATP bioluminescence.

A bottle of fresh BacTiter-Glo buffer was completely thawed out (about 1 hour), equilibrated to room temperature (23 – 25 °C) and then reconstituted with the included reaction substrate. The buffer and substrate were mixed by gentle vortexing to obtain a homogenous solution (i.e., reagent). Ten mL of the resulting reagent (buffer plus substrate) were subdivided into 20 microcentrifuge vials in equal volumes (500 µL) and stored in a freezer at −20 °C for up to 14 days. A 10 µM ATP standard (Lonza Inc., Walkersville, MD), a specialized aqueous preparation of adenosine triphosphate with a known amount of ATP, was also subdivided into 20 microcentrifuge vials in equal volumes (500 µL) and placed at −20 °C.

Experiments were performed with ATP standards and ATP reagents (BacTiter-Glo buffer and substrate) of different age (1, 3, 5, 7, 9, and 14 days) to investigate a potential decay in reagent activity with storage time. During each predetermined day, six vials of reconstituted ATP reagents and six vials ATP standard (500 µL each) were removed from the freezer and completely thawed out (approximately 1 hr). Once completely thawed, the vials were placed in a refrigerator (4 °C) and removed at predetermined time intervals (0, 1, 2, 4, and 6 hours) to measure bioluminescence intensity of the ATP standard. Then, 100 µL of ATP reagent was mixed with 100 µL of ATP standard, the contents were briefly vortexed, left at room temperature for 1 min, and measured for the resulting bioluminescence reaction by a luminometer (model 20/20n, Turner Biosystems Inc., Sunnyvale, CA). The mixing and bioluminescence measurements were performed in triplicate for each investigated time period.

2) Measurement of the luminescence signal as a function of bioaerosol sample age

These tests were designed to investigate if the intensity of luminescence changes when there is time lag between when the bioaerosol sample is collected and eventually analyzed. The tests were conducted with gram-positive Bacillus atrophaeus bacterial cells (ATCC 49337, American Type Culture Collection, MD) and gram-negative Pseudomonas fluorescens (ATCC 13525). Both microorganisms are commonly used as test microorganisms (Hill et al., 1999; Horner et al., 1995; Johnson et al., 1994; Madelin, 1994) and their mean aerodynamic diameters are 0.82 and 0.89 µm (Han et al., 2010), respectively. For each experiment, fresh bacterial cells were grown in trypticase soy broth (Becton Dickinson Microbiology Systems, Cockeysville, MD) at 26°C and 30°C, respectively, for 18 hrs in a gyratory shaking incubator (Amerex Instruments Inc., Lafayette, CA) (Han et al., 2010). The used growth time yielded optimum ATP signal (data not shown). The bacterial cells were harvested by centrifugation at 7,000 rpm (6140g) for 5 min at 4°C (BR4; Jouan, Winchester, VA) and then washed 4 times with sterile deionized water. Prior to experiments, the resulting cell pellets were resuspended in sterile deionized water.

To simulate bioaerosol samples of different age, the final liquid suspension of each bacteria was subdivided into centrifuge tubes with equal volume (10 mL), 8 tubes in total, and stored at room temperature for 0.08, 0.25, 1, 2, 3, 4, 5, and 6 hours. The luminescence signal was measured immediately after each aging time by combining 100 µL of bacterial suspension with 100 µL of Bactiter-Glo reagent and measuring the resulting signal with a luminometer. For each aging time, the luminescent signal at the time t= 0 (as soon as the suspension was prepared) was measured as a reference. The effect of aging of bioaerosol sample was determined by calculating the ratio of the luminescence value (i.e., RLUi) at each time period (i.e., 0.08, 0.25, 1, 2, 3, 4, 5, and 6 hours) to the initial luminescence value (i.e., RLUo) at t = 0.

3) Effect of bioaerosol aerosolization and sampling parameters on luminescence signal Experimental setup

Fig. 1 illustrates the experimental setup used to investigate how different bioaerosol sampling methods and sampling times (i.e., 15 and 60 min) affect the measured bioluminescence intensity. The test system consists of a flow control system, a particle generation system, an air-particle mixing system, and a particle monitoring system. The system was housed inside a Class II Biosafety cabinet (NUAIRE Inc., Plymouth, MN). The relative humidity during the experiments was approximately 40 – 50%.

A three-jet Collison nebulizer (BGI Inc., Waltham, MA) with either a glass or polycarbonate jar was used to aerosolize test particles (i.e., B. atrophaeus and P. fluorescens) from a liquid suspension at a flow rate (QA) of 5 L/min (pressure of 12 psi). A HEPA-filtered dilution air flow, QD (105 L/min), provided by an in-house compressor, was controlled by a pressure regulator and monitored by a mass flowmeter (TSI Inc., Shoreview, MN). The aerosolized particle stream and dilution air were combined and passed through a 2-mCi Po-210 charge neutralizer to reduce aerosolization-induced particle charges to Boltzmann equilibrium. The electrically neutralized particles then passed through the first mixing box which improved the uniformity of particle distribution across the flow cross-section (Han et al., 2005). A second mixing box and a U-type duct connector enhanced turbulence further and improved particle mixing. A flow straightener was placed at the exit of the second mixing box to eliminate large scale turbulence and flow swirl generated by the mixing boxes. Finally, the well-mixed aerosol was introduced into a 0.152 m (6 inches) diameter test duct. Each test sampler was placed 6 duct diameters downstream of the exit of the flow straightener in order to provide a uniform crosssectional concentration of the test particles. The aerosol concentration was monitored by an optical particle counter (model 1.108, Grimm Technologies Inc., Douglasville, GA) (The counter is not shown in Fig. 1).

The tested bioaerosol samplers included a Button aerosol sampler (SKC Inc., Eighty Four, PA), a BioSampler (SKC Inc.) and a newly developed electrostatic precipitator with a superhydrophobic surface (EPSS) (Han et al., 2015a; Han et al., 2011). The Button aerosol sampler and the BioSampler were chosen for this study because they operate based on filtration and liquid impingement collection mechanisms, respectively, which are two of the most commonly used bioaerosol sampling methods (Reponen et al., 2011). Also, collected samples could then be analyzed using culture-independent methods and compared the newly developed EPSS (Han et al., 2010). The Button aerosol sampler was tested with three different filters: a 0.6 µm pore size polycarbonate filter (Millipore, Billerica, MA), a 0.5 µm pore size PTFE membrane filter (Millipore), and a 0.45 µm pore size cellulose nitrate filter (Millipore). It was operated at its nominal flow rate of 4 L/min. For the BioSampler, either 5 or 20 mL of collection fluid was placed in appropriate sampling cups, and the sampler was operated at its nominal flow rate of 12.5 L/min. The EPSS was operated with charging/collection voltage of −9 kV and a sampling flow rate of 10 L/min. The bacteria collected by the EPSS were removed by a 40 µL rolling water droplet for analysis.

Our earlier study showed that when bacteria are aerosolized by a Collison nebulizer, different jar material, i.e., polycarbonate vs glass, produces different effect on the integrity of bacterial cells, likely due to higher flexibility of polycarbonate compared to glass (Zhen et al., 2013). Therefore, to assess the effect of aerosolization by the Collison nebulizer on the physiological state of the test bacteria and the resulting bioluminescence signal, glass and polycarbonate jars were used and the resulting bioluminescence signals were compared..

Comparison metricsEffect of aerosolization

10 mL of freshly prepared suspension of each bacterial species was placed in the Collison nebulizer and aerosolized for 0, 1, 5, 15, and 60 min. After aerosolization, two 100 µL aliquots were transferred from the suspension remaining in the Collison nebulizer into two different 1.5 mL centrifuge tubes: one aliquot for the ATP analysis and the other for the Acridine Orange epifluorescence microscopy (AOEM) (Han et al., 2010). For each aerosolization time, the ATP bioluminescence intensity (RLU) was normalized to the total number of cells counted by microscopy (NAOEM) and ratio R was calculated: Ro=[RLUNAOEM]oorRi=[RLUNAOEM]i, where Ro is the ratio for t = 0 min (i.e., a reference ratio) and Ri is the ratio after a certain aerosolization time (i.e., 1, 5, 15, and 60 min). These ratios indicate ATP content per cell and allow to account for day to day variations in cell concentration when performing replicate experiments. The effect of aerosolization on the physiological state of bacteria was then determined by comparing the ratio after the nebulizer operation, Ri, to the initial ratio Ro (reference): RR(Collison)i=RiR0=[RLUNAOEM]i[RLUNAOEM]o, where RR(Collison)i is a dimensionless ratio comparing the ATP output per cell before and after aerosolization for a certain time i. It can also be used to investigate the effect of different Collison jar material on the ATP output per cell.

Effect of sampling methods

The biological particles were aerosolized and then sampled using the BioSampler (12.5 L/min), the Button aerosol sampler (4 L/min), and the EPSS (10 L/min). After sampling was completed by the BioSampler, the remaining collection liquid was transferred into a centrifuge tube. For the Button sampler, bacteria were eluted from the filter into sterile deionized water using the procedure described elsewhere (Wang et al., 2001). We used this procedure to determine the elution efficiency for each filter type and bacterial species since the elution efficiency affects the total number of bacteria per sample and the resulting bioluminescence reaction. For B. atrophaeus, the elution efficiency was 97 ± 14% for polycarbonate filter, 96 ± 5% for PTFE, and 98 ± 12% for cellulose nitrate filter; for P. fluorescens, the elution efficiency was 102 ± 8% for polycarbonate filter, 86 ± 3% for PTFE, and 107 ± 8% for cellulose nitrate filter. For the EPSS, once sampling was completed, a 40 µL water droplet was used to remove the collected bacteria as described elsewhere (Han et al., 2010), and the droplet containing particles was transferred into a centrifuge tube, where its volume was increased to 100 µl by adding sterile deionized water. Similar to the assessment of the effects of aerosolization stress, the effect of sampling method and time on bioluminescence intensity was investigated by comparing the RLU output per bacterium in a sample collected by a particular sampler with the RLU output per bacterium before aerosolization (suspension in Collison nebulizer): RR(sampler)k,l=Rk,lR0=[RLUNAOEM]k,l[RLUNAOEM]0 where k is sampler and l is sampling time.

Statistical analysis

The measured bioluminescence values (RLU), dimensionless ratios RR(Collison) and RR(sampler) were analyzed as a function of storage time, particle type, aerosolization time, Collison jar type, and sampler type using ANOVA (SigmaPlot 2011, Version 12.3, Systat Software Inc., San Jose, CA). The differences between individual pairs of variables were examined by using Holm-Sidak method, which takes into account multiple comparisons. The p < 0.05 was considered significant at alpha = 0.05.

RESULTS AND DISCUSSION

Fig. 2 shows the strength of bioluminescence signal (RLU) as a function of ATP reagent and ATP standard storage time (i.e., 0, 1, 2, 4, and 6 hours) at 4 °C once they were completely thawed after having been stored at −20 °C for 1, 3, 5, 7, 9, and 14 days. The results show that within 1 hour of the ATP reagent and ATP standard thawing out, the luminescence intensity of their reaction decreased on average by 36 ± 9%. However, the luminescence intensity did not change substantially over the next 5 hours indicating a time window when bioaerosol quantification by this method should be performed. Storing the reagents at −20 °C for up to 14 days did not significantly affect the reaction intensity (p > 0.05). A two-way ANOVA indicated statistically significant effect of both variables (storage time of frozen reagents at −20 °C (days) and storage time of thawed out reagents at 4 °C (hours). However, pairwise comparison using Holm-Sidak method showed that there was not significant effect when thawed out reagents were kept at 4 °C for 2 hours or longer. Within this time frame (2 – 6 hours), the effect of storing frozen reagents at −20 °C was also not statistically significant with the exception of hour 2 between the 5th and 14th day. Thus, the results indicate how to design a protocol for ATP bioluminescence method to include storage conditions that do not significantly affect the intensity of the bioluminescence signal. The reagent manufacturer also indicated that to achieve maximum sensitivity, equilibration time longer than 15 may be required.

Fig. 3 presents change of bioluminescence intensity per bacterium after a certain time in suspension relative to the bioluminescence intensity at t = 0 (ATP intensity measured immediately after preparing bacterial suspension) for B. atrophaeus and P. fluorescens. Firstly, one can observe that the RLUi / RLU0 ratio at t = 0.08 hr (5min) is above > 1 (approximately 1.48 for P. fluorescens and 1.39 for B. atrophaeus) and then begins to decrease as the bacterial suspensions are stored longer. Since the total number of bacteria in each sample does not change with time, the observed change in RLUi / RLU0 ratio could be interpreted as natural degradation of the ATP signal intensity produced by bacteria over time. Over the six hour period, the RLUi / RLU0 ratio for B. atrophaeus decreased gradually to approximately 0.97 of that at t = 0, while at 4 – 5 hours the RLUi / RLU0 ratio for P. fluorescens was still about 21% higher than that at t = 0; it did decrease to approximately 1 after 6 hours. One can also observe that after the initial decay over approximately 1 – 2 hours, the observed bioluminescence intensity does not change substantially. The pairwise comparisons of RLUi / RLU0 ratios for B. atrophaeus from 2 to 6 hours and for P. fluorescens from 1 to 5 hours were not statistically different (p > 0.05). The difference between the two bacteria was statistically significant (p < 0.001). When pair-wise comparisons of bacteria were performed for each time point, the differences were statistically significant for all time points except t = 1 and 6 hours. Thus, if bioaerosol samples are collected at different time points for analysis by ATP bioluminescence at a later time, they should be stored within the above indicated time frames to avoid bias in sample quantification.

Fig. 4 presents the effect of aerosolization on RLU output by B. atrophaeus and P. fluorescens as a function of aerosolization time (i.e., 1, 5, 15, and 60 min). The data also compares the effects due to different jar material of the Collison nebulizer: glass versus polycarbonate. One can see that the RR(Collison) for B. atrophaeus was at or below 1, depending on the Collison jar material. When aerosolizing B. atrophaeus with a glass jar, the RR(Collison) values were 0.81 ± 0.01, 0.71 ± 0.05, 0.60 ± 0.02, and 0.72 ± 0.05 for 1, 5, 15, and 60 min of aerosolization time, respectively. According to pair-wise comparison, the differences among RR(Collison) for all time points were not statistically significant. When a polycarbonate jar was used, the RR(Collison) values were close to unity for the first 15 min (1.03 ± 0.17, 1.02 ± 0.07, and 0.97 ± 0.02, for 1, 5, and 15 min, respectively) and then decreased to 0.72 ± 0.21 after 60 min of aerosolization time. According to the pair-wise comparison, the RR(Collison) after 60 min of aerosolization was statistically different from that at other aerosolization times (p < 0.05). From 1 to 15 min aerosolization time, the average RR(Collison) value for polycarbonate jar (1.01 ± 0.03) was statistically different from that for a glass jar (0.71 ± 0.11) (p < 0.001). However, for the 60 min nebulization time, the results between the glass and the polycarbonate jars were not statistically different (p = 0.989).

Conversely, the RR(Collison) values of P. fluorescens were above unity for both jar materials. For the glass jar, the values were 1.17 ± 0.12, 1.12 ± 0.19, 1.15 ± 0.05, and 1.52 ± 0.08 for 1, 5, 15, and 60 min of aerosolization time, respectively. The RR(Collison) values for 60 min were statistically different from those at other time points (p < 0.05). For the polycarbonate jar, the RR(Collison) values for the same time points were 4.65 ± 0.91, 6.29 ± 1.18, 8.37 ± 2.58, and 10.29 ± 2.33, respectively. According to pair-wise comparisons, the RR(Collison) value at 60 min was statistically different from that at 1 min (p < 0.05). All other comparison pairs were not statistically different. For each aerosolization time of P. fluorescens, the difference between the glass and the polycarbonate jars was statistically significant (p < 0.05).

Except when aerosolizing P. fluorescens with a polycarbonate jar, the RR(Collison) values averaged over time were close to unity: 0.71 ± 0.09 for B. atrophaeus and glass jar, 0.94 ± 0.14 for B. atrophaeus and polycarbonate jar, and 1.24 ± 0.19 for P. fluorescens and glass jar. For P. fluorescens and polycarbonate jar, however, the average RR(Collison) value was much higher: 7.40 ± 2.45.

The data with B. atrophaeus show that for up to 15 min of aerosolization, the polycarbonate Collison jar enables up to 30% stronger bioluminescence signal per bacterium compared to the glass jar. Surprisingly, for P. fluorescens, bioluminescence output per bacterium increased slightly when using a Collison nebulizer with a glass jar compared to the fresh suspension. Even more surprising was the substantial increase in the bioluminescence output per bacterium when the polycarbonate jar was used. We speculate that mechanical stress during the aerosolization of P. fluorescens with both the glass and polycarbonate jars causes damage to bacterial membranes, including the release of intercellular material (Zhen et al., 2014); this partial lysis results in the release of intracellular ATP, which is then available to participate in bioluminescence reaction.

However, when a glass jar is used, the ATP molecules suffer substantial mechanical stress due to their multiple impacts into the rigid glass wall and become unable to participate in bioluminescence reaction. Polycarbonate material, on the other hand, is less rigid (more flexible) and absorbs part of the impact energy thus resulting in lower stress to the ATP molecules, which are now more abundant in the suspension, leading to a higher ATP signal per bacterium. An increase in ATP bioluminescence signal from gram negative Escherichia coli following their sonication was observed in another study (Turner et al., 2010). At the same time, even if the bacterial membrane is damaged, the bacterium still retains its physical shape and can be counted under a microscope. Some fraction of bacteria might be completely damaged and not counted under a microscope thus also leading to a higher bioluminescence output per bacterium. Damage to the integrity of the bacterial membrane was also observed in our previous study of DNA release from Escherichia coli (also a gram-negative organism) after aerosolization (Zhen et al., 2013). Our previous aerosolization studies with P. fluorescens observed damage to bacterial culturability due to prolonged aerosolization (Mainelis et al., 2005). Since B. atrophaeus is a gram positive bacterium with a thick outer wall, the mechanical stress of aerosolization and sampling does not result in the cell membrane rupture (Zhen et al., 2013), but apparently the stress is still sufficient to reduce RLU output per bacterium as illustrated with RR(Collison) values being below unity. Here again, since the polycarbonate jar is able to absorb part of the impaction energy from the bacteria, it produces lower stress and higher RR(Collison) compared to the glass jar. The above-mentioned study by Turner at al. (2010) also observed minimum change in ATP bioluminescence signal when gram-positive Staphylococcus aureus was sonicated.

Fig. 5 presents RR(sampler) ratio for three different samplers and their sampling parameters, sampling times and two different bacteria.

Data for <italic>B. atrophaeus</italic>

For this bacterium, the RR(sampler), when sampled with the BioSampler, ranged from 0.07 to 0.21; when sampled with the Button sampler, the RR(sampler) ranged from 0.10 to 0.52; the EPSS showed the highest values for each time point ranging from 0.45 to 0.68. One can also observe that the RR(sampler) values for this bacterium in Fig. 5 are lower than RR(Collison) data with the polycarbonate jar in Fig. 4 where the aerosolization effects are shown. The difference was statistically significant (p < 0.001) for each tested sampler and for both sampling times. The lower values RR(sampler) compared to RR(Collison) indicate that B. atrophaeus is subject to additional stress during sampling, which leads to lower bioluminescence output per bacterium.

Overall, for B. atrophaeus data in Fig. 5, the effects of sampler and sampling time were statistically significant (p < 0.001). According to Holm-Sidak pair-wise comparison, the RR(sampler) was statistically different (p < 0.001) for all sampler pairs, except the following pairs: BioSampler with 5 mL vs. BioSampler with 20 mL; BioSampler with 5 and 20 mL vs. Button sampler with cellulose filter; and Button sampler with polycarbonate filter vs. Button sampler with cellulose filter. The effect of sampling time was statistically significant (p < 0.005) for all samplers except BioSampler with 20 mL and Button sampler with cellulose filter. Also, for each sampling time, the EPSS had the highest RR(sampler) values compared to other samplers. The differences ranged from 0.16 (Button sampler with Teflon filter) to 0.53 (Button sampler with cellulose filter) and were significant when compared with all samplers (p < 0.001).

Data for <italic>P. fluorescens</italic>

On the other hand, the average RR(sampler) for P. fluorescens were clearly above unity and ranged from 4.13 ± 0.90 (BioSampler with 5 mL) to 6.15 ± 2.32 (Button sampler with polycarbonate filter). For 15 min of sampling, the values of RR(sampler) were statistically significantly lower than RR(Collison) (p < 0.05) for BioSampler with 5 mL and the Button sampler with polycarbonate filter. For 60 min sampling, the values of RR(sampler) were statistically significantly lower than RR(Collison) for all samplers except the Button sampler with polycarbonate filter. When 15 and 60 min data were pooled together to minimize experimental variability, then RR(sampler) values were statistically significantly lower (p < 0.005) than RR(Collison) for all samplers.

Overall for P. fluorescens data set in Fig. 5, the effect of sampler on RR(sampler) was statistically significant (p < 0.001). According to Holm-Sidak pair-wise comparison, RR(sampler) for EPSS and Button sampler with polycarbonate filter were statistically significantly higher than that for BioSampler with either 5 or 20 mL of collection fluid (p < 0.05). The effect of time on the overall P. fluorescens data set was not statistically significant (p = 0.80); however, Holm-Sidak pair-wise comparison showed differences in RR(sampler) between 15 and 60 min sampling time for BioSampler with 20 mL and Button sampler with polycarbonate filter.

In general, Fig. 5 shows that, when using B. atrophaeus and P. fluorescens, the RLU output per bacterium decreased after sampling compared to RLU output per bacterium just after aerosolization. This indicates that the sampling stress and, possibly, stress during bioaerosol transport form the aerosolization point to the sampling point, results in lower bioluminescence output per bacterium. This is in agreement with other studies reporting stress to bioaerosols during sampling, including their lower culturability (Chen & Li, 2005). This study, however, focuses on the effects of sampling on bioluminescence output per bacterium. For B. atropheus, and both 15 and 60 min sampling times, the lowest difference between the RR(Collison) and RR(sampler) was for EPSS. For P. fluorescens and both 15 and 60 min sampling times, the EPSS yielded the second lowest difference between the RR(Collison) and RR(sampler). If Button sampler data with different filter materials are pooled together to minimize experimental variability, then the EPSS yields the lowest difference between the RR(Collison) and RR(sampler) among the tested sampler types. These observations suggest that of the tested samplers the EPSS is among the least stress inducing devices when it comes to bioaerosol sampling for analysis by the ATP-based bioluminescence.

Overall the data presented here show that the ATP-based bioluminescence could be used to rapidly analyze performance of bioaerosol samplers. However, one has to take into account that the RLU output depends on bacterial species, as well as sampling and aerosolization methods. Once the effects of aerosolization and sampling in a particular setup are known, the ATP-based bioluminescence analysis provides a quick and convenient way to analyze bioaerosol samples, especially during bioaerosol sampler development and testing.

CONCLUSION

We found that multiple factors affected ATP-based bioluminescence intensity when analyzing bioaerosols: i.e., reagent decay over time, natural degradation of the signal produced by the tested bioaerosol (i.e., B. atrophaeus and P. fluorescens) liquid suspension over time, nebulization method (Collison nebulizer with glass and polycarbonate jars) and also bioaerosol sampling method. Surprisingly, P. fluorescens showed higher bioluminescence intensity per bacterium after aerosolization compared to P. fluorescens suspension before aerosolization. An opposite effect was observed for B. atrophaeus: bioluminescence intensity per bacterium decreased after aerosolization compared to that before aerosolization. For both microorganisms, the EPSS affected the bioluminescence output per bacterium the least among the investigated sampling methods.

Overall, this work shows that the ATP-based bioluminescence method could be applied for rapid analysis of bioaerosol samples; however, researchers should take into account different bioluminescence output by different species as well as changes to that output depending on aerosolization and sampling techniques and parameters. Such changes might be especially important in field studies where multiple organisms aerosolized through different pathways are encountered. It is hoped that the results of this study will contribute toward wider application of this technique in bioaerosol research.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACKNOWLEDGEMENTS

This research was supported in part by the CDC/NIOSH grant R01OH009783 “Design of Advanced Electrostatic Sampler for Total Bioaerosols” from CDC-NIOSH and the New Jersey Agriculture and Experiment Station (NJAES) at Rutgers University. J. Therkorn was supported by EPA STAR Graduate Fellowship (F13A10035) and the Air and Waste Management Association’s Milton Feldstein Memorial Scholarship. M. Wren was supported by NIEHS Training Grant in Exposure Science 1T32ES019854 (PI: Weisel, C.P.). Contents of the manuscript are solely the responsibility of the authors and do not necessarily represent the official view of the funding agencies.

REFERENCESBerchebruLRameilPGaudinJ-CGaussonSLarigauderieGPujolCRamisseVNormalization of test and evaluation of biothreat detection systems: Overcoming microbial air content fluctuations by using a standardized reagent bacterial mixtureJournal of Microbiological Methods20141050141145doi: http://dx.doi.org/10.1016/j.mimet.2014.07.01325038460ChenP-SLiC-SSampling Performance for Bioaerosols by Flow Cytometry with FluorochromeAerosol Science and Technology2005393231237ChoiAParkJ-SJungH-ISolid-medium-integrated impedimetric biosensor for real-time monitoring of microorganismsSensors and Actuators B: Chemical20091371357362doi: http://dx.doi.org/10.1016/j.snb.2008.12.062EbersoleRCMillerJAMoranJRWardMDSpontaneously formed functionally active avidin monolayers on metal surfaces: a strategy for immobilizing biological reagents and design of piezoelectric biosensorsJournal of the American Chemical Society1990112832393241FranklinBABrookRArden Pope IiiCAir pollution and cardiovascular diseaseCurrent Problems in Cardiology0doi: http://dx.doi.org/10.1016/j.cpcardiol.2015.01.003FranklinBABrookRArden Pope IiiCAir Pollution and Cardiovascular DiseaseCurrent Problems in Cardiology2015405207238doi: http://dx.doi.org/10.1016/j.cpcardiol.2015.01.00325882781HanTAnHRMainelisGPerformance of an Electrostatic Precipitator with Superhydrophobic Surface when Collecting Airborne BacteriaAerosol Science and Technology2010445339348HanTFennellDMainelisGDevelopment and Optimization of the Electrostatic Precipitator with Superhydrophobic Surface (EPSS) Mark II for Collection of BioaerosolsAerosol Science and Technology2015a494210219HanTNazarenkoYLioyPJMainelisGCollection efficiencies of an electrostatic sampler with superhydrophobic surface for fungal bioaerosolsIndoor Air201121211012021204982HanTO'NealDLMcFarlandARHaglundJOrtizCAEvaluation of Mixing Elements in an L-Shaped Configuration for Application to Single-Point Aerosol Sampling in DuctsHVAC&R Research2005114657672HanTZhenHFennellDEMainelisGDesign and evaluation of the field-deployable electrostatic precipitator with superhydrophobic surface (FDEPSS) with high concentration rateAerosol and Air Quality Research, (submitted)2015bHillSCPinnickRGNilesSPanY-LHollerSChangRKFeatherGReal-time measurement of fluorescence spectra from single airborne biological particlesField Analytical Chemistry & Technology199934–5221239doi:10.1002/(SICI)1520-6521(1999)3:4/5<221::AID-FACT2>3.0.CO;2-7HornerWEHelblingASalvaggioJELehrerSBFungal allergensClinical Microbiology Reviews1995821611797621398JohnsonBMartinDDResnickIGEfficacy of Selected Respiratory Protective Equipment Challenged with Bacillus subtilis subsp. nigerApplied and Environmental Microbiology19946062184218616349306KarlDMCellular nucleotide measurements and applications in microbial ecologyMicrobiological Reviews19804447397967010116KovářDFarkaZSkládalPDetection of Aerosolized Biological Agents Using the Piezoelectric ImmunosensorAnalytical Chemistry201486178680868625088715LavoieJMarchandGCloutierYHalleSNadeauSDuchaineCPichetteGEvaluation of bioaerosol exposures during hospital bronchoscopy examinationsEnvironmental Science: Processes & Impacts2015[10.1039/C4EM00359D]LeeSJParkJSImHTJungH-IA microfluidic ATP-bioluminescence sensor for the detection of airborne microbesSensors and Actuators B: Chemical20081322443448doi: http://dx.doi.org/10.1016/j.snb.2007.10.035LimDVSimpsonJMKearnsEAKramerMFCurrent and Developing Technologies for Monitoring Agents of Bioterrorism and BiowarfareClinical Microbiology Reviews200518458360716223949LinC-JWangY-THsienK-JTsaiYIKungP-YChyanJ-MIn Situ Rapid Evaluation of Indoor Bioaerosols Using an ATP Bioluminescence Assay2013133922931Retrieved from http://ir.chna.edu.tw/handle/310902800/27860MadelinTMFungal aerosols: A reviewJournal of Aerosol Science199425814051412doi: http://dx.doi.org/10.1016/0021-8502(94)90216-XMainelisGBerryDReoun AnHYaoMDeVoeKFennellDEJaegerRDesign and performance of a single-pass bubbling bioaerosol generatorAtmospheric Environment2005391935213533doi: http://dx.doi.org/10.1016/j.atmosenv.2005.02.043MandalJBrandlHBioaerosols in Indoor Environment - A Review with Special Reference to Residential and Occupational LocationsThe Open Environmental & Biological Monitoring Journal201148396MarćMTobiszewskiMZabiegałaBGuardiaMdlNamieśnikJCurrent air quality analytics and monitoring: A reviewAnalytica Chimica Acta20158530116126doi: http://dx.doi.org/10.1016/j.aca.2014.10.01825467453MempinRTranHChenCGongHKim HoKLuSRelease of extracellular ATP by bacteria during growthBMC Microbiology2013131301[10.1186/1471-2180-13-301]24364860PakpourSLiD-WKlironomosJRelationships of fungal spore concentrations in the air and meteorological factorsFungal Ecology2015130130134doi: http://dx.doi.org/10.1016/j.funeco.2014.09.008PanY-LDetection and characterization of biological and other organic-carbon aerosol particles in atmosphere using fluorescenceJournal of Quantitative Spectroscopy and Radiative Transfer201515001235doi: http://dx.doi.org/10.1016/j.jqsrt.2014.06.007ParkCWParkJ-WLeeSHHwangJReal-time monitoring of bioaerosols via cell-lysis by air ion and ATP bioluminescence detectionBiosensors and Bioelectronics2014520379383doi: http://dx.doi.org/10.1016/j.bios.2013.09.01524080217RamjiRChiongCFHirataHRahmanARALimCTRapid Quantification of Live Cell Receptors Using Bioluminescence in a Flow-Based Microfluidic DeviceSmall2014n/a-n/aReponenTWillekeKGrinshpunSBiological particle samplingAerosol Measurement: Principles, Techniques, and Applications2011Wiley and SonsSchneiderDAGourseRLRelationship between Growth Rate and ATP Concentration in Escherichia coli: A BIOASSAY FOR AVAILABLE CELLULAR ATPJournal of Biological Chemistry200427998262826814670952SeshadriSHanTKruminsVFennellDEMainelisGApplication of ATP bioluminescence method to characterize performance of bioaerosol sampling devicesJournal of Aerosol Science2009402113121doi: http://dx.doi.org/10.1016/j.jaerosci.2008.10.002ShahamatMLevinMRahmanIGrimCHeidelbergJStelmaGColwellREvaluation of media for recovery of aerosolized bacteriaAerobiologia1997134219226ThomasRJWebberDHopkinsRFrostALawsTJayasekeraPNAtkinsTThe Cell Membrane as a Major Site of Damage during Aerosolization of Escherichia coliApplied and Environmental Microbiology201177392092521148696TurnerDEDaugherityEKAltierCMaurerKJEfficacy and Limitations of an ATP-Based Monitoring SystemJournal of the American Association for Laboratory Animal Science : JAALAS201049219019520353694VarshneyMLiYInterdigitated array microelectrodes based impedance biosensors for detection of bacterial cellsBiosensors and Bioelectronics2009241029512960doi: http://dx.doi.org/10.1016/j.bios.2008.10.00119041235WangZReponenTAGrinshpunSLGónyRWillekeKEffect of sampling time and air humidity on the bioefficiency of filter samplers for bioaerosol collectionJournal of Aerosol Science2001325661674YangLLiYGriffisCLJohnsonMGInterdigitated microelectrode (IME) impedance sensor for the detection of viable Salmonella typhimuriumBiosensors and Bioelectronics2004191011391147doi: http://dx.doi.org/10.1016/j.bios.2003.10.00915046744YoonKYParkCWByeonJHHwangJDesign and Application of an Inertial Impactor in Combination with an ATP Bioluminescence Detector for In Situ Rapid Estimation of the Efficacies of Air Controlling Devices on Removal of BioaerosolsEnvironmental Science & Technology20104451742174620143821ZhaoYAarninkAJADoornenbalPHuynhTTTKoerkampPWGGde JongMCMLandmanWJMInvestigation of the Efficiencies of Bioaerosol Samplers for Collecting Aerosolized Bacteria Using a Fluorescent Tracer. I: Effects of Non-sampling Processes on Bacterial CulturabilityAerosol Science and Technology2011453423431ZhenHHanTFennellDEMainelisGRelease of Free DNA by Membrane-Impaired Bacterial Aerosols Due to Aerosolization and Air SamplingApplied and Environmental Microbiology201379247780778924096426ZhenHHanTFennellDEMainelisGA systematic comparison of four bioaerosol generators: Affect on culturability and cell membrane integrity when aerosolizing Escherichia coli bacteriaJournal of Aerosol Science20147006779doi: http://dx.doi.org/10.1016/j.jaerosci.2014.01.002

Schematic diagram of the experimental setup showing testing with the Button aerosol sampler. Insert in the right bottom corner shows connections when testing with the BioSampler and the electrostatic precipitator with superhydrophobic surface (EPSS).

Bioluminescence intensity (RLU) as a function of ATP reagent and ATP standard storage time at −20 °C (1, 3, 5, 7, 9, and 14 days). Error bars denote one standard deviation, based on triplicate measurements.

Change in the amount of bioluminescence activity per bacterium (luminescence intensity per bacterium at certain time =i vs reference luminescence intensity per bacterium at time = 0) as a function of B. atrophaeus and P. fluorescens suspension storage time (0.08, 0.25, 1, 2, 3, 4, 5, and 6 hours). Error bars denote one standard deviation, based on triplicate measurements.

Effect of aerosolization on RLU output per bacterium for B. atrophaeus and P. fluorescens as a function of aerosolization time (i.e., 1, 5, 15, and 60 min) and Collison jar material (i.e., glass and polycarbonate jars). The RR(Collison) ratio represents a dimensionless ratio of the ratio Ri = (RLU/NAOEM)i after a certain aerosolization t = i over the reference ratio R0 = (RLU/NAOEM)0 at t = 0. Error bars denote one standard deviation, based on triplicate measurements.

Comparison of the RR(sampler) ratio for three different samplers (i.e., BioSampler, Button sampler, and EPSS), two different sampling times (i.e., 15 and 60 min) and two different bacteria (B. atrophaeus and P. fluorescens). The RR(sampler)k,l ratio represents a dimensionless ratio of the ratio Rk,l = (RLU/NAOEM)k for a certain sampler k after a certain sampling time l over the reference ratio R0 = (RLU/NAOEM)0. The latter is determined in microbial suspensions before aerosolization. A Collison nebulizer with a polycarbonate jar was used in all experiments. Error bars denote one standard deviation, based on triplicate measurements.

Summary of the variables tested in this research.

VariablesVariable valuessTested effects onbioluminescenceintensity
Bacterial aerosol species

B. atrophaeus

P. fluorescence

 		Gram positive vs. gram negative bacterial speciesAerosolization stressBioaerosol sampling stress
Collison nebulizer jar material and aerosolizing time

Glass jar

Polycarbonate jar

 		1, 5, 15, and 60 min
Bioaerosol sampling method and sampling time

Button sampler (with 0.6 εm pore size polycarbonate filter, 0.5 εm pore size PTFE membrane filter, and 0.45 εm pore size cellulose nitrate filter)

BioSampler (5 and 20 mL of collection liquid)

Electrostatic precipitator with superhydrophobic surface (EPSS)

 		15 and 60 min
ATP reagent storage time at −20 °C1, 3, 5, 7, 9, and 14 daysProtocol development and quality control
Storage time of thawed out ATP reagent at 4 °C1, 2, 4, and 6 hours
Storage time of bacterial suspension0.08, 0.25, 1, 2, 3, 4, 5, and 6 hours
Research Highlights

Various factors affecting the ATP bioluminescence intensity were investigated

Optimal conditions for the ATP reagent storage and preparation determined

The bioluminescence intensity was significantly affected by mechanical stress

The bioluminescence intensity significantly depended on sampling method

Electrostatic sampler was least injurious to bacteria among investigated samplers