Article pubs.acs.org/est
Filterable Redox Cycling Activity: A Comparison between Diesel Exhaust Particles and Secondary Organic Aerosol Constituents Robert D. McWhinney,*,† Kaitlin Badali,† John Liggio,‡ Shao-Meng Li,‡ and Jonathan P. D. Abbatt† †
Department of Chemistry, University of Toronto, 80 St. George Street, Toronto, Ontario, Canada Air Quality Processes Research Section, Environment Canada, 4905 Dufferin Street, Toronto, Ontario, Canada
‡
S Supporting Information *
ABSTRACT: The redox activity of diesel exhaust particles (DEP) collected from a light-duty diesel passenger car engine was examined using the dithiothreitol (DTT) assay. DEP was highly redox-active, causing DTT to decay at a rate of 23−61 pmol min−1 μg−1 of particle used in the assay, which was an order of magnitude higher than ambient coarse and fine particulate matter (PM) collected from downtown Toronto. Only 2−11% of the redox activity was in the water-soluble portion, while the remainder occurred at the black carbon surface. This is in contrast to redox-active secondary organic aerosol constituents, in which upward of 90% of the activity occurs in the water-soluble fraction. The redox activity of DEP is not extractable by moderately polar (methanol) and nonpolar (dichloromethane) organic solvents, and is hypothesized to arise from redox-active moieties contiguous with the black carbon portion of the particles. These measurements illustrate that “Filterable Redox Cycling Activity” may therefore be useful to distinguish black carbon-based oxidative capacity from water-soluble organic-based activity. The difference in chemical environment leading to redox activity highlights the need to further examine the relationship between activity in the DTT assay and toxicology measurements across particles of different origins and composition.
■
INTRODUCTION Particulate matter (PM) is known to be a major contributor to adverse effects of air pollution on human health, both acute1−3 and chronic.4−6 It has been postulated that one of the major toxicological mechanisms is through the induction of oxidative stress.7,8 Redox-cycling reactions, wherein components of PM catalytically transfer electrons from reducing agents to dissolved oxygen, can both deplete antioxidants and generate reactive oxygen species (ROS) through the general reaction shown in Scheme 1. Participation of redox-active transition metals and quinone species in such reactions has been proposed as one path to oxidative stress induction.9,10
Diesel exhaust particles (DEP) are one class of relevant particles that have been studied using the DTT assay and identified as redox active.15,17,20−25 In addition to their activity in the DTT assay, superoxide generation has been observed in aqueous suspensions of DEP,26,27 and these particles have been shown to induce direct adverse effects on both cell tissue cultures20,28 and organisms.27,29−33 However, there is some discrepancy as to the nature of the chemical moieties that lead to the redox activity, ROS generation, and overall toxicity, particularly with respect to their solubility in water and organic solvents. There have been observations that the chemical species responsible for toxicity or ROS generation remain bound to the particle surface in aqueous solution,20,24,26,27 but the results for organic extractability of toxic components of DEP have been mixed. With respect to toxicology of both the black carbon particle and the organic-extractable fraction of DEP, both have been shown to be toxic and generate ROS in different manners, with particulate components tending to stimulate cellular release of ROS and organic components generating intracellular ROS.34 In one example of organicsoluble ROS generation, both superoxide generation and toxicity to mice by intratracheal instillation were substantially
Scheme 1. General Redox Cycle for Particulate Matter (PM) and a Reducing Agent (Red)
The dithiothreitol (DTT) assay is an acellular technique used specifically for measuring the redox cycling capacity of PM by measuring how fast DTT is oxidized in the presence of PM in aqueous solution.11 The results of the assay have been correlated to measures of oxidative stress12,13 and have been used to compare the potential toxicity of numerous ambient and laboratory particles.14−19 © 2013 American Chemical Society
Received: Revised: Accepted: Published: 3362
November 15, 2012 March 4, 2013 March 7, 2013 March 7, 2013 dx.doi.org/10.1021/es304676x | Environ. Sci. Technol. 2013, 47, 3362−3369
Environmental Science & Technology
Article
depressed when DEP were washed with methanol.27 In another, macrophage cells showed generation of reactive oxygen radical species in the presence of DEP and their methanol extracts, but not in methanol-washed particles.35 In contrast, organic extractions are sometimes found to not alter the toxicity (or surrogate properties of toxicity) of the black carbon particle core. One study found that the ability of DEP to generate superoxide in the presence of an added reducing agent such as DTT or ascorbate was unaffected when particles were washed with dichloromethane, suggesting that redox-cycling reactions in particular arise from compounds that are either very strongly adsorbed to or chemically contiguous with the particle surface.26 Additionally, the DTT decay rate of a dichloromethane extract of diesel exhaust particles has been illustrated to be only about 28% of the activity of the full aqueous suspension.24 In previous work, we found that the species responsible for the redox activity of a DEP standard reference material are associated with the particle surface under aqueous extraction; filtering insoluble material greatly reduces the redox activity of these particles and the activity remains at the particle surface after repeated aqueous washes.20 While many of the toxicity studies have used methanol washes to extract organic constituents from DEP, to our knowledge the effect of a moderately polar organic solvent washing on DEP redox activity has not yet been examined. The effect of the engine operating conditions on DEP toxicity is also important to understanding the chemical nature of the toxic mechanism. Physicochemical properties of DEP have been demonstrated to change under different engine operating conditions, with lower engine loads being associated with higher ratios of organic to elemental carbon.36 There is some evidence in direct toxicity studies that DEP toxicity changes when engine operating conditions fluctuate; higher engine loads have been found to be associated with larger toxicological effects when normalized to particle mass.30 With respect to redox cycling, however, there are few data on how engine operating conditions change the activity of DEP, with one study noting similar redox cycling under load and no-load conditions.25 In this study, we examine the redox-cycling activity of DEP collected from a light-duty diesel engine equipped with a diesel oxidation catalyst (DOC) and operated under three different driving modes, representing low-, medium-, and high-load modes during the Diesel Engine Emission Research Experiments 2012 (DEERE 2012). Given the conflicting reports of previous studies examining DEP redox activity and respiratory toxicity, we aim to examine the nature of the redox-active species in DEP and their extractability in water, methanol, and dichloromethane to determine if the active redox-cycling species remain strongly bound to the black carbon surface in solvents with a range of polarities. We also compare the nature of the DEP to other particles, including nondiesel-derived black carbon and secondary organic aerosol (SOA) systems, with respect to their solubility behavior. Finally, we examine if significant changes in the redox activity of these particles are observed under the three different engine operating conditions. We suggest that the “Filterable Redox Cycling Activity” may be a means of establishing the degree to which DEP contributes to the rates of oxidative redox cycling.
of DEERE 2012. A diesel enginea Volkswagen TDi engine recovered from a Volkswagen Jetta passenger vehicle and mounted on a dynamometerwas operated under three steady conditions representing the average output of the engine under low-, medium-, and high-load driving conditions. A DOC aftertreatment was used, which removes a portion of semivolatile hydrocarbons from the exhaust.37 The engine operating conditions are listed in Table 1. An ultralow sulfur diesel fuel Table 1. Summary of Engine Operating Mode Conditions operating mode
RPM
torque/ft lb
high-load medium-load low-load
2000 1800 1600
72 60 40
was used, and the engine was allowed to run for approximately one hour to reach a steady-state prior to collection of filter samples. Particles were drawn onto 47-mm Teflo filters (Pall Corporation, polytetrafluoroethylene [PTFE], 2 μm pore size) and were sampled directly from the output of a Constant Volume Sampler (CVS), which takes in the engine exhaust, mixes the exhaust with filtered laboratory air and maintains a constant volume flow [see ref 38 for details]. Filter samples were collected at a flow rate of 50 L per minute for a total of 30−60 min. The DEP samples were kept in a freezer at −20 °C until extraction. The PTFE filter samples were extracted into 20 mL of phosphate buffer (0.1 M, pH 7.4) by sonication for 15 min before being put through the DTT assay.39 Some particles remained clumped on the filter during extractions, so in all cases the filter was removed following extraction, rinsed gently in deionized water, dried, and reweighed to find the extracted DEP mass. After extraction, the samples were refrigerated until the assay was performed. Aliquots of these concentrated DEP suspensions were diluted to a 1-mL volume to yield an appropriate DTT decay rate (less than approximately 40% decay from the initial DTT concentration) over a 60-min time period. DTT (Sigma Aldrich) was added to the diluted particle suspensions (final DTT concentration of 0.1 mM) and the reaction was quenched at 15-min intervals over 60 min by mixing a 0.25 mL aliquot of the reaction mixture with 1 mL of 0.25 mM 5,5′-dithiobis-(2-nitrobenzoic acid) (Alfa Aesar). In contrast to most studies, the assay is performed at room temperature. DTT concentrations were quantified by absorbance of the resulting product at 412 nm and were normalized to the mass concentration of particles in the assay. Absorbance calibration curves were calculated from standard DTT solutions and calculated each time the assay was performed. DTT decay rates for the engine operating conditions are reported as the average ± one standard deviation for triplicate filters. Redox activity of each filter extract was measured 4−5 times. Blank corrections were performed with a phosphate buffer blank each time the assay was performed; average blank DTT decays were typically 0.06 ± 0.03 μM min−1. To determine the total redox activity of the water-soluble fraction, the particle suspensions were passed through a 0.2-μm PTFE syringe filter. After a 1-mL wash, the filter did not change the DTT activity of pure 1,2- and 1,4-naphthoquinone or 9,10phenanthrenequinone solutions. This result verifies that the filter does not sorb water-soluble redox-active organic species. The water-soluble portion of the DEP extract was carried through the DTT assay without further dilution.
■
EXPERIMENTAL SECTION DEP filters were collected at Environment Canada’s Environmental Technology Centre in Ottawa, Ontario, Canada as part 3363
dx.doi.org/10.1021/es304676x | Environ. Sci. Technol. 2013, 47, 3362−3369
Environmental Science & Technology
Article
For the activity of the organic extractable fraction, one filter of each driving type was divided into quarters; one segment was sonicated with 10 mL of dichloromethane or methanol and filtered using a PTFE syringe filter. After the organic solvent was evaporated, the residue was reconstituted in phosphate buffer and carried through the DTT assay. An additional highload DEP filter was carried through the methanol extraction to measure particle-bound redox activity; the filter was washed and sonicated in methanol and the particles were isolated by centrifugation and decanting the supernatant. The particles were washed in this manner three times and resuspended in 10 mL of phosphate buffer before DTT assay analysis. The same measurement was attempted for a dichloromethane extract, however the measurement was not made due to difficulty decanting the supernatant without a large loss of particles. Organic carbon (OC) and elemental carbon (EC) of the water-insoluble portion of aqueous suspended particles were measured using a Sunset Laboratory OC-EC analyzer. The Teflon filter samples were extracted in water and pulled through quartz fiber filters (Whatman) for analysis. Several other particle types were used for comparison and were prepared in phosphate buffer in a manner similar to the DEP suspensions. First, a commercial black carbonRegal 400R (Cabot Corporation, Pampa, TX, USA), a furnace carbon black for use as a pigmentwas used as a contrasting black carbon from a nondiesel source. Second, oxidized two-stroke engine exhaust (without removal of VOCs)16 was collected on Zefluor filters (Pall Corportation, PTFE, 2 μm pore size) for a mixed primary and secondary organic particle source. Third, we collected chamber secondary organic aerosol from hydroxylradical initiated oxidation of naphthalene. The naphthalene oxidation will be described in more detail in forthcoming work,40 but briefly, the particles were generated without a seed aerosol by photooxidation of naphthalene and hydrogen peroxide in a PTFE chamber and collected on Zefluor filters. And last, ambient coarse (2.5 μm < dp < 10 μm) and fine (0.3 μm < dp < 2.5 μm) particles from downtown Toronto were collected on Zefluor filters using previously described ambient particle concentrators.41,42 Ambient particles were collected in late February/early March; average redox activities were calculated based on nine coarse samples and eight fine samples.
The DEP were highly redox-active under every driving condition examined, with average DTT consumption rates ranging from 23 to 61 pmol of DTT per minute per μg of DEP (Figure 1), depending on the engine conditions. As has been
Figure 1. Average mass-normalized DTT decay rates; error bars are ± one standard deviation. Shown, left to right, are DEP from high-, medium-, and low-load driving conditions; two standard reference materials, SRM2975 (DEP) and SRM1648a (urban PM);20 average activities for coarse and fine particles collected in downtown Toronto; naphthalene-derived secondary organic aerosol (NSOA; mass loading >100 μg m−3); and a commercial black carbon Regal 400R.
previously observed,20,24 the redox activity in the aqueous particle suspensions is associated with the particles themselves; filtration of particles from the extract removed most (89−98%) of the activity in the DTT assay (see Figure 2). The amount of water-soluble redox activity did not increase over time, suggesting that there was no slow dissolution of redox active constituents.
■
RESULTS AND DISCUSSION Redox Activity of DEP and Carbon Black. Linear decay of DTT over a 60-min time period was observed in all diesel exhaust particle suspensions. We are interpreting this loss of DTT as occurring through catalytic redox cycles with active species in the DEP, but we note that we cannot fully rule out other noncatalytic mechanisms of loss, such as slow adsorption or reaction with the surface. We view these processes as unlikely given that the number of active surface sites on the soot is likely much lower than the total number of DTT molecules lost. As well, these processes would have to be irreversible over the time scale of several hours for the DTT not to react with the DTNB reagent and be measured by absorbance. Furthermore, superoxide generation has been observed by DEP in the presence of DTT, while electron paramagnetic resonance measurements have suggested the presence semiquinone-like radicals on the surface of DEP, which are likely to be redox active.26 The DTT assay activity from DEP has, in the past, been interpreted as a redox cycling mechanism (e.g., 24), and so in the remainder of the paper we will continue to use this mechanistic interpretation.
Figure 2. Water-soluble fraction of redox activity for eight particle classes: a commercial black carbon, Regal 400R; high-, medium-, and low-load DEP; two standard reference materials, SRM2975 (DEP) and SRM1648a (urban PM);20 ozone-oxidized two-stroke engine particles (engine/O3); and naphthalene-derived secondary organic aerosol (NSOA).
To determine how much of the redox activity arose from organic-extractable species, each driving condition DEP sample was extracted with both methanol and dichloromethane. The DTT assay was carried out on the water-soluble residue of the filtered extracts after evaporation of the solvent. Neither the methanol nor dichloromethane extracts exhibited observable 3364
dx.doi.org/10.1021/es304676x | Environ. Sci. Technol. 2013, 47, 3362−3369
Environmental Science & Technology
Article
reproduce from run-to-run, either from differences in how well flocculated particles broke apart during sonication, or how quickly particles recoagulated after sonication. As a result of this variability, limited conclusions can be drawn for how redox activity is influenced by the driving condition. The DEP collected during the high-load driving mode were the most redox-active, followed by the low-load driving mode, and the least amount of activity was exhibited under the medium-load conditions (averages shown in Figure 1). This is partially consistent with observations that diesel particles are more toxic under higher loads,30 as the high-load driving mode is the most redox-active DEP. However, without measuring the surface area of the DEP in the aqueous suspensions, it cannot be determined whether there are differences in the chemical environment on the surface or if a simple physical metric such as surface area per unit mass (or a combination of the two) are driving this effect. Chemical environment does not seem to account for differences in redox activity; the trend observed for gas phase NOx emissions and CO/CO2 ratios, as well as OC/EC ratios of the insoluble DEP, all follow trends corresponding to engine torque (Figure 3; CO/CO2 and OC/EC ratios decrease with increasing engine torque, while NOx emissions increase with increasing torque). Intrafilter replicates of the DTT assay were performed on the same extract up to 35 days after the initial extraction of the filter, which allowed examination of time trends in the redox activity. There is the potential for the DTT activity to follow time-dependent trends following initial extraction. For example, if redox-active moieties in the DEP decayed over time, redox activity would be expected to fall over time; on the other hand, DTT activity may increase with extraction time if active sites on the particle take time to be wetted and become more accessible. However, as presented and discussed in the Supporting Information, there is no evidence of a clear temporal trend. One noteworthy aspect of the time trends, however, is that DEP displayed considerable inherent redox activity even one month after aqueous extraction. This suggests the moieties that catalyze redox activity have the potential to be long-lived and could be important for a considerable time after particles are inhaled, should redox cycling be an important mode of toxicity for the particles. There is a potential for these moieties to behave differently in the body however, given the lack of refrigeration to preserve any potentially unstable species, and the potential for modification of the particle surface through the many additional biological reagents available. We conclude that for the DEP, redox-cycling activity, as measured by the DTT assay, is established primarily through reactions occurring at the black carbon surface that cannot be removed by a wide range of polar and nonpolar solvents. At this point, we cannot determine the exact nature of the active moieties from this study. One possibility is that small, typically water-soluble species such as quinones or transition metals are present in the DEP but are so strongly adsorbed to the particle that even strong organic solvents are unable to wash them away. Since strong organic solvents have been used to extract quinone species from the DEP for quantification,39,43 it seems somewhat unlikely that small redox-active quinone species such as naphthoquinone or phenanthrenequinone are responsible. We have also previously determined that a standard reference material DEP exhibited high redox capacity despite very low concentrations of transition metals,20 indicating that black carbon itself is likely acting as the active species. Another possibility is that quinone-like moieties, which have been
DTT decay above that of the blank for DEP from any of the three driving conditions or the Regal 400R carbon black (corresponding to an upper-bound DTT decay rate of
