Rapid conversion of isoprene photooxidation products in terrestrial plants

Isoprene is emitted from the biosphere into the atmosphere, and may strengthen the defense mechanisms of plants against oxidative and thermal stress. Once in the atmosphere, isoprene is rapidly oxidized, either to isoprene-hydroxy-hydroperoxides (ISOPOOH) at low levels of nitrogen oxides, or to methyl vinyl ketone (MVK) and methacrolein at high levels. Here we combine uptake rates and deposition velocities that we obtained in laboratory experiments with observations in natural forests to show that 1,2-ISOPOOH deposits rapidly into poplar leaves. There, it is converted first to cytotoxic MVK and then most probably through alkenal/ one oxidoreductase (AOR) to less toxic methyl ethyl ketone (MEK). This detoxification process is potentially significant globally because AOR enzymes are ubiquitous in terrestrial plants. Our simulations with a global chemistry-transport model suggest that around 6.5 Tg yr− of MEK are re-emitted to the atmosphere. This is the single largest MEK source presently known, and recycles 1.5% of the original isoprene flux. Eddy covariance flux measurements of isoprene and MEK over different forest ecosystems confirm that MEK emissions can reach 1–2% those of isoprene. We suggest that detoxification processes in plants are one of the most important sources of oxidized volatile organic compounds in the atmosphere.


Experimental Design
The experimental setup is illustrated in Supplementary Figure 1. Gray poplar plants (see Methods -plant material) were placed in an enclosure system consisting of a glass desiccator (Schott Duran ©) of 17.3 L volume turned upside-down. The inner surface of the desiccator was coated with Teflon (PFC 801A, Cytonix, USA) in order to minimize surface deposition of oxidized VOCs (OVOCs). The enclosure was placed on two PTFE ground plates equipped with a groove and tongue and ports for air inlet and air sampling, respectively. The plant stem and a Teflon coated K type thermocouple were fed through a central notch in the plates. Possible leaks were sealed with a Teflon tape. A Teflon coated fan (Propeller Stirrer Shaft, 6.5 mm chucking diameter, Bohlender GmbH, Grünsfeld, Germany) connected to a 12 VDC motor was used for turbulent air mixing inside the enclosure. For the entire setup, only chemically inert materials such as PTFE, PFA and PEEK were used. All tubing was light-shielded with pipe insulation to exclude unwanted photolysis of 1,2-ISOPOOH. The plants were fumigated with synthetic air (5.0 grade, Messer Austria GmbH, Gumpoldskirchen, Austria) that was mixed with CO 2 (4.8 grade, Messer Austria GmbH, Gumpoldskirchen, Austria) resulting in a CO 2 volume mixing ratio of on average ~450 ppm. Before entering the enclosure, the air was flushed through a liquid calibration unit (LCU, see below) (Ionicon Analytik, Innsbruck, Austria) to humidify the air and to add a defined quantity of a solution of a synthetic ISOPOOH standard (1) in deionized water (2.9 µL ISOPOOH in 100 mL deionized water (v/v)). The ISOPOOH standard consisted entirely of the 1,2-ISOPOOH isomer (kindly provided by the Frank Keutsch group, Harvard University, Boston MA) (1). Air composition at the inlet and outlet of the enclosure was alternately analyzed with the SRI-ToF-MS (see below) and an infrared-gas-analyzer (IRGA, see below).
Characterization of 1,2-ISOPOOH deposition to the empty enclosure In order to quantify possible surface deposition and decomposition of 1,2-ISOPOOH on the Teflon-coated surface of the empty glass enclosure and on Teflon tubing material, we performed 1,2-ISOPOOH fumigation experiments with the empty enclosure. Before starting the plant fumigation experiments, we fumigated the empty cuvette with 1,2-ISOPOOH to condition the inner surfaces. Subsequently, the 1,2-ISOPOOH loss to the surface of the empty enclosure was measured for each individual experiment. The enclosure was flushed with humidified air (RH ~35 %) at room temperature containing 7.8±1.0 ppb 1,2-ISOPOOH. For an estimation of the 1,2-ISOPOOH loss to the empty enclosure, we modelled the deposition rate to the surfaces according to (2): (1) where ?@,ABCDCCE [nmol mol -1 ] is the volume mixing ratio (VMR) of 1,2-ISOPOOH determined in the enclosure inlet air, F'G,ABCDCCE [nmol mol -1 ] is the VMR of 1,2-ISOPOOH measured at the enclosure outlet and τ [s] represents the residence time for a single exchange of the air in the enclosure (see Supplementary Table 1 and Supplementary Figure 2 for an overview). Under well-mixed conditions, which were achieved with the Teflon fan, the residence time can be expressed as the ratio of the enclosure volume #@+JF&'(# to the enclosure inlet gas flow (3): According to (4) it takes 5 τ to exchange 99% of the air in the enclosure. The residence time in our fumigation experiments was on average 4.5 min, thus requiring 22.5 min to reach 99% air exchange.
Supplementary Figure 2 depicts a typical change in 1,2-ISOPOOH deposition rate "#$,&'()*+# over time. At the beginning of the experiment, when surfaces are virtually free of 1,2-ISOPOOH, "#$,&'()*+# is 0.25 molecules s -1 . Subsequently, the deposition rate decreases exponentially over time. 60 min after starting the 1,2-ISOPOOH fumigation the deposition rate becomes smaller than 0.01 molecules s -1 and is neglected for further analysis. The characterization of the empty enclosure system revealed a far slower adaption to new experimental conditions for highly water-soluble compounds such as 1,2-ISOPOOH compared to more volatile and less soluble compounds. For example, MVK and MEK are typically adjusted after 23 minutes, which is the time required to completely (> 99 %) exchange the gas in the enclosure. We used equation (1) to estimate the time required to reach steady state conditions in our enclosure setup. The 1,2-ISOPOOH deposition rate to the enclosure surface (k dep,wall ) becomes negligible after 91 min, corresponding to 20 residence times. In the case of MVK this occurs after 23 min (5 residence times). All experiments were performed in such a way that ample time was allowed in order to reach steady-state conditions before analyzing deposition rates of 1,2-ISOPOOH to plants. To minimize the effects of 1,2-ISOPOOH reemission from surfaces due to changes in humidity, only steady-state conditions were considered in each respective step for further analyses.
Photolytic loss of 1,2-ISOPOOH We tested possible photolysis losses of 1,2-ISOPOOH (caused by irradiation) in the empty enclosure. For this purpose, we fumigated the empty enclosure with 1,2-ISOPOOH in darkness.
Once the 1,2-ISOPOOH signals had stabilized after ~120 minutes, the light was switched on. We observed, however, no effect of the radiation on the 1,2-ISOPOOH volume mixing ratios in the enclosure.
Four-step protocol for the 1,2-ISOPOOH fumigation experiment Each poplar plant was installed in the enclosure system shown in Supplementary Figure 1 several hours before starting an individual experiment, allowing the plant to adapt to the enclosure environment and recover from possible stress during the installation process. Before starting the fumigation run, we measured the "default" emissions of the gray poplars during darkness and illumination. We performed 1,2-ISOPOOH fumigation experiments following a four-step protocol for each plant (Supplementary Table 2). During step (A), 1,2-ISOPOOH from the liquid calibration unit (LCU) was analyzed directly with the SRI-ToF-MS via a bypass system, while the plant enclosure was fumigated with catalytically generated clean air (zeroair, ). During steps (B), (C) and (D), the SRI-ToF-MS sampled air at the enclosure outlet. At step (B), a poplar plant was fumigated with 1,2-ISOPOOH, typically for 9-12 hours, under dark conditions (no light), followed by fumigation of the illuminated plant (step C) for another 9-12 hours. After performing steps (A), (B) and (C), the plant was removed from the enclosure in order to conduct background measurements of the empty enclosure (step D). An asterisk following the step label (e.g. A*) indicates that the air mixture was passed through a 1 m × ¼ inch (length × diameter) stainless steel tube kept at room temperature before being analyzed by the SRI-ToF-MS. As reported previously (1, 5, 6), under these conditions 1,2-ISOPOOH is converted efficiently to MVK and C 5 -diols. For further analysis, we used the OVOC data averaged over 30-60 min at the end of each experimental step when OVOC signals had reached steady state conditions.
Representative poplar fumigation experiment Supplementary Figure 3 depicts a typical fumigation cycle for a gray poplar plant, performed according to the four-step protocol. In step A, we enriched the air with 7.8±1.0 ppbv of 1,2-ISOPOOH across all replicates. The air flow leaving the LCU also contained 2.0±0.4 ppbv MVK and 0.2±0.1 ppbv C 5 -diols as contaminants in the 1,2-ISOPOOH standard. Comparing the 1,2-ISOPOOH concentration in the outlet air of the fumigated empty enclosure (step D) with the corresponding 1,2-ISOPOOH values measured in the inlet air before the plant fumigation experiment (step A) revealed only minor (if any) losses to the enclosure system. On average only 6±5% of the input gaseous 1,2-ISOPOOH was lost to the enclosure surface when equilibration times of more than 2 hours were permitted. Upon starting the fumigation of the plant under dark conditions (step B), the mixing ratio of 1,2-ISOPOOH increased slowly over the course of several hours, reaching steady state after 8-10 hours. The mixing ratio of MVK reached steady state conditions (equal to the input concentration) after approximately 22 min, which is consistent with the time it takes for >99% air exchange of the enclosure (5τ ≈ 22 minutes). Instantaneously upon illumination the concentration of 1,2-ISOPOOH started to decrease. At the same time, the MVK signal showed a slight initial burst, followed by a decrease in emission, which leveled off to a ~1.5-fold increase from the original MVK concentration. MVK emissions were accompanied by a simultaneous increase in the MEK signal. When passing the air flow through metal tubing in steps A*, B*, C* and D*, the MVK and C 5 -diol signal increased due to 1,2-ISOPOOH conversion on metal surfaces, while the MEK signal remained unchanged.
Calculation of emission dynamics The deposition velocity " (cm s -1 ) is commonly used to describe trace gas deposition to vegetation from the atmosphere (7), and is defined as the ratio between the flux Φ ? (representing the amount of compound deposited to a unit surface area per unit time) and the local concentration ? .
Similarly, as described in (8), for enclosure measurements the deposition velocity ",? for compound can be estimated from the flux Φ ? to the system (plant + enclosure surface) and the concentrations ?,F'G measured at the enclosure outlet: To take into account possible losses to the enclosure surfaces, we adapted the commonly known formula for emission rates (Equation 5) to the following form (Equation 6) which follows the considerations of (11): where F'G,?,ef is the volume mixing ratio of substance at the enclosure outlet during fumigation of the empty enclosure (step D, see below). Both background and plant experiments were performed under identical conditions in terms of 1,2-ISOPOOH mixing ratios and relative humidity in the inlet air. This background correction minimizes any potential errors caused by an underestimation of surface sinks.
Liquid Calibration Unit (LCU) A liquid calibration unit (LCU, Ionicon Analytic GmbH, Innsbruck, Austria) allows quantitative evaporation of liquid standards into a gas stream to generate calibration mixtures.
Here an LCU was used to evaporate an aqueous solution of 1,2-ISOPOOH into a synthetic air stream fed into the plant enclosure. The liquid flow of the 1,2-ISOPOOH solution was regulated and kept constant at 10 µL min -1 . Additionally, 40 µL min -1 of purified water were introduced into the synthetic air stream resulting in a relative humidity of approximately 35% at room temperature. The temperature in the evaporation chamber of the LCU was set to 50°C to avoid thermally-induced dissociation of 1,2-ISOPOOH.  (12). 1,2-ISOPOOH decomposes at elevated temperatures (e.g., during GC analysis; (1)) or on stainless steel surfaces even at room temperature (6) to MVK and 3-methyl-1-butene-3,4-diol (C 5 -diol, (5) (14). To distinguish MEK from isomers such as butanal and 2-methyl propanal, we used the reagent ion NO + . Our instrument allows fast switching between different reagent ions. According to (15) 4 + , 1,2-ISOPOOH) were below the detection limit. During plant enclosure experiments observed ion signals are substantially higher than these back-ground signals.

Supplementary Methods -Construction of AOR Phylogenetic Trees and AOR Gene Expression Analysis
To detect AOR genes, sequence similarity searches (BLAST; e-value cutoff 1e-4) were conducted against the NCBI nr, dbEST and SRA databases using known amino acid sequences of genes encoding the chloroplast and cytosolic plant AOR proteins. Reciprocal BLAST searches were conducted to ensure that AOR represents the closest homologs to the identified bacterial and Bryophyta sequences in Angiosperms. To recover AOR sequences from Quercus robur and Fagus sylvatica, BLAST searches (e-value threshold 1e-30) were conducted against a local database containing whole genome and predicted protein sequences for these species (16,17). Sequences for Picea abies and Pinus taeda were obtained from Congenie (http://congenie.org); Salix purpurea AOR sequences were obtained from Phytozome (https://phytozome.jgi.doe.gov). Multiple sequence alignments containing plant and bacterial (Supplementary Figure 10) and only plant ( Figure 3) AOR genes were constructed using MUSCLE (18) and refined manually using Mesquite v. 3.51 (19). Maximum Likelihood (ML) AOR phylogenetic trees were reconstructed using RaxML (20) with bootstrapping (100 bootstrap replicates) and the LG amino acid substitution model. To discriminate between genes encoding cytosolic and plastid-targeted AOR, a targeting peptide prediction was conducted using TargetP 1.1 Server (21). Expression information for the gray poplar chloroplasic and cytocolic AOR genes was derived from (22). Only samples collected in the light phase from plants maintained under normal control scenarios (AC: ambient CO 2 ; EC: enhanced CO 2 ) conditions were considered for analysis.

Supplementary Methods -Eddy Covariance VOC Flux Measurements
Calibration and data analysis of the PTR3 Isoprene and MEK sensitivities were calibrated as a function of humidity using a gas standard (Apel-Riemer Inc., Broomfield, USA), which was diluted in air with changing humidity. The humidity was varied from dry (synthetic air bottle) to tens of ppth during calibration covering the typical ambient humidity conditions. Water transport in eddies causes the sample humidity to correlate with vertical wind speed. This can induce artifacts in EC measurements if the analyzer shows humidity-dependent sensitivity. In order to accurately calibrate the PTR3-TOF VOC signals, a fast water-sensitive tracer (N 2 H + ), which is produced in the PTR3-TOF, was regularly cross-calibrated against ambient humidity measurements with the IRGA (time resolution of seconds). This calibration of the PTR3-TOF signals was done at 10 Hz, resulting in a fast humidity trace. Data analysis was performed using multi-peak analysis routines (23), which only rely on single-ion counting within specified mass ranges and a subsequent correction of cross talk from neighboring mass peaks. This is important since typical mass spectra show fully developed peak shapes at the 10 Hz acquisition rate that can be peak-fitted. The resulting time-traces were calibrated as described above and used as input for "InnFlux", an eddy covariance flux routine developed by the group of Thomas Karl at the Faculty of Geoand Atmospheric Sciences, University of Innsbruck, Austria.

Statistical analysis
Biological replication: for enclosure measurements with 1,2-ISOPOOH fumigation we used gas phase data for five biological replicates (for plant details see Supplementary Table 4). For AOR analysis we took samples from seven and five poplars (ISOPOOH and MVK fumigation, respectively). From each extract, three technical replicates were analyzed.

Supplementary Figure 1. Experimental design of enclosure measurements
A poplar plant was placed in the enclosure equipped with a Teflon fan for turbulent mixing. Gas flows of synthetic air and CO 2 were controlled by mass flow controllers (MFC). The liquid calibration unit (LCU) humidified the air stream (RH 35% at room temperature) and added ~7 ppbv 1,2-ISOPOOH. Switching valves directed the gas stream either before or after the enclosure to the SRI-TOF-MS with or without passing through a metal line. CO 2 and H 2 O concentrations were analyzed with an infrared gas analyzer (IRGA) either before or after the enclosure. Plants were illuminated with a true light lamp. Infrared radiation was blocked by a water bath. The enclosure could be flushed with zero air (0-air). replicates. The enzyme assays were performed according to (25).

Supplementary Figure 10. AOR phylogenetic tree
Evolutionary history of genes encoding plastid and cytosolic AOR proteins. The AOR phylogeny was reconstructed using the maximum likelihood (ML) method and LG amino acid substitution model. The bacterial clade (black color) was used as outgroup. Gray coloring of the branch labels indicates genes encoding cytosolic AOR; green coloring indicates genes encoding choroplast AOR. Numbers at the tree branches indicate node bootstrap support. Basal nodes of the plant AORchl, AORcyt-I and AORcyt-II and bacterial NADPH: quinone reductase ortholog clusters are labeled with solid circles. Scale bar below the tree shows branch length. This phylogeny implies a bacterial origin of the plant AOR genes and several gene duplication events in Embryophyte. The first duplication event, which gave rise to the chloroplast and cytosolic AOR gene copies, is likely to have occurred during the early evolution of land plants.
An additional cytosolic AOR copy has arisen via duplication of the chloroplast AOR copy in Angiosperms.