Experimental evidence for direct sesquiterpene emission from soils



[1] The biogenic organic precursors which are of relevance in secondary organic aerosol formation are primarily isoprene, monoterpenes, sesquiterpenes (SQTs) and their derivatives. Atmospheric volatile organic compound inventories consider plant emissions as the exclusive sources of natural secondary organic aerosol precursors. Very recently we have shown in pure culture experiments that abundant soil fungal strains are capable of producing significant emission fluxes of SQTs and have implied that soils might be globally important SQT emission sources. In the present follow-up study we determined sesquiterpene emission fluxes in the laboratory directly from unperturbed soil samples and from selected soil samples with the top litter layer removed. We also characterized basic soil parameters and native fungal strains in the samples, and quantified ergosterol as a proxy of bulk fungal biomass. The measured SQT emission fluxes (6–1980 ng m−2 h−1; median 109 ng m−2 h−1) were of the same magnitude as inferred from culture experiments confirming that soil biota can be an important sesquiterpene emission source that has to be included in future VOC inventories.

1. Introduction

[2] Sesquiterpenes (SQTs) are considered to be important biogenic secondary organic aerosol (SOA) precursors in the atmosphere among the volatile organic compounds emitted by vegetation [Griffin et al., 1999; Duhl et al., 2008]. The main source of sesquiterpenes in global VOC inventories is the emission from coniferous vegetation [Guenther et al., 1995], though a recent study estimated somewhat lower emission factors for the region of the U.S. [Sakulyanontvittaya et al., 2008]. However, in none of these studies microbial biomass in soils were ever considered to be sesquiterpene emission source. Motivated by the discovery that several sesquiterpenes were detected among the typical metabolites of abundant microscopic fungi in indoor air quality studies [Kuske et al., 2005], we have very recently quantified sesquiterpene emission fluxes from cultures of several fungal strains in a flow-through apparatus in the laboratory [Horváth et al., 2011]. Sesquiterpenes are secondary metabolites in fungi, the role of which have not been fully understood. For certain other metabolites, such as mycotoxins or antibiotics the advantage of production is obvious. In other cases secondary metabolism may eliminate harmful intermediers resulting from primary metabolism. Other known secondary metabolites of fungi can be glucose derivatives, aromatic compounds, polyacetides, polyacetilenes, nitrogen-containing compounds and steroids. By assigning a given average fungal biomass for soils we inferred by back-of-the-envelope calculations that sesquiterpene emission from soil fungal communities may not be negligible compared to the estimated emission from abundant coniferous vegetation in a selected region and time of the year [Helmig et al., 2006]. However, it was highly uncertain whether emission factors measured from pure culture measurements can be applied for natural soils. We implied the possibility that sesquiterpenes as highly reactive compounds of moderate volatility might not escape from soils even if they are produced there by microscopic fungi. In other terms, the question was whether the rate of production under ambient conditions can exceed the rate of potential losses in soils. The primary concerns for potential losses are the adsorption on soil particles followed by metabolism by soil microorganisms and oxidation reactions in soil interstitial air. The aim of the present follow-up study is to determine sesquiterpene emission fluxes directly from undisturbed soil samples and relate these fluxes to measured soil and microbial factors with a view to establish the potential significance of soils as sesquiterpene emission sources in global VOC inventories.

2. Sampling and Analysis

2.1. Soil Sampling and Analysis

[3] 32 topsoil samples were collected in Veszprém County (Hungary) with a sampler designed for this purpose. The sampling sites and conditions are listed in Table 1. The upper 3 cm of soil layers were sampled, the diameter of the sampled soil cakes was 9 cm. At each sampling site four samples were taken, two of them undisturbed (native, with vegetation and/or litter layer on the surface), and the other two with the vegetation and topmost litter layer removed with tweezers on the site prior to sampling. The samples were immediately placed into Petri dishes of exactly the same size and were taken to the laboratory within less than an hour in all cases. One of each sample pair was placed into the same aseptic flow-through apparatus that was used previously for measuring SQT emissions of pure fungal cultures [Horváth et al., 2011]. The samples were thermostated to their typical soil temperature in order to minimize thermal stresses during the measurements. Laboratory measurements started immediately after equilibration of the samples, in order to maintain ambient sample integrity and conditions. Despite all precautions, it might be possible that some changes in microbial activity occurred during sample collection, handling and measurements. However, disturbance of the natural systems cannot be avoided in the case of in situ (e.g., enclosure) measurements, either.

Table 1. Soil and Vegetation Types, Sampling Details as Well as the Physical and Chemical Properties of the Soil Samplesa
Soil TypebLocation and GPS CoordinatesDate of SamplingBulk Density (g cm−3)Water-Holding Capacity (%)pH (H2O)OC (g kg−1)TN (g kg−1)
  • a

    OC: organic carbon content, TN: total nitrogen content.

  • b

    Vegetation is given in parentheses.

Cambisol (sunflower)Veszprém 47°04.288′ 17°52.756′7 June 20111.4349.
Cambisol (barley)Veszprém 47°04.288′ 17°52.793′25 May 20111.3647.46.814.11.5
Chernozem (pea)Csajág 47°02.63′ 18°09.32′18 July 20111.2261.76.828.13.3
Chernozem (barley)Csajág 47°02.63′ 18°09.18′18 July 20111.2964.46.929.52.8
Leptosol (oak)Királyszentistván 47°05.507′ 18°01.989′25 May 20111.0441.27.441.73.8
Leptosol (hornbeam and oak)Vörösberény 47°02.917′ 17°59.967′25 May 20111.0843.
Luvisol (pasture)Litér 47°05.809′ 18°01.207′7 June 20111.4759.65.924.32.3
Luvisol (weed)Vörösberény 47°02.917′ 17°59.982′7 June 20111.2652.

[4] For the microbial characterization of backup soil samples Penicillin (100 μg g−1) and Streptomycin (150 μg g−1) were added to each microcosm for selective inhibition of soil bacteria. Bulk density of the soils was calculated as the dry weight (determined by drying at 105°C to constant weight) of soil divided by its volume. For determination of water holding capacity closed funnels were filled with the soil samples, surplus water was added and the funnels were kept at room temperature for 30 min. After removing the plug, surplus water was drained for 30 min. Water content was determined as the mass loss by drying at 105°C to constant weight. Soil pH was measured in soil-water suspensions (1:1 soil to water ratio) using a combination glass electrode. Organic carbon and total nitrogen content of the soils were determined using an elemental analyzer (FISONS Instruments NA 1500 NC).

2.2. Estimation of Fungal Biomass in Soil Samples

[5] The concentrations of ergosterol were also measured in the collected soil samples and in their aliquots. The concentrations of this endogenous sterol were used to estimate the total fungal biomass in the samples since ergosterol is specific to fungal cell membranes (besides certain microalgae and protozoa) and cannot be found in plant tissues [Weete, 1989; Newell, 1992]. Several methods have been developed to extract and measure ergosterol in soils. Alkaline or non-alkaline extraction can be used for this purpose [Martin et al., 1990]. Ruzicka et al. [1995]combined non-alkaline extraction with ultrasonic treatment and a simplified method was developed byGong et al. [2001] using only methanol as extractant and disrupting the fungal mycelium with glass beads. It is not yet clear whether ergosterol measurements can be used to distinguish between living and dead fungal mycelia. Despite the uncertainties the ergosterol method is commonly used as a standard procedure to estimate fungal biomass in soil. Montgomery et al. [2000] proposed a conversion factor of 250 μg dry fungal biomass μg−1 ergosterol based on the average ergosterol concentration of six fungal species.

[6] The mass concentrations of ergosterol were determined from the backup soil samples according to the method of Bååth [2001]. As a small modification of the method the soil samples were first dried to constant weight at 105°C, then grounded and homogenized in a porcelain mortar before sample preparation. Four ml 10% KOH in methanol was added to one gram of the dried and homogenized soil samples then vortexed in a 15 ml polypropylene screw-cap centrifuge tube for 10 s. After 15 min ultrasonic treatment the samples were kept at 70°C for 90 min. Following the addition of 1 ml of distilled water and 2 ml of n-hexane and vortexing for another 30 s the tubes were centrifuged at 4000 rpm for 10 min. After removing the top phase the extraction was repeated with 2 ml n-hexane and the hexane fractions were evaporated in a water bath at 45°C to dryness. The precipitates were redissolved in 2 × 200μl methanol and filtered through a 0.22 μm PTFE syringe filter. The sample solutions were analyzed with HPLC equipped with a C18 Novapak column (Serial number: 197) and a Jasco UV-970 (UV: 282 nm) detector. Methanol was used as eluent at a flow rate of 1 ml min−1. Run time: 10 min, rate: 5.00 points min−1. Calibration standards of ergosterol were prepared at concentrations of 0.02–0.5 μg ml−1 by diluting the pure solid standard in methanol. To calibrate the instrument response 100 μl of each standard solution was injected into the HPLC apparatus.

[7] The total number of fungal propagula as a microbiological measure of the amount of soil fungi was determined with standard quantitative microbiological culturing method. Stock suspension of tenfold dilution (10 g of homogenized soil samples stirred in 90 ml physiological saline) was sonicated to destroy the biotic aggregates. Further tenfold dilution series were prepared with physiological saline. Aliquots of 100 μl of every dilution step were spread on the plate surface of mycological nutrient medium potato dextrose agar (Merck Ltd., Germany) containing penicillin. The plates were incubated under aerobic conditions at 26°C for one week. As the majority of fungal propagula forms a single particle in the suspensions the number of grown colonies could be used for calculation of the number of the colony forming units (CFU).

2.3. Sesquiterpene Emission Flux Measurements

[8] The sesquiterpene emission fluxes of the samples were determined by the same method as described in details in our previous paper [Horváth et al., 2011]. Briefly, the emitted metabolites were extracted by solid phase microextraction in a sterile flow-through apparatus designed in the laboratory to sample the headspace of three soil samples simultaneously. Each sampling vial was made of borosilicate glass (9 cm in diameter, 4 cm in height) and was provided with a sampling port sealed with a silicone rubber, PTFE-protected septum (Duran Group). The headspace of the vials was constantly flushed with purified compressed air at a flow rate of 1 ml min−1 in order to avoid the accumulation of carbon dioxide and thus the inhibition of fungal activity. The flush air was purified upstream with an activated carbon column (activated charcoal, apparent density: 0.48 g cm−3, particle size: 100 mesh) and a HEPA (Balston DFU, Grade AQ) filter to remove both volatile organic compounds and aerosol particles. The temperature of the cultures was held constantly at 26°C by immersing the sample vials into a thermostat (Lauda E100). Light conditions were not controlled in the laboratory. Emission flux measurement with zero flush air is a reliable representation of natural processes since sesquiterpenes are highly reactive compounds in the atmosphere for which redistribution between atmosphere and soil particles (and thus dry deposition) is not possible. In our experiments three SPME fibers coated with 100 μm thick polydimethylsiloxane were used. The fibers were exposed to the headspace for a duration of 1 h. The fibers were injected into an Agilent 6890 gas chromatograph coupled to a quadrupole mass spectrometer 5973 MSD (mass selective detector, Agilent) fitted with a ZB-5 MS capillary column (30 m × 0.25 mm, 0.25μm film thickness). The operating conditions were as follows: the injector was held at 250°C (splitless mode), the transfer line at 250°C. The MSD was used in electron impact ionization mode at 70 eV. Helium was used as carrier gas at a flow rate of 1 ml min−1. The following temperature program was used: oven temperature start at 40°C, hold 2 min, then programmed from 40°C to 120°C at 10°C min−1, from 120°C to 200°C at 4°C min−1, from 200°C to 250°C at 15°C min−1, hold 5 min. The analysis was performed in selective ion mode (SIM), m/z = 161 and m/z = 204 mass fragment ions of the sesquiterpenes were used for semiquantitative analysis. Calibration standards of all commercially available sesquiterpenes (α-cedrene, Aldrich 22133;β-caryophyllene, Sigma 22075; thujopsene, Aldrich 89235; and humulene, Aldrich 53675) were prepared at solute concentrations of 0.005–10 ppm by diluting the pure liquid standards in cyclohexane. To calibrate the instrument response multipoint calibration was performed in which 1μl of each standard solution was injected into the gas chromatograph in splitless mode and measured with the temperature program detailed in our previous paper [Horváth et al., 2011]. It has to be admitted that this type of external calibration introduces uncertainties due to imprecise sample volumes and some venting losses as well as the differences in the shape of chromatographic peaks compared to direct gas-phase calibration which was not possible in our laboratory for lack of gas standards and adequate gas-dilution infrastructure. The peak areas of the sesquiterpenes in the ion chromatograms of 161 m/z and the calibration curves of the available standard compounds were used to calculate the amounts of the analytes adsorbed on the fiber during the sampling period of 1 h. Only the amount of four standard compounds were directly determined from their respective calibration curves; for all other sesquiterpenes the calibration curves to be used were selected on the basis of highest mass spectral similarities with one of the four available standard compounds. These proxy calibration curves were used to estimate the adsorbed amount of these unidentified compounds assuming identical mass spectral distributions and ionization efficiencies with the standard. Therefore the results are semiquantitative only and based on cross-calibration of the four available standard compounds they are loaded with uncertainties of about a factor of 2 (50–200% relative to the reported result).

[9] As an additional source of uncertainty, the distribution of SQTs between the fiber and the headspace was estimated on the basis of available thermodynamic data as detailed in our previous paper [Horváth et al., 2011]. However, it has to be stressed that the experimental setup was exactly that same that was used for the estimation of SQT emissions from pure fungal cultures [ibid.]

3. Results and Discussion

3.1. Total Sesquiterpene Emission Fluxes From the Soil Samples

[10] Individual sesquiterpenes other than the commercially available standard compounds were not possible to identify. Based on library search of full scan mass spectra α-Gurjenene, (+)-Valencene,α-helmiscopene,β-bisabolene andβ-cadinene were tentatively identified in Leptosol with oak vegetation from Királyszentistván.

[11] The measured total emission fluxes of all detected sesquiterpenes for the different soil samples are summarized in Table 2. The sesquiterpene emission flux was determined directly in units of ng m−2 h−1. All the results in Table 2 are estimated to be uncertain at least by a factor of 2 due to the analytical limitations discussed in section 2.3. Since it was not possible to conduct repeated measurements from the same soil samples in the laboratory without possibly changing their natural state due to the limited capacity of the experimental apparatus, standard errors of the sampling and measurement procedures were not determined. The reported results are deemed semiquantitative only. The SQT emission fluxes for different soil/vegetation combinations differ by 3 orders of magnitude. The maximum value was measured in the case of Leptosol soil covered with hornbeam and oak and a minimum value for a Chernozem soil with barley. Concerning the limited set of soil/vegetation combinations, median emission fluxes (with standard deviation) were 109 (660) ng m−2 h−1 and 64 (87) ng m−2 h−1 for undisturbed samples and samples with litter/vegetation removed, respectively. In most cases there were significant differences between the sesquiterpene emissions of the same soil types with different vegetations, except for Chernozem soils from which emission fluxes were in the same order of magnitude for different vegetations. Removal of litter/vegetation affected emission fluxes to either direction, i.e., increasing or decreasing SQT emission fluxes on a per mass basis. However, high emission fluxes in the order of several hundred ng m−2 h−1 measured in undisturbed samples all dropped significantly when litter/vegetation was removed.

Table 2. Total Sesquiterpene Emissions of Different Soils in Units of ng m−2 h−1
Soil TypeaSQTb
Undist.Without Litter
  • a

    Vegetation is given in parentheses.

  • b

    Undist.: undisturbed. Without litter: litter/vegetation removed.

Cambisol (sunflower)83030
Cambisol (barley)740
Chernozem (pea)2540
Chernozem (barley)660
Leptosol (oak)43014
Leptosol (hornbeam and oak)198015
Luvisol (pasture)4030
Luvisol (weed)130290

3.2. Fungal Biomass in the Soil Samples

[12] The dry fungal biomass as estimated from the measured ergosterol concentrations is given in Table 3. The highest fungal biomass was found in a Leptosol soil with hornbeam and oak which also yielded the highest sesquiterpene emission flux. Although one would expect higher sesquiterpene emission for a higher value of fungal biomass, sesquiterpene emission fluxes are also influenced by the metabolic activity of different fungal species, similarly to sporulation which is associated with secondary metabolism [Walker and White, 2011]. In accordance with the results of our previous culture experiments and those of indoor measurements, sesquiterpene emission fluxes from the soil samples spanned a range of about three orders of magnitude [Horváth et al., 2011; Fischer et al., 1999; Kuske et al., 2005]. In some cases the fungal biomass of the samples was not proportional to the measured SQT fluxes. For Leptosol with oak and Luvisol with pasture fungal biomass of 35 and 200 μg g−1 dry soil corresponded to SQT fluxes of 430 and 40 ng m−2h−1, respectively. From the limited number of samples and conditions, no clear relationship can be established between the dry fungal biomass and SQT emission. As found in our previous paper [Horváth et al., 2011], various fungal cultures of the same size yielded very different emission fluxes. In the case of soil samples this fact is aggravated by the differences in soil properties and conditions as well as the differences between native fungal strains and their development stage. Furthermore, though strongly decayed vegetation litter is not expected to directly emit residual sesquiterpenes at the times and locations of sampling, soil biota other than fungi (e.g., bacteria) may also contribute to SQT emission from soil. Therefore it cannot be guaranteed that all of the SQTs come from de novo production by soil fungi. The concentrations of fungal propagules and dominant fungal genera in the soil samples are also listed in Table 3. The number of cultured fungal propagula of the soil Ca1 resulted in 9.7 × 104 CFU g−1 as one of the most dominant fungal genera proved to be a low sporulating genus, Mortierella. The relatively low number of propagules (2.9 × 105 CFU g−1) in the soil Lu1 can be explained by the dominance of Rhizoctoniagenus having spore-free mycelia, while the high number of propagula of the soilLu2 with a relatively high fungal biomass could be caused by the rich sporulation properties of the dominant genera Paecilomyces, Torula and Trichoderma.

Table 3. Concentration of Fungal Biomass, Fungal Propagules, and Dominant Fungal Genera in the Soil Samples
Soil TypeaSoil SymbolDry Fungal Biomassb (μg g−1)Fungal Propagules (CFU g−1)Dominant Fungal Genera
  • a

    Vegetation is given in parentheses.

  • b

    NA: not available.

Cambisol (sunflower)Ca187.8 × 105Cladosporium sp. Mortierella sp. Paecilomyces sp.
Cambisol (barley)Ca2259.7 × 104Cladosporium sp. Mortierella sp. Paecilomyces sp.
Chernozem (pea)Ch1351.5 × 105Acremonium sp. Alternaria sp. Verticillium sp.
Chernozem (barley)Ch22003.4 × 105Acremonium sp. Alternaria sp. Fusarium sp.
Leptosol (oak)Le1381.9 × 106Acremonium sp. Arthrobotrys sp. Penicillium sp.
Leptosol (hornbeam and oak)Le289.0 × 105Acremonium sp. Cladosporium sp. Penicillium sp.
Luvisol (pasture)Lu1NA2.9 × 105Fusarium sp. Penicillium sp. Rhizoctonia sp.
Luvisol (weed)Lu2505.6 × 105Paecilomyces sp. Torula sp. Trichoderma sp.

3.3. Atmospheric Implications

[13] In our previous study, based on the medium SQT emission factor (for the temperature of 26°C) of Trichoderma harzianum Mt 29 in the culture experiments and assumed average fungal biomass in typical soils, SQT emission flux of 8–50 ng m−2 h−1 was inferred from soils [Horváth et al., 2011]. This estimate was obtained by assuming the same rates of SQT production in soils as those measured in pure cultures and no loss of SQTs on soil particles or in soil interstitial air. The SQT emission fluxes of selected soils, as measured in the present study fall into the range of 6–1980 ng m−2 h−1 (median 109 ng m−2 h−1) for undisturbed soils and of 14–290 ng m−2 h−1 (median 64 ng m−2 h−1) for soil samples with vegetation/litter layer removed. These measured soil fluxes are generally in the same order of magnitude as (or in some cases exceed) the range of 8–50 ng m2 h−1 inferred by calculations on the basis of laboratory measurements [Horváth et al., 2011]. High emission fluxes are always measured from undisturbed soils indicating that the presence of litter may enhance SQT production or reduce SQT losses in soil under favorable conditions.

[14] Our results imply that soils close to their natural conditions do emit SQTs at rates comparable to those inferred by our previous fungal culture experiments [Horváth et al., 2011]. These findings corroborate our previous hypothesis that sesquiterpene emission from soils may be important compared to that of vegetation in certain regions and time of the year and has to be included in VOC inventories. As a matter of course, the limited span of our study in terms of soil and vegetation types and conditions is just the first step toward this goal and further studies are clearly warranted.


[15] The authors are grateful for the financial support of the grants TÁMOP-4.2.2/B-10/1-2010-0025 ‘Supporting research schools at the University of Pannonia’ and TÁMOP-4.2.1.B-11/2/KMR-2011-0003 supporting researches at the University of Szent István, Gödöllő. The Projects are supported by the European Union and co-financed by the European Regional Development Fund.