Elevated Atmospheric Carbon Dioxide Concentrations Amplify Alternaria alternata Sporulation and Total Antigen Production

Background Although the effect of elevated carbon dioxide (CO2) concentration on pollen production has been established in some plant species, impacts on fungal sporulation and antigen production have not been elucidated. Objective Our purpose was to examine the effects of rising atmospheric CO2 concentrations on the quantity and quality of fungal spores produced on timothy (Phleum pratense) leaves. Methods Timothy plants were grown at four CO2 concentrations (300, 400, 500, and 600 μmol/mol). Leaves were used as growth substrate for Alternaria alternata and Cladosporium phlei. The spore abundance produced by both fungi, as well as the size (microscopy) and antigenic protein content (ELISA) of A. alternata, were quantified. Results Leaf carbon-to-nitrogen ratio was greater at 500 and 600 μmol/mol, and leaf biomass was greater at 600 μmol/mol than at the lower CO2 concentrations. Leaf carbon-to-nitrogen ratio was positively correlated with A. alternata spore production per gram of leaf but negatively correlated with antigenic protein content per spore. At 500 and 600 μmol/mol CO2 concentrations, A. alternata produced nearly three times the number of spores and more than twice the total antigenic protein per plant than at lower concentrations. C. phlei spore production was positively correlated with leaf carbon-to-nitrogen ratio, but overall spore production was much lower than in A. alternata, and total per-plant production did not vary among CO2 concentrations. Conclusions Elevated CO2 concentrations often increase plant leaf biomass and carbon-to-nitrogen ratio. Here we demonstrate for the first time that these leaf changes are associated with increased spore production by A. alternata, a ubiquitous allergenic fungus. This response may contribute to the increasing prevalence of allergies and asthma.


Research
Anthropogenic increases in global atmo spheric carbon dioxide (CO 2 ) concentra tion have been shown to stimulate earlier and greater production of allergenic pollen (LaDeau and Clark 2006;Rogers et al. 2006;Ziska et al. 2003), as have warming tem peratures (Wan et al. 2002;YliPanula et al. 2009). The effects of anthropogenic climate change on the production of airborne fungal spores are not as well documented, but the implications for allergic disease are no less important. As with pollen, exposure to fungal spores is associated with allergy and asthma symptoms [Institute of Medicine (IOM) , 2004Salo et al. 2006], although the specifics of the relationship are not completely understood (Portnoy et al. 2008). Among patients with asthma from six regions of the world, 11.9%, on average, were sensitized to Alternaria alternata, with the proportion as high as 28.2% in Portland, Oregon; in addition, sensitivity to A. alternata was more prevalent among patients with more severe asthma (Zureik et al. 2002). Allergenic fun gal spores may be increasingly abundant in some areas of the globe as well. In Derby, United Kingdom, mean seasonal airborne spore concentrations of the genus Alternaria have increased over the years 1970-1998, as have the number of days with spore counts > 50 per cubic meter of air, and the start of the Alternaria spore season has advanced from the end to the beginning of June (Corden and Millington 2001).
Rising atmospheric CO 2 concentration is well documented [Intergovernmental Panel on Climate Change (IPCC) 2007]. Large increases in atmospheric CO 2 concentration (e.g., doubled or greater increases in concen tration) have been shown to alter both plant biomass and chemistry in many plant spe cies (Ainsworth and Long 2005;. The responses of individual plant spe cies to large increases in atmospheric CO 2 concentration are variable but often include greater total biomass production (de Graaf et al. 2006) and greater carbontonitrogen ratios (C:N) of plant tissues (Taub and Wang 2008). Primary consumers, including aller genic fungi that grow on living or dead plant materials, are therefore likely to encounter changes in substrate quality or quantity as global atmospheric CO 2 concentration increases (Reich et al. 2006).
The responses of plant-fungal interactions to increased atmospheric CO 2 concentration may be more complex and difficult to elu cidate than the observed changes in pollen production and allergenicity. Fungi play sev eral different roles in plant biology, acting as plant pathogens, saprobes (decomposers of dead plant material), or plant symbionts, with neutral, beneficial, or negative impacts on living plant hosts. There may be a number of indirect effects on fungal growth and repro duction as a result of plant changes under increased atmospheric CO 2 concentration as well as feedback and interactions among plants, fungi, and abiotic (e.g., environmen tal, nonorganismal) factors (Rillig 2007;Stiling and Cornelissen 2007).
In a field experiment using opentop chambers to grow individual poplar trees in ambient or doubled atmospheric CO 2 con centration, Klironomos et al. (1997) found that air and leaf litter in chambers with doubled CO 2 concentration contained sig nificantly more fungal spores than those in ambient chambers. An indirect effect of doubled atmospheric CO 2 concentration on sporulation, mediated by changes in plant chemistry, was suggested by their findings; however, a direct effect of atmospheric CO 2 concentration or of related changes such as soil moisture and relative humidity could not be ruled out.
To directly assess the quantitative and qualitative impacts of rising atmospheric CO 2 concentration on allergenic fungi, we grew the perennial C 3 monocot timothy grass (Phleum pratense), a common fodder used for hay, at four levels of atmospheric CO 2 concentra tion approximating preindustrial, current, and potential/projected future concentrations (300, 400, 500, and 600 µmol/mol, respectively). We used leaves from these plants to separately volume 118 | number 9 | September 2010 • Environmental Health Perspectives grow two fungal species. The first species, A. alternata, is a ubiquitous, facultatively plantpathogenic or saprobic species known to produce allergenic airborne conidial spores (Sanchez and Bush 2001). The second species, Cladosporium phlei (syn. Heterosporium phlei), is a specific pathogen of timothy that has not been specifically shown to produce aller genic spores; other species within the genus Cladosporium are known to cause allergies in humans (Kurup 2003;Poll et al. 2009). We also examined the sporulation of A. alternata on clippings from fieldgrown grasses to see if responses were similar on other grass species. Our objective was to quantify changes in sporulation on plant leaves grown at varying atmospheric CO 2 concentrations and elucidate the mechanism that drives such changes.

Methods
General approach. We used two growth chambers to provide atmospheric CO 2 con centration at 300, 400, 500, or 600 µmol/mol to experimental timothy plants for 60 days of growth. These approximate atmospheric CO 2 concentrations at the beginning of the 19th century, current ambient levels, and those projected by the years 2025 and 2040, respec tively (IPCC 2007). During repeated runs over time, each of the two growth chambers was used for each level of CO 2 concentra tion twice. During each run, each chamber contained 10 timothy plants. After 60 days of growth, leaf subsamples from each plant were used to grow A. alternata and C. phlei separately for 7 days inside loosely capped media bottles in an incubator. Media bottles were then volumetrically flooded with water, shaken, and subsampled for spore counts and, for A. alternata only, for antigenic protein extraction and quantification via ELISA using polyclonal antibodies.
Growth, harvest, and C:N quantification of timothy grass. The study was conducted using two controlled environment chambers (EGC Corp., Chagrin Falls, OH) at Beltsville, Maryland, with a given chamber set at one of four atmospheric CO 2 concentration set points (300, 400, 500, and 600 µmol/mol) for 24 hr/day. The atmospheric CO 2 concen tration of the air within each chamber was controlled by adding either CO 2 or CO 2 free air to maintain the set concentration. Actual average 24hr atmospheric CO 2 concentra tion values (± SD) were 315 ± 19, 395 ± 11, 491 ± 10, and 589 ± 12 µmol/mol. Injection of CO 2 and CO 2 free air was controlled by a TC2 controller using input from an absolute infrared gas analyzer (WMA2, PP Systems, Haverhill, MA). Chamber tempera tures were altered diurnally (in steps) from an overnight low of 20°C to a maximum after noon value of 30°C. Photosynthetically active radiation (PAR) was altered concurrently with temperature, with the highest PAR value (900-1,000 µmol/m 2 per sec) occur ring during the afternoon (1,200-1,500). The duration of daily PAR was 14 hr, which was supplied by a mixture of highpressure sodium and metal halide lamps.
Seeds of timothy grass (P. pratense, var. Climax) were obtained from Pawnee Buttes Seed (Greeley, CO). We sowed three seeds in each 4inch pot used and thinned to one seedling per pot 6-8 days after emergence. Pots were watered to the drip point daily with a complete nutrient solution contain ing 14.5 mmol/L nitrogen. Ten pots were grown in each of the two chambers during each run. In eight runs over time, each level of CO 2 concentration was run a total of four times, twice in each chamber. Plants from the third run were stunted and grew poorly for unknown reasons; this run was omitted, resulting in three runs for CO 2 concentration levels of 300 and 500 and four runs for CO 2 concentration levels of 400 and 600.
Plants were harvested 60 days after sowing. All leaves were separated from stems above the collar and weighed. Ten leaves from each plant were used for determination of total area, mass, and nitrogen and carbon content, which was determined using a PerkinElmer 2400 Series II CHNS/O analyzer (PerkinElmer, Waltham, MA). Leaf subsamples of approxi mately 2 g fresh weight were removed from each plant for fungal growth, refrigerated, and processed within 3 hr of harvest.
All solutions, suspensions, and rinses were made using sterilized, deionized, distilled water (ddw) with Tween 80 (Fisher Scientific, Fairlawn, NJ) added at the rate of 5 drops/L (ddw+T). Leaves were surfacesterilized by immersion and agitation in 1 L of 0.3% sodium hypochlorite solution for 3 min, fol lowed by three successive 1min rinses. All solutions and rinses were changed after leaves from all plants in a chamber were immersed to avoid transfer of any leaf material among treatments. Surfacesterilized leaves were oven dried at 68°C for 48 hr.
Fungal inocula were prepared immedi ately before inoculating leaves and prepared the same way for both fungal species. Two Petri plates with conidia covering most of the agar surface were selected. Conidia were removed from these plates by pouring small aliquots of ddw+T onto the surface of the agar in each plate, gently scraping the agar surface with a ceramic scraper, pouring the water with removed conidia through four lay ers of sterilized cheesecloth (Fisher Scientific) into a sterilized 1L beaker, and repeating three more times. The resulting spore sus pensions were vortexed (Vortex Genie2, Daigger, Vernon Hills, IL) at maximum speed for 30 sec to separate clusters of spores and subtending hyphae. The total volume was brought up to 1 L with additional ddw+T. Inoculum strength was standardized visually: A 10mL subsample was removed with a glass widemouth pipette, filtered with a 47mm diameter, 0.45µm pore size, gridded mixed cellulose filter (Fisher Scientific) in a glass vac uum filtration apparatus, and examined at 5× magnification under a dissecting microscope (SMZ 1500; Nikon Precision Inc., Belmont, CA). If spores were more than one layer deep on the filter, ddw+T was added to the inocu lums suspension in 200mL increments until subsamples resulted in filters densely covered with a single layer of spores. Five hundred milliliters of the final inoculum suspension was placed into two sterilized 1L beakers; each beaker was used to inoculate leaves from a single chamber.
Weighed leaf subsamples (0.5 ± 0.075 g dry weight) from each single plant were inocu lated separately by immersion and agitation in the inoculum suspension for 1 min. Inoculated leaves were placed in a sterilized media bottle (250mL size used for A. alternata, 500mL size for C. phlei) (VWR, West Chester, PA) with caps placed loosely, then incubated for 1 week in an environmental controller (Percival, Perry, IA) set at 21°C with a 12hr light/dark cycle and ambient CO 2 concentration.
Spore harvest and A. alternata antigen extraction. After 1 week, 200 mL sterilized ddw+T was added to each media bottle. Bottles were vortexed at maximum speed for 30 sec to dislodge spores from leaves and sub tending hyphae. For A. alternata only, antigen was then extracted: media bottles were left at room temperature for 2 hr; bottles were then inverted several times by hand. With a sterile syringe, 20 mL water with suspended spores was removed from each bottle and passed through a syringe filter (0.22µm pore size, 33mm diameter, polyethersulfone; Millipore, Billerica, MA). The resulting spore free extract was frozen inside a 50mL plastic centrifuge tube (Corning Inc., Corning, NY) and lyophilized on a Freezone 4.5 freeze dry system (Labconco, Kansas City, MO). For both fungal species, media bottles were then inverted several times by hand; 10mL sub samples were removed using glass widemouth pipettes, placed on filters, and examined as described above. If spore density on the filter was too sparse (or too crowded) for accurate counting, a larger (or smaller) subsample was obtained, a new filter was made, and spore density was rechecked. The final volume used for each subsample ranged from 1 to 100 mL and was recorded for extrapolation of the total number of spores produced per gram of dried leaf tissue.
Spore quantification. Filters were rewet ted from below with several drops of sterilized ddw and placed on a compound light micro scope stage (Axioplan2 Imaging Microscope; Carl Zeiss Microimaging Inc., Thornwood, NY). Eight 300 × 400 µm fields of view were digitally photographed (Axiocam digi tal camera and Axiovision imaging software; Carl Zeiss Microimaging, Inc.) at random positions over one half of each wetted filter. Spores within each of the eight fields were counted manually from the photographs; all whole, uncollapsed spores were counted regardless of size. For A. alternata only, the first 10 spores laying flat on the filter were also measured for length and width.
Antigen quantification. Lyophilized extracts were reconstituted in 3 mL ddw and assayed for A. alternata antigen using a com petitive inhibition ELISA described by Salo et al. (2005). The assay kit consisted of a rab bit polyclonal antibody (lot ZA44; Greer Laboratories, Lenoir, NC) produced using whole mycelial extracts of A. alternata (lot XPM1X10); the A. alternata mycelial extract was also used as the standard in the assay. This assay has been shown to detect multiple A. alternata antigens including the allergen Alt a 1 (Portnoy et al. 1993;Salo et al. 2006). Briefly, the standard A. alternata extract was bound to plastic micro titer plate wells, excess was washed away, and unbound sites were blocked with 1% bovine serum albumin solu tion. Dilution series of the sample extracts were pipetted into the wells and followed immediately by the rabbit antiA. alternata antibody. After incubation, we aspirated solutions, washed the wells, and added a per oxidaselabeled goat antirabbit antibody solu tion (SigmaAldrich, St. Louis, MO). Finally, we added substrate and meas ured the color change at 405 nm using a microtiter plate reader (Molecular Devices, Sunnyvale, CA). We compared the resulting reaction rates against a dilution series of standard A. alternata antigen to determine antigen concen tration. Results are reported as micrograms A. alternata antigen per spore or per plant.
Data analysis. All statistical analy ses were conducted using R version 2.8.1 (R Development Core Team 2008). The R package "lme4" (Bates et al. 2008) was used to fit and evaluate generalized linear mixed models of the following effects: atmospheric CO 2 concentration level (fixed at 300, 400, 500, or 600 µmol/mol), run (random effect of seven runs over time), and chamber (random effect of the two growth chambers used). All fungal responses were also modeled with the covariate of leaf C:N of each plant. Factors were sequentially removed from models and model fit was evaluated using Akaike's infor mation criterion; for each response, the sim plest statistical model with good fit was used. We used the R package "languageR" (Baayen 2008) to estimate pvalues for the fixed effect of atmospheric CO 2 concentration level in each model using Markov Chain Monte Carlo resampling and to generate highest posterior density confidence intervals (CIs) for means at each atmospheric CO 2 concentration level. We considered pvalues ≤ 0.05 to be statistically significant. The responses evaluated were the following: a) A. alternata spore counts (natu ral log transformed to meet statistical model assumptions that variance does not increase with group mean); b) mean length and width of A. alternata spores; c) quantity of antigenic protein produced (squareroot transformed to linearize relationships with model factors and to meet statistical model variance assump tions); and d) C. phlei spore counts (natural log transformed, as above).
The random effect of chamber was neither significant nor needed to improve model fit for any responses and is not discussed further. The random effect of run was very important for modeling all plant and fungal responses; modeladjusted means, which account for vari ation due to run, are presented for responses at individual levels of CO 2 concentration. When the fixed effect of CO 2 concentration was determined to be statistically significant, the R package "multcomp" (Hothorn et al. 2008) was used to run pairwise Tukey's compari sons of CO 2 concentration level means. The contribution of each factor toward explaining response variance was calculated as described by Kramer (2005).
Supplemental experiment with fieldcollected plant material. Nine samples of plant material were collected during routine mowing from three different sites at each of three local golf courses. A mixture of peren nial ryegrass (Lolium perenne L.) and annual poa (Poa annua), both cool season grasses with C 3 photosynthetic metabolism, is grown at all sites, but fertilizer application rates vary. The samples were collected on 8 January 2008 and refrigerated within 3 hr of mowing. Leaves were treated and used as growth substrate for the two fungal species as described above. Resulting C:N and spore counts were not statistically analyzed because of small sample size but are shown for visual comparison (Figure 1).

Plant responses to atmospheric CO 2 concentration levels.
Leaf C:N was significantly higher in plants grown at 500 and 600 µmol/mol atmospheric CO 2 concentrations than at the two lower concentrations (p = 0.017) ( Table 1). Leaf dry weight per plant was significantly greater at 600 µmol/mol atmo spheric CO 2 concentration than at the three lower concentrations (p < 0.001) ( Table 1).  Table 1. Adjusted means and 95% CIs by CO 2 concentration level.
Fungal sporulation on leaves grown at four atmospheric CO 2 concentration levels. The logtransformed counts of A. alternata spores produced per gram leaf were positively cor related with leaf C:N (p < 0.001; adjusted R 2 = 0.25; Figure 1), with an orderofmagnitude increase across the range of leaf C:N of experi mental plants. A. alternata sporulation on field collected grass leaves (supplemental experi ment) shows a similar relationship (Figure 1). Mean length and width of A. alternata spores did not change significantly with leaf C:N (not shown). In contrast to the positive asso ciation with spore numbers, the quantity of A. alternata antigen per spore (squareroot transformed) decreased as leaf C:N increased (p < 0.001; adjusted R 2 = 0.34; Figure 2). Despite the decreased antigen content per spore at higher C:N, the quantity of A. alternata spores as well as the total antigen produced on a perplant basis were positively associated with atmospheric CO 2 concentration levels beyond the variation attributed to leaf C:N, with nearly three times the number of spores and twice the amount of antigen produced on plants grown at 500 and 600 µmol/mol atmospheric CO 2 concentrations (p < 0.001; Table 1) than at the lower two concentrations.
The logtransformed counts of C. phlei spores produced per gram leaf also increased with C:N (p < 0.001; adjusted R 2 = 0.22; Figure 3). Spores of C. phlei were produced in much lower numbers than A. alternata spores, although the slopes of the relationships with leaf C:N were not significantly different between the two fungi (T = 1.21; p > 0.20). In contrast with A. alternata spores, however, the mean number of C. phlei spores produced on a perplant basis was not significantly dif ferent among plants grown at different CO 2 concentration levels (p = 0.78; Table 1).

Discussion
Climate change and urbanization are expected to increase the prevalence of asthma and aller gies (Schmier and Ebi 2009). Concomitant rises in temperature, atmospheric CO 2 con centration, and pollen abundances over recent decades have been suggested as causes of increasingly prevalent and severe asthma and allergy symptoms observed over the same time period (Beggs and Bambrick 2005;Shea et al. 2008). Our findings indicate that, as with pol len production, the sporulation of allergenic fungi is likely to be amplified as atmospheric CO 2 concentration increases and therefore is also likely to contribute to increasing preva lence and severity of asthma and allergies.
We found significant positive relationships between leaf C:N and A. alternata and C. phlei spore production per gram leaf, although only A. alternata sporulation was associated with CO 2 concentration level after accounting for changes in leaf C:N. These relationships sug gest that considerable increases in sporulation of both species will occur, and may already be occurring, if rising atmospheric CO 2 concen tration increases plant C:N in the field. The response of C. phlei, however, may not be as strong as that of A. alternata. More study of this species is needed.
Although C:N was correlated with increased sporulation, a large proportion of the variability in spore production remains unex plained. Some of this variability in sporulation may result from inherent variation in growth among individual plants, which can be of a magnitude large enough to obscure responses to elevated atmospheric CO 2 concentration treatments (Poorter and Navas 2003). Other aspects of environmental variability, which deserve additional scrutiny, may also contrib ute to variation in spore numbers and qual ity. Temperature was kept constant during fungal growth in this experiment, but rela tive humidity was not controlled. Additional plant attributes such as leaf tissue concentra tions of mineral elements (other than nitro gen), stomatal density, and other leaf surface properties, and the chemistry of leaf carbon compounds were not measured. All of these factors may contribute to the variation in spo rulation, and all are also likely to be altered by anthropogenic climate change (Denman et al. 2007). For example, plant transpira tion is likely to decrease under elevated atmo spheric CO 2 concentration, which often leads to increased moisture in soils and overlying litter (Johnson et al. 2003); warming tempera tures and altered precipitation patterns will further impact moisture in soil and plant litter layers. Global climate change factors encom pass many direct and indirect effects on plants and fungi, as well as complex interactions and feedbacks. The net effects of global changes are difficult to capture with a single study. Long term (multidecadal) observational studies are needed to distinguish the impact of multiple abiotic changes on allergenic spore production in toto.
Our results corroborate the findings of Klironomos et al. (1997), who found large increases (~ 100-150%) in Alternaria spore abundance in chambers with poplar trees growing in elevated atmospheric CO 2 concentration. Our results also suggest the potential ubiquity, across plant species, of increased spore production as a function of increasing atmospheric CO 2 concentration. Klironomos et al. (1997) also found a large increase in the abundance of spores from the genus Cladosporium at elevated atmospheric CO 2 concentration (~ 225-350% increases), which was not apparent in our study. A strong response to atmospheric CO 2 concentration may relate to a saprobic (growing on dead plant tissues) lifestyle. A. alternata is facul tatively pathogenic or saprobic. Although C. phlei can grow on dead plant tissues, as it did in this experiment, it is described primar ily as a pathogen of living timothy grass. Other  species within Cladosporium, however, are sap robic (Zalar et al. 2007). Specializations that allow fungal species to grow on specific host plants, such as unique metabolic pathways, may cause these fungi to be limited more by other plant characteristics than by carbon content alone.
Although only a few studies have exam ined fungal sporulation on plants grown at elevated atmospheric CO 2 concentration, there has been extensive work examining spo rulation on artificial growth media and on plants with different levels of nitrogen con tent. In some pathogenic fungal species, spo rulation on living leaves is greatest at medium C:N (Walters and Bingham 2007), but other species sporulate maximally at high C:N. The C:N that promotes the most sporulation varies among fungal species and even strains (Elson et al. 1998;Gao et al. 2007). These varying trends among fungal taxa may reflect different reproductive strategies, specializa tions, and functional tradeoffs needed to thrive in different environmental and host conditions (Pariaud et al. 2009). There is experimental evidence that while the great est numbers of spores may be produced at high C:N in some fungal taxa, spore protein content, germination success, and hostplant infectivity are greater when more nitrogen is available (Jackson and Schisler 1992;Yu et al. 1998). In our study, the decreasing content of antigenic protein in A. alternata spores grown at higher C:N suggests increasing nitrogen limitation of the fungus. It is not yet clear if this nitrogen limitation is associated with compromised germination, growth, or aller genicity of individual A. alternata spores.
Metaanalyses of plant responses to ele vated atmospheric CO 2 concentration reveal clear trends despite variation among plant spe cies and experimental conditions. The mean increase in C:N across many studies of grass species grown at doubled atmospheric CO 2 concentration was 24.4% (Taub and Wang 2008). The responses of plants to lessthan doubled increases in atmospheric CO 2 con centration have not been well studied, so expectations for C:N of plant tissues grown in realworld, gradually increasing atmospheric CO 2 concentration cannot be predicted with certainty. However, an observational study of herbarium leaf specimens found that leaf nitrogen content decreased by a mean of 17% across samples from the early, mid, and late 20th century for 11 C 3 plant species (Penuelas and Estiarte 1997) [recorded global levels of atmospheric CO 2 concentration increased from < 320 in 1958 to 388 µmol/mol in July 2009 (National Oceanic and Atmospheric Administration 2009)]. The authors attribute the decreased leaf nitrogen content to the combination of direct and indirect (warming) effects of rising atmospheric CO 2 concentra tion, and their findings are corroborated by other observational and experimental meas urements (Denman et al. 2007;Gifford et al. 2000;Taub and Wang 2008). Thus, this evidence of increased plant C:N associated with rising atmospheric CO 2 concentration and climatic change is consistent with increased sporulation of allergenic fungi, as observed in the current study, and has impli cations for increasing the prevalence and sever ity of allergy and asthma symptoms (Beggs and Bambrick 2005;D'Amato et al. 2005).