Effects of chronic elevated ozone concentration on antioxidant capacity, photosynthesis and seed yield of 10 soybean cultivars


E. A. Ainsworth. 147 Edward R. Madigan Laboratory, 1201 W. Gregory Drive, Urbana, IL 61801, USA. Fax: +217 244 4419; e-mail: lisa.ainsworth@ars.usda.gov


Crops losses to tropospheric ozone (O3) in the United States are estimated to cost $1–3 billion annually. This challenge is expected to increase as O3 concentrations ([O3]) rise over the next half century. This study tested the hypothesis that there is cultivar variation in the antioxidant, photosynthetic and yield response of soybean to growth at elevated [O3]. Ten cultivars of soybean were grown at elevated [O3] from germination through maturity at the Soybean Free Air Concentration Enrichment facility in 2007 and six were grown in 2008. Photosynthetic gas exchange, leaf area index, chlorophyll content, fluorescence and antioxidant capacity were monitored during the growing seasons in order to determine if changes in these parameters could be used to predict the sensitivity of seed yield to elevated [O3]. Doubling background [O3] decreased soybean yields by 17%, but the variation in response among cultivars and years ranged from 8 to 37%. Chlorophyll content and photosynthetic parameters were positively correlated with seed yield, while antioxidant capacity was negatively correlated with photosynthesis and seed yield, suggesting a trade-off between antioxidant metabolism and carbon gain. Exposure response curves indicate that there has not been a significant improvement in soybean tolerance to [O3] in the past 30 years.


Tropospheric ozone concentrations ([O3]) have more than doubled in the past 50 to 100 years (Staehelin et al. 1994; Vingarzan 2004) and O3 is the most damaging air pollutant to plants (Ashmore 2005; EPA U.S. 2006). Crop losses to O3 damage in 2000 have been estimated to cost $1.8 to $3.9 billion in the United States and $3.0 to $5.5 billion in China (Van Dingenen et al. 2009). These significant costs to world agriculture will grow with the predicted 25% increase in background [O3] expected for the next 30 to 50 years (Meehl et al. 2007). However, this challenge also provides an opportunity to develop improved O3 tolerance in sensitive crop species (Fiscus, Booker & Burkey 2005; Ainsworth, Rogers & Leakey 2008).

Soybean [Glycine max (L.) Merr.] is among the more O3-sensitive crops (Heagle 1989; Morgan, Ainsworth & Long 2003). A meta-analysis of 53 published studies investigating O3 effects on soybean showed that when grown at 60 ppb [O3], soybean above-ground biomass was reduced by 40% when compared to plants grown in charcoal-filtered air (Morgan et al. 2003). Supporting that synthesis, an increase in 7 h mean daytime [O3] from 30 to 60 ppb was estimated to reduce soybean yields by ∼16% (Ashmore 2002). Recent results from Free Air Concentration Enrichment (FACE) of soybean with elevated [O3] suggest that soybean yields are even more sensitive to elevated [O3] than previously predicted (Morgan et al. 2006). Seed yield of the commercial cultivar Pioneer 93B15 was reduced by 20% when [O3] was increased from an 8 h growing season average of 50 to 63 ppb (Morgan et al. 2006).

The sensitivity of soybean to [O3] is well established, yet there is also significant intraspecific variation in the response to elevated [O3] (Lee et al. 1984; Mulchi et al. 1988; Foy et al. 1995; Heagle, Miller & Pursley 1998; Robinson & Britz 2000; Cheng et al. 2007). Howell, Koch & Rose (1979) reported significant cultivar differences in seed yield response to ambient [O3]. Ozone tolerance has been linked to differences in antioxidant capacity (Gupta, Alscher & McCune 1991; Skarby et al. 1998; Chernikova et al. 2000; Robinson & Britz 2000; Cheng et al. 2007) and flux of O3 into the leaf (Fiscus et al. 2005; Dizengremel et al. 2009). Flux of O3 into the leaf is controlled by stomatal conductance (gs), which is often decreased by elevated [O3] as a result of reduced photosynthesis (McKee, Farage & Long 1995; Fiscus et al. 1997). While much research has focused on studying physiological and biochemical responses that provide O3 tolerance, no specific markers for O3 tolerance have emerged (Fiscus et al. 2005). There have been some recent successes in identifying quality trait loci associated with O3 tolerance in rice (Frei, Tanaka & Wissuwa 2008; Frei et al. 2010), but to date, little progress has been made in understanding the physiological or genetic basis for O3 tolerance in soybean.

In the current study, 10 different soybean cultivars were grown at elevated [O3] from germination through maturity at the Soybean FACE (SoyFACE) experiment in 2007 and a follow-up study of six of those cultivars was done in 2008. These cultivars are adapted to central Illinois growing conditions and showed variability in yield response to O3 in preliminary tests (R. Nelson, unpublished results). The hypothesis that there is significant cultivar variation in the response of soybean physiology, antioxidant capacity and yield to growth at elevated [O3] was tested. An additional aim of the study was to identify the interrelationship of growth, metabolic and yield responses to [O3]. Therefore, photosynthetic gas exchange, chlorophyll fluorescence, leaf area index (LAI), leaf chlorophyll content and leaf antioxidant capacity were measured during two growing seasons, and correlated with changes in seed yield. Using data from 2007 and 2008, ozone exposure response functions were defined for six soybean cultivars.


Experimental site and plant growth conditions

The SoyFACE facility is located on 32 ha near Champaign, IL, USA (40°02′N, 88°14′W, 228 m above sea level; http://soyface.illinois.edu). Soybean [G. max (L.) Merr.] and maize (Zea mays) are each planted on half of the area and rotated annually. As per standard regional agronomic practice, the maize crop was fertilized each year, but no fertilizer was added to the soybean crop. Meteorological data, including air temperature (Fig. 1a,b), incident maximum photosynthetic photon flux density (PPFD; Fig. 1c,d) and rainfall (Fig. 1e,f), were measured on site (described in Leakey et al. 2004).

Figure 1.

Meteorological data collected at the Soybean Free Air Concentration Enrichment facility in Urbana, Illinois during the 2007 and 2008 soybean growing seasons. Maximum (black line) and minimum (grey line) daily temperature in 2007 (a) and 2008 (b), daily maximum photosynthetic photon flux density (PPFD) in 2007 (c) and 2008 (d), daily precipitation in 2007 (e) and 2008 (f).

The experiment was a randomized complete block design (n = 4), with each block containing two treatments: 20-m-diameter octagons at current ambient [O3] or at elevated [O3]. The fumigation system, based on the design of Miglietta et al. (2001), added O3 as a fixed proportion of the instantaneous background [O3] measured in the control plots (Morgan et al. 2004). The fumigation target was twice the ambient concentration (Fig. 2a,b), with the maximum fumigation concentration capped at 150 ppb. The 8 h growing season average [O3] was 46.3 and 37.9 ppb in ambient plots in 2007 and 2008 and 82.5 and 61.3 ppb in elevated plots in 2007 and 2008, respectively (Fig. 2a,b). Ozone was not added at night or when leaves were wet. This treatment significantly increased the cumulative dose of O3 experienced by the crop, as measured by the accumulated exposure over a threshold of 40 ppb (AOT40; Fig. 2c,d) and sum of hourly average [O3] greater than or equal to 60 ppb (SUM06; Fig. 2e,f).

Figure 2.

Ozone treatment at the SoyFACE facility in 2007 and 2008. Eight-hour daily mean [O3] in 2007 (a) and 2008 (b). AOT40 (average concentration over a threshold of 40 ppb) in 2007 (c) and 2008 (d) and Sum06 (sum of exposure over 60 ppb) in 2007 (e) and 2008 (f). Grey lines show the ambient treatment, and black lines show the elevated [O3] treatment. Cumulative metrics, AOT40 and Sum06 were calculated according to Mauzerall & Wang (2001).

In 2007, 10 indeterminate soybean cultivars (Table 1) were planted in plots eight rows wide and 2.7 m long, with 0.38 m row spacing on 22 May (DOY 142). In 2008, six of those cultivars (Table 1) were planted on 17 June (DOY 169) in plots that were eight rows wide and 5.4 m long, with 0.38 m spacing. After maturity, the plots were trimmed to 2.1 m in 2007 and 4.9 m in 2008 and the centre six rows were harvested on 2 October 2007 and 29 October 2008 to measure seed yield. Ozone fumigation began on 4 June 2007 (DOY 155) and 11 July 2008 (DOY 193) and ended on 21 September 2007 (DOY 264) and 13 October 2008 (DOY 292). Average daytime maximum temperature was 28.6 and 26.7 °C during the growing season in 2007 and 2008, with night-time minimum temperature averaging 16.0 and 15.4 °C in 2007 and 2008, respectively (Fig. 1a,b). Average maximum PPFD was 1885 and 1778 µmol m−2 s−1 in 2007 and 2008 (Fig. 1c,d), and total seasonal precipitation was 233 and 432 mm in 2007 and 2008, respectively (Fig. 1e,f).

Table 1.  List and description of soybean cultivars used in the study
CultivarYear of releaseMaturity groupFemale parentMale parent
  1. Cultivars shown in italics were only grown in 2007.

HS93-41182000IVIA 2007DSR 304
IA 30101998IIIJ285S29-39

LAI, relative chlorophyll content, photosynthetic gas exchange and chlorophyll fluorescence

Throughout the growing seasons of 2007 and 2008, LAI, relative chlorophyll content, photosynthetic gas exchange and chlorophyll fluorescence were measured. Approximately every two weeks, LAI was measured non-destructively with a plant canopy analyser (LAI-2000, Li-Cor, Lincoln, NE, USA), which calculates LAI using a fisheye optical sensor that measures radiation attenuation through the canopy (Welles & Norman 1991). Measurements were made just after sunrise or before sunset when radiation was diffuse. Readings for each plot were made along a pair of diagonal transects between the rows using the 45° view cap. Each transect consisted of one above canopy reading followed by three below canopy readings made at evenly spaced intervals along and across the row. The first transect was made with the sensor looking parallel to the row, and the second transect was made with the sensor looking perpendicular to the row.

Every two weeks, relative chlorophyll content was measured on three fully expanded leaves at the top of the canopy within each cultivar and plot with a SPAD meter (Minolta SPAD-502, Osaka, Japan). In order to calibrate the SPAD values with actual chlorophyll content, SPAD readings and leaf samples from 60 randomly selected plants from all genotypes were taken. Chlorophyll was extracted from leaves in chilled 100% methanol (Porra, Thompson & Kreidemann 1989). Absorbance of methanol extracts was measured in a 96-well plate reader (HT-Synergy, Bio-Tek, Winooski, VT, USA) at 666, 653 and 470 nm. Chlorophyll a and b content were calculated according to Lichtenthaler (1987). The relationship between SPAD and total chlorophyll content was best described by an exponential function, y = 0.089e(0.0411x), where y is the chlorophyll content (g m−2) and x is the SPAD reading (Uddling et al. 2007). This function was subsequently used to calculate chlorophyll content from the SPAD readings.

Midday gas exchange and chlorophyll fluorescence of fully expanded leaves at the top of the canopy were measured on four dates in 2007, spanning vegetative and reproductive development, using four to six open gas exchange systems with integrated modulated chlorophyll fluorometers (LI-6400 and LI-6400-40; Li-Cor) following the methods of Bernacchi et al. (2006) and Leakey et al. (2006). Gas exchange systems were calibrated every 30–45 d using certified gas with a known CO2 concentration with 21% oxygen and nitrogen as balance (S.J. Smith, Champaign, IL, USA), and controlled water vapour concentrations generated with a humidification system (LI-610 Portable Dew Point Generator; Li-Cor). In the field, one gas exchange system was operated within each of the four experimental blocks, which consisted of one ambient and one elevated [O3] treatment. Two systems were first used in ambient plots, while the other two were first used in the elevated [O3] plots. Each system was then moved to the alternate [O3] treatment within the block. Three plants of each cultivar were measured in each plot. Measurements of chlorophyll fluorescence and gas exchange were made at ambient [CO2] (∼380 ppm), ambient air temperature and incident PPFD. Leaf photosynthesis (A), gs, and intercellular [CO2] (ci) were calculated using the equations of von Caemmerer & Farquhar (1981). The photochemical efficiency of photosystem II (Fq/Fm) was determined by measuring steady-state fluorescence and maximum fluorescence during a light-saturating pulse of ∼6500 µmol m−2 s−1 following the procedures of Genty, Briantais & Baker (1989). The efficiency of CO2 assimilation (ΦCO2) was determined according to Naidu & Long (2004) using average leaf absorbance measurements determined at SoyFACE by Rascher et al. (2010).

In 2008, gas exchange was measured during vegetative growth on 31 July 2008 (vegetative stage 7; Fehr et al. 1971). The most recently fully expanded leaves at the top of the canopy and also more mature leaves, approximately two nodes down from the most recently fully expanded leaves were measured. Both cohorts of leaves were at or near the top of the canopy, receiving direct light.

Tissue sampling and biochemical analyses

Leaf tissue samples for measuring total antioxidant capacity and phenolic content were taken once during vegetative growth (6 July 2007, DOY 187) and once during the seed filling stage of reproductive growth (21 August 2007, DOY 233). Leaf discs (∼1.2 cm2) were excised from fully expanded leaves at the top of the canopy, plunged immediately into liquid N and then stored at −80 °C. Five plants per cultivar per plot were sampled for total antioxidant capacity measurements and three plants per cultivar per plot were sampled for phenolic content. Additional discs from the same plants were removed and oven-dried at 70 °C for calculation of specific leaf mass.

Total antioxidant capacity was assessed with the oxygen radical absorbance capacity (ORAC) assay, which measures antioxidant inhibition of peroxyl radical-induced oxidations, according to the methods of Gillespie, Chae & Ainsworth (2007). Total phenolic content was measured with a Folin–Ciocalteu assay following the protocol of Ainsworth & Gillespie (2007).


Statistics were performed on plot means for all variables. The effect of elevated [O3] on different cultivars was analysed with a randomized complete block split-plot mixed model analysis of variance (anova), with the Satterthwaite option (Proc MIXED, SAS 9.1, SAS Institute, Cary, NC). In all tests, [O3] and cultivar were fixed effects and block a random effect. In tests of physiological and biochemical variables, day of year was considered as a fixed effect, although when analysed with repeated measures, the statistical outputs were similar. For the 2008 photosynthesis data, leaf age was also considered as a fixed effect in the anova model. The relationship between seed yield and physiological or biochemical variables, and seed yield and AOT40 was assessed by linear regression (Proc REG, SAS 9.1).


Seed yield

Elevated [O3] significantly reduced soybean seed yield from 3237 to 2782 kg ha−1 (17.2%) when averaged across six cultivars and two growing seasons (F = 102.9, P < 0.0001; Table 2). Yields were significantly higher in 2008 compared to 2007 (F = 111.0, P < 0.0001). When both years were included in the statistical model, there was not a statistically significant cultivar × O3 (F = 1.81, P = 0.12) or year × cultivar × O3 interaction (F = 1.08, P = 0.378). However, in 2007, when 10 cultivars were available for the analysis, there was a significant cultivar × O3 interaction for the seed yield response of soybean cultivars to elevated [O3] (F = 2.12, P < 0.05), with yield decreases ranging from 11% in Loda to 37% in IA-3010 (Table 2). In both 2007 and 2008, IA-3010 was the most sensitive to elevated [O3], but there was little consistency in the ranks of the other cultivars (Table 2).

Table 2.  Seed yield (kg ha-1) of soybean cultivars exposed to ambient and elevated [O3] in 2007 and 2008
Ambient [O3]Elevated [O3]% ChangeAmbient [O3]Elevated [O3]% ChangeAmbient [O3]Elevated [O3]% Change
  1. Significant differences between ambient and elevated [O3] within a cultivar and year are shown with asterisks. *P < 0.05, **P < 0.01, ***P < 0.001.

A31272667 ± 1812108 ± 75**21.0      
Clark2100 ± 1021564 ± 159**25.5      
Dwight3052 ± 2342306 ± 173***24.43502 ± 1193091 ± 130*11.73277 ± 1482699 ± 179***17.6
Holt2319 ± 931911 ± 216*17.6      
HS93-41182956 ± 1732273 ± 86***23.13554 ± 1483286 ± 1657.53255 ± 1552779 ± 210***14.6
IA-30103433 ± 1112170 ± 149***36.83353 ± 1442815 ± 115**16.03393 ± 862493 ± 150***26.5
LN97-150762817 ± 672093 ± 119***25.73443 ± 1003079 ± 143*10.63130 ± 1312586 ± 205***17.4
Loda2763 ± 612452 ± 9711.33400 ± 1762983 ± 131*12.33082 ± 1482718 ± 125**11.8
NE33992448 ± 791777 ± 274***27.4      
Pana3113 ± 1972552 ± 71**18.03451 ± 663010 ± 106*12.83282 ± 1162782 ± 125***15.2
Average2767 ± 1262121 ± 96***23.13451 ± 292906 ± 63***11.83237 ± 462782 ± 105***17.2

Exposure response relationships were analysed for seed yield versus AOT40 for the six cultivars that were studied in both 2007 and 2008 (Fig. 3). The slope of the linear regression indicated that IA-3010 had the greatest yield loss per increase in AOT 40, supporting the results from the anova. The yield response relationships also indicated that Loda and Pana were more tolerant to [O3] than IA-3010, and the slopes of the regressions were different from the slope of IA-3010 (P < 0.1).

Figure 3.

AOT-40-yield response functions for six soybean cultivars grown in 2007 (circles) and 2008 (triangles). White symbols show ambient [O3], and black symbols show elevated [O3]. Relationships between AOT40 and yield were determined by linear regression.

Physiological responses

All soybeans cultivars in both ambient and elevated [O3] reached peak LAI on approximately 1 August 2007 (DOY 214; Supporting Information Fig. S1) and 17 August 2008 (DOY 230; Supporting Information Fig. S2). There was a significant effect of O3 on LAI when averaged across all time points and cultivars in 2007 (Table 3), but peak LAI was not affected by elevated [O3] in any cultivar except Holt. In contrast, in 2008, peak LAI was reduced by elevated [O3] in IA-3010, LN97-15076, Loda and Pana (Supporting Information Fig. S2). There was also a significant interaction of [O3] × time in 2007 (Table 3). Dwight, Holt, HS93-41118, LN97-15076, Loda and NE3399 showed significant decreases in LAI after the peak in August, while Holt and IA3010 showed reduced LAI earlier in the season (Supporting Information Fig. S1). Other cultivars (A3127, Clark and Pana) showed very little response of LAI to elevated [O3] in 2007 (Supporting Information Fig. S1). Similar variability in cultivar LAI responses was measured in 2008 (Supporting Information Fig. S2).

Table 3.  Analysis of variance of leaf area index (LAI), chlorophyll content, photosynthetic rate (A), stomatal conductance to water vapour (gs), operating efficiency of photosystem II (Fq/Fm), efficiency of CO2 assimilation (ΦCO2), total antioxidant capacity (ORAC) and total phenolic content measured in 2007
F, PF, PF, PF, PF, PF, PF, PF, P
  1. A mixed model analysis of variance was used with [O3], cultivar and time as fixed effects and block as a random effect. Significant effects are shown in bold.

[O3]33.42, 0.00145.98, <0.0124.21, 0.01624.07, <0.0016.13, 0.08023.64, 0.0161.46, 0.3101.65, 0.205
Cultivar27.08, <0.0017.78, <0.0011.94, 0.0491.86, 0.0590.82, 0.6031.94, 0.0481.65, 0.1252.66, 0.013
[O3] × Cultivar0.48, 0.8902.08, 0.0470.96, 0.4780.97, 0.4700.39, 0.9400.83, 0.5880.88, 0.5530.88, 0.553
Time742.5, <0.001298.4, <0.001103.9, <0.001160.3, <0.001143.6, <0.001226.4, <0.0015.94, 0.0180.81, 0.371
[O3] × Time2.12, <0.00110.44, <0.0018.26, <0.0010.84, 0.6955.11, 0.0027.77, <0.0016.33, 0.0154.49, 0.039
Cultivar × Time2.99, 0.0057.07, <0.0011.83, 0.0110.55, 0.5491.27, 0.1771.72, 0.0192.33, 0.0261.12, 0.366
[O3] × Cultivar × Time0.58, 0.9931.21, 0.1590.54, 0.9700.54, 0.9720.46, 0.9910.55, 0.9651.11, 0.3720.68, 0.724

Chlorophyll content in upper canopy leaves peaked in August (DOY 220-240) for all cultivars in 2007 and slightly later (DOY 240-260) in 2008 (Supporting Information Figs S3 & S4). At the beginning of September 2007 (DOY 248), significant differences in chlorophyll content and senescence between ambient and elevated [O3] were apparent in some cultivars (Supporting Information Fig. S3). There was significant intraspecific variation in the response of relative chlorophyll content (Table 3). Leaves of Dwight, IA3010, LN97-15076, Loda, NE3399 and Pana all showed increased rates of chlorophyll degradation at the end of the growing season when grown at elevated [O3] compared to ambient conditions (Supporting Information Fig. S3). A3127, Clark and Holt showed no significant differences in relative chlorophyll content at any time in the growing season (Supporting Information Fig. S3). Similar trends in chlorophyll content were measured in 2008 (Supporting Information Fig. S4).

Elevated [O3] reduced net carbon assimilation (A) and gs by 11 and 15%, respectively, when averaged across all cultivars and time points in 2007 (Fig. 4, Table 3). There was a trend towards a decrease in Fq/Fm at elevated [O3], and elevated [O3] resulted in a significant decrease in ΦCO2 (Table 3). Cultivars also differed in their photosynthetic rate (Table 3); however, there was no significant [O3]–cultivar interaction in any of the gas exchange or fluorescence parameters (Table 3). Photosynthetic differences between ambient and elevated [O3] were greatest at the end of the growing season when there was evidence for increased rates of senescence in the elevated [O3] plots (Fig. 4).

Figure 4.

Photosynthetic carbon uptake of 10 soybean cultivars exposed to ambient (white symbols) or elevated [O3] (black symbols) over the 2007 grown season. Mean values ± standard error are shown, and significant differences within a cultivar and time point are illustrated with asterisks (*P < 0.05, **P < 0.01).

In 2008, A was measured in two leaf cohorts during vegetative growth, and there was a significant [O3]–leaf cohort interaction (F = 13.01, P < 0.001; Fig. 6). In ambient [O3], there was no difference in A in the youngest most fully expanded leaves and in older leaves two nodes down, while in elevated [O3], the older leaves had significantly lower rates of A (Fig. 5). While the three-way interaction of cultivar × [O3] × leaf cohort was not significant, individual pairwise comparisons within cultivars indicated that IA-3010, the most sensitive cultivar in terms of seed yield, was also relatively sensitive to [O3] in terms of photosynthetic capacity (Fig. 5).

Figure 6.

Linear regression of seed yield (kg ha−1) versus physiological and biochemical parameters and photosynthesis versus total antioxidant capacity. Chlorophyll content was measured on 5 September 2007. Photosynthetic rate (A), stomatal conductance to water vapour (gs) and total antioxidant capacity of leaf extracts [oxygen radical absorbance capacity (ORAC)] measured as Trolox equivalents (TE) were taken on 21 August 2007. A in lower canopy leaves was measured on 31 July 2008. Open symbols represent mean values for each cultivar at ambient [O3], and closed symbols represent mean values for each cultivar at elevated [O3].

Figure 5.

Photosynthetic carbon uptake of six soybean cultivars exposed to ambient (white bars) or elevated [O3] (grey bars) measured on 31 July 2008. The most recently fully expanded leaves (open bars) and older leaves, approximately two nodes down (hatched bars) were measured. Mean values ± standard error are shown, and significant differences within a cultivar are illustrated with different letters.

Antioxidant capacity and total phenolic content

The total antioxidant capacity and total phenolic content of sunlit leaves of soybean cultivars was measured once during vegetative growth (6 July 2007) and once during reproductive growth (21 August 2007; Table 4). There was a significant effect of time on the total antioxidant capacity and total phenolic content (Table 3). There was also a significant time × [O3] interaction on total antioxidant capacity (Table 3). During vegetative growth, total antioxidant capacity was not affected by growth at elevated [O3], but in three of the 10 cultivars, total antioxidant capacity was higher in leaves exposed to elevated [O3] when sampled during reproductive growth (Table 4). Cultivars also showed significant differences in total phenolic content, but the O3 treatment had no significant effect on total phenolic content (Table 3).

Table 4.  Antioxidant capacity of soybean cultivars exposed to ambient and elevated [O3]
CultivarTotal antioxidant capacity (mmol trolox equivalents/g DW)Total phenolic content (mmol gallic acid equivalents/g DW)
6 July 200721 Aug 20076 July 200721 Aug 2007
ControlElevated [O3]ControlElevated [O3]ControlElevated [O3]ControlElevated [O3]
  1. Total antioxidant capacity was measured with the ORAC assay as Trolox equivalents per gram of dry weight. Total phenolic content was measured with a Folin–Ciocalteu assay and expressed as gallic acid equivalents per gram of dry weight. Samples were taken once during vegetative growth (6 July 2007) and once during reproductive growth (21 August 2007). Significant differences between control and elevated [O3] within a cultivar and time point are shown in bold text.

A31270.71 ± 0.030.73 ± 0.030.54 ± 0.120.58 ± 0.090.152 ± 0.0130.140 ± 0.0360.134 ± 0.0090.148 ± 0.003
Clark0.68 ± 0.100.66 ± 0.140.69 ± 0.060.67 ± 0.090.177 ± 0.0250.147 ± 0.0180.138 ± 0.0020.111 ± 0.033
Dwight0.75 ± 0.090.69 ± 0.100.50 ± 0.050.72 ± 0.110.115 ± 0.0140.108 ± 0.0070.140 ± 0.0030.119 ± 0.015
Holt0.66 ± 0.020.77 ± 0.090.71 ± 0.110.69 ± 0.120.094 ± 0.0250.099 ± 0.0220.115 ± 0.0080.153 ± 0.004
HS93-41180.67 ± 0.130.59 ± 0.130.63 ± 0.100.85 ± 0.040.150 ± 0.0400.141 ± 0.0310.155 ± 0.0060.153 ± 0.009
IA 30100.50 ± 0.130.60 ± 0.050.55 ± 0.070.60 ± 0.110.152 ± 0.0290.136 ± 0.0370.138 ± 0.0030.135 ± 0.006
LN97-150760.71 ± 0.080.65 ± 0.030.63 ± 0.100.79 ± 0.090.138 ± 0.0140.126 ± 0.0250.147 ± 0.0070.130 ± 0.016
Loda0.57 ± 0.040.49 ± 0.070.60 ± 0.110.73 ± 0.120.119 ± 0.0290.069 ± 0.0030.097 ± 0.0100.120 ± 0.007
NE33990.57 ± 0.070.77 ± 0.050.60 ± 0.110.83 ± 0.110.137 ± 0.0300.141 ± 0.0180.125 ± 0.0070.162 ± 0.013
Pana0.82 ± 0.030.75 ± 0.020.56 ± 0.100.67 ± 0.080.148 ± 0.0580.079 ± 0.0010.120 ± 0.0050.120 ± 0.003

Correlations with seed yield

A number of variables showed significant correlations with seed yield (Fig. 6). Late-season measurements of relative chlorophyll content, gs, A and Fq/Fm showed positive correlations with seed yield in 2007 (Fig. 6). Total antioxidant capacity (ORAC), when measured during reproductive growth (21 August 2007), showed a significant, negative correlation with both photosynthetic rate and seed yield in 2007 (Fig. 6), indicating a trade-off between antioxidant metabolism and carbon gain. In 2008, A of older leaves measured during vegetative growth was significantly correlated with seed yield (Fig. 6).


The SoyFACE experimental site provided the opportunity to grow 10 soybean cultivars under field conditions at elevated [O3]. Under these unique conditions, significant variability among these cultivars was demonstrated, indicating the presence of genetic variation in soybean seed yield tolerance to [O3]. This is consistent with the findings of previous studies, many of which have compared soybean yields at elevated [O3] to yield in charcoal filtered air (Mulchi et al. 1988; Heagle et al. 1998; Morgan et al. 2003; Jaoude et al. 2008). A recent investigation of 30 ancestral soybean lines reported significant variability in foliar injury with growth at elevated [O3], with Midwestern cultivars predicted to have a wider spread of [O3] injury than Southern cultivars (Burkey & Carter 2009). Our study of 10 Midwestern cultivars revealed that an elevation of [O3] to 82 ppb significantly reduced soybean seed yield by an average of 23% across all 10 cultivars. In 2008, yields were not reduced by as much, only 12%, but the season average elevated [O3] was more than 20 ppb lower than in 2007. In both years of the study, IA-3010 was the most sensitive cultivar to [O3], with reductions in yield of 1264.5 kg ha−1 in 2007 and 537.5 kg ha−1 in 2008. Regression analysis of the exposure response of IA-3010 also indicated that it was the most sensitive to the O3 concentrations applied in this study (Fig. 3). The exposure response analysis also indicated that Loda and Pana were more tolerant to [O3] than IA-3010.

Several measured physiological and biochemical variables showed significant correlation with absolute seed yield, including A, gs, chlorophyll content and total antioxidant capacity (Fig. 6). However, these variables were measured late in the growing season, and therefore, have limited utility as markers of O3 tolerance. In 2008, photosynthesis was measured on two cohorts of leaves early in the growing season, and a significant correlation between seed yield and photosynthesis in older leaves was apparent (Fig. 6). Plants were at approximately the V7 stage of development when measurements were taken in 2008, but it is possible that A in mature leaves may correlate with yield even earlier in the growing season. Therefore, analysis of mature leaves on young plants may provide a good system for investigating the mechanisms of O3 tolerance. A previous analysis of 12 soybean cultivars from maturity groups III, IV and V demonstrated significant correlations among net assimilation rate, relative growth rate and leaf expansion rate in beans exposed to elevated [O3], but none of those parameters correlated with percent change in seed yield (Mulchi et al. 1988). It is possible that the measures taken by Mulchi et al. (1988) were less sensitive than the measures used in the current study. Additionally, many of the parameters in the current study that correlated with yield were measured later in the growing season, or on older leaves, which had been exposed to the ozone treatment for a longer period of time.

In 2007, the effect of O3 on photosynthesis was greatest at the end of the growing season when eight of the 10 cultivars had significantly lower photosynthetic rates at elevated [O3]. In 2008, older leaves had a lower rate of photosynthesis on average across six cultivars during vegetative growth (Fig. 4). Previous studies have also found that photosynthetic damage was more pronounced when plants and/or leaves were older (Mulchi et al. 1992; Nie, Tomasevic & Baker 1993; McKee et al. 1995; Booker et al. 1997; Fiscus et al. 1997; Morgan et al. 2006) and elevated [O3] accelerated the process of senescence (Pell, Schlagnhaufer & Arteca 1997; Miller, Heagle & Pursley 1998; Long & Naidu 2002). The chlorophyll data support the observation that elevated [O3] accelerated senescence in the majority of cultivars that were investigated in 2007 and 2008 (Supporting Information Figs S1–S4). Further evidence for the role of accelerated senescence in determining the response of photosynthesis to elevated [O3] comes from the cv. Clark, in which photosynthetic rate was not significantly different in Clark when measured at ambient and elevated [O3] at the end of the season, nor was there any change in leaf chlorophyll content.

Three cultivars with the greatest yield loss in elevated [O3] in 2007, IA-3010, LN97-15076 and NE3399, also showed significant decreases in photosynthetic carbon gain at the end of the season (Fig. 5). These decreases in photosynthesis occurred during seed filling simultaneous with accelerated senescence in all three cultivars, and decreases in LAI in LN97-15076 and NE3399. Fewer live leaves and decreased photosynthesis in the late grain filling period appeared to drive the O3-induced losses in production and yield in another Midwestern soybean cultivar (Pioneer 93B15; Morgan et al. 2006). Surprisingly, many of the more tolerant cultivars also showed decreases in photosynthesis, chlorophyll content and/or LAI during seed filling. In fact, Loda, the most tolerant cultivar in 2007, showed decreases in all three parameters. This suggests that there are various mechanisms of tolerance to O3 within the soybean germ plasm. Overall, the data support the hypothesis that relatively subtle depressions in A accumulate over the season to reduce soybean capacity for remobilization of photosynthate and protein for seed filling.

Total antioxidant capacity (ORAC), when measured during reproductive growth (21 August 2007, DOY 233), showed a significant, negative correlation with photosynthesis and seed yield (Fig. 6), suggesting a trade-off between antioxidant metabolism and carbon gain. Previous work with wheat showed that ΦCO2 was decreased in old leaf tissue, while Fq/Fm increased, possibly to provide increased energy for reducing equivalents for free radical-scavenging systems (Nie et al. 1993). A positive intercept of the relationship between Fq/Fm versus ΦCO2 under photorespiratory conditions is consistent with alternative electron sinks, and the intercept was significantly higher (0.035 versus 0.064) in soybean grown at elevated [O3] (Supporting Information Fig. S5). While photorespiration likely constitutes the largest alternative electron sink (Ort & Baker 2002), it seems unlikely that photorespiration would increase at elevated [O3], and recent work suggests that O3 inhibits photorespiration in some species (Bagard et al. 2008). Therefore, the shift in the relationship between Fq/Fm versus ΦCO2 supports the hypothesis that there is a trade-off between energy expended on carbon gain and antioxidant metabolism.

The total antioxidant capacity of cells is collectively constituted by the range of enzymatic and non-enzymatic antioxidant mechanisms that plants have evolved (Larson 1988; Ghisseli et al. 2000). Apoplastic antioxidant capacity is a primary protectant from O3 damage (Kangasjarvi, Jaspers & Kollist 2005) and a common response to O3 is an increase in a plant's capacity to scavenge active oxygen species, both inside and outside the cell (Long & Naidu 2002). Variation among species and genotypes in tolerance to various environmental stresses has been linked to leaf antioxidant capacity and metabolism (Blokhina, Virolainen & Fagerstedt 2003; Scebba et al. 2003; Huang & Guo 2005; Nayyar & Gupta 2006). ORAC is a measure of general antioxidant capacity (Cao, Alessio & Cutler 1993) and has the potential to be a useful phenotypic marker of stress tolerance (Gillespie et al. 2007), although it has so far only been predictive when measured late in the growing season. Total phenolic content was not significantly correlated with yield across cultivars and O3 treatments, but there was a significant reduction in total phenolic content in two of the more tolerant cultivars early in the growing season (Table 3). Therefore, there may be some potential for developing that assay as an early-season marker of tolerance.

In conclusion, this study demonstrated significant intraspecific variability in the soybean yield response to elevated [O3]. Many physiological and biochemical parameters correlated with soybean yield when measured late in the growing season, which is not ideal for rapid screening of germ plasm, but photosynthetic carbon uptake in older leaves measured during vegetative growth correlated with seed yield in 2008. Previous study of a single soybean cultivar found that leaves formed during the vegetative growth stage did not show a significant O3-induced loss of photosynthetic capacity as they aged (Morgan et al. 2004), but this is likely due to variability in O3 exposure and dose within a growing season. The O3 exposure response relationships determined in field-grown soybean cultivars in this study indicate that there has not been a significant improvement in O3 tolerance since the National Crop Loss Assessment Network assessments in the 1980s (Heagle 1989).


We thank Don Ort, Steve Long, Andrew Leakey, Tim Mies, Jamie Howard, Jesse McGrath, Stacy Alikakos, Jessica Chiang, Tiphaine Feray, Sharon Gray, Sara Kammlade, Cody Markelz, Matt Nantie, Katie Richter, David Rosenthal, Chris Rudisill, Matt Siebers, Reid Strellner and Diana Umali for assistance at SoyFACE. We acknowledge the editor and two anonymous reviewers for helpful suggestions for improving the original manuscript. This research and the SoyFACE experimental facility were supported by the Illinois Council for Food and Agricultural Research, USDA ARS and the Archer Daniels Midland Company.