Author for correspondence: Andreas Heinemeyer Tel: +44 1904 43 2991 Fax: +44 1904 432898 Email: firstname.lastname@example.org
• Although arbuscular mycorrhizal (AM) fungi are a major pathway in the global carbon cycle, their basic biology and, in particular, their respiratory response to temperature remain obscure.
• A pulse label of the stable isotope 13C was applied to Plantago lanceolata, either uninoculated or inoculated with the AM fungus Glomus mosseae. The extra-radical mycelium (ERM) of the fungus was allowed to grow into a separate hyphal compartment excluding roots. We determined the carbon costs of the ERM and tested for a direct temperature effect on its respiration by measuring total carbon and the 13C:12C ratio of respired CO2. With a second pulse we tested for acclimation of ERM respiration after 2 wk of soil warming.
• Root colonization remained unchanged between the two pulses but warming the hyphal compartment increased ERM length. δ13C signals peaked within the first 10 h and were higher in mycorrhizal treatments. The concentration of CO2 in the gas samples fluctuated diurnally and was highest in the mycorrhizal treatments but was unaffected by temperature. Heating increased ERM respiration only after the first pulse and reduced specific ERM respiration rates after the second pulse; however, both pulses strongly depended on radiation flux.
• The results indicate a fast ERM acclimation to temperature, and that light is the key factor controlling carbon allocation to the fungus.
Elevated CO2 concentrations in the atmosphere are leading to an increase in global mean temperature, and substantial climatic changes are expected over the next century (Houghton et al., 2001). Elevated CO2 generally increases plant growth (Ceulemans et al., 1999; Norby & Luo, 2004). Temperature effects, and especially the effects of soil temperature, have been less studied in the context of global environmental change.
Most experiments on the impacts of rising temperature on ecosystems have been conducted under controlled conditions and have used nonmycorrhizal plants, although most plants live in association with mycorrhizal fungi (Trappe, 1987) and two-thirds of plants are in symbiosis with arbuscular mycorrhizal (AM) fungi (Fitter & Moyersoen, 1996). AM fungi (Glomeromycota) grow both inside roots and in soil as an extra-radical mycelium (ERM), which directly experiences wide variations in soil environment such as pH, soil moisture and temperature. The AM symbiosis appeared contemporaneously with land plants (Remy et al., 1994) and must therefore have experienced major climate changes in the past. However, we still do not understand the basic ecology of this ubiquitous symbiosis (Fitter et al., 2000; Bever et al., 2001; Fitter et al., 2004). The fungi are obligate symbionts but form a mutual relationship with the host plant in which the fungus receives its entire carbon requirement from the host plant but provides the host with nutrients that are poorly mobile in soil, such as phosphate, or with other benefits (Newsham et al., 1995; Clark & Zeto, 2000). Hence, this symbiosis plays a key role in linking above- and below-ground carbon cycling (Finlay & Söderström, 1992).
Most of the carbon supply to the fungus seems to be recently fixed carbon (Jakobsen & Rosendahl, 1990), and the biochemical pathways involved are becoming clear (Bago et al., 2003), yet we do not know whether plants regulate this process (Graham et al., 1997). Changes in the environment that affect plant photosynthesis, for example reduced light availability (Tester et al., 1986), affect fungal growth indirectly. Yet the fungus itself might respond to changes in the soil environment directly, independently of host plant responses; soil temperature is likely to be an important factor affecting growth and respiration. As fungal carbon supply can consume up to 20% of net photosynthesis (Smith & Read, 1997), any temperature responses of fungal growth and respiration might therefore have a significant influence on carbon input and cycling in the soil.
Elevated CO2 often increases internal colonization as a result of an increase in plant growth. Consequently, the amount of ERM in the soil may also increase (Staddon et al., 1998). Temperature also increases internal colonization, both indirectly, through increased plant growth (Staddon et al., 2002), and directly, through the fungal response (Heinemeyer & Fitter, 2004). Yet, to date, hyphal respiration rates have never been measured directly for AM fungi (Rillig & Allen, 1999) and it is not known how much ERM respiration contributes to the carbon budget of the plant and whether it shows similar acclimation to temperature changes as do roots (Gifford, 1995; Atkin et al., 2000) and soil respiration (Luo et al., 2001).
Ecological studies increasingly use the stable isotope 13C to follow a carbon signal through different trophic layers (Ostle et al., 2000; Radajewski et al., 2000), including AM fungi (Staddon et al., 1999a; Miller & Kling, 2000). We combined a compartment study with the application of two 13CO2 pulse labels during a period of warming the ERM. We aimed to measure (i) the sensitivity of the respiration of the ERM of an AM fungus to temperature and whether it displayed acclimation; (ii) how much of the carbon allocated to the fungus from the root carbon pool is recently fixed carbon; and (iii) the carbon costs for the host plant of the ERM and the dynamics of carbon flux from the roots to the ERM.
Materials and Methods
Ten open-top Perspex boxes (23 × 23 × 10 cm) were divided in half using Perspex plates. These halves were each further divided into two compartments (A and B) with nylon mesh (pore size 20 µm; Staniar P/N°25T11-20, John Staniar & Co., Manchester, UK), allowing passage of extra-radical hyphae only, not roots, into hyphal compartment B (Fig. 1). The growth medium in each compartment (1200 g) consisted of a 1 : 1 mixture [volume/volume (v/v)] of coarse builders’ dried silica sand and Terragreen® (an attapulgite clay soil conditioner; Turfpro Ltd, Staines, UK). Bonemeal was mixed into the medium as a long-term phosphate source (0.25 g l−1 soil). After germination on moist filter paper (Whatman® no. 1) for 2 d (20°C in the dark), two ribwort plantain (Plantago lanceolata L.) seedlings (Emorsgate Seeds, Norfolk, UK) were planted at the same position in each plant compartment A, 3.5 cm from the mesh, and thinned to one seedling after 2 wk. Compartment A was evenly inoculated (150 g kg−1 soil) with fresh inoculum of Glomus mosseae ((Nicol. & Gerd.) Gerd. & Trappe, isolate UY 316, ex P. Bonfante, Torino) consisting of c. 1 cm pieces of P. lanceolata and Trifolium repens (L.) roots within the same substrate. Control pots received the same amount of uninoculated roots of the same species in the same substrate, and 150 ml of a filtrate (Whatman® no. 3M) of a further batch of the inoculum. Boxes were placed in a glasshouse for the first 61 d after planting (dap) at 14 : 25 ± 3°C (day:night), 60–80% relative humidity and a mean photosynthetically active radiation (PAR) of 450 ± 100 µmol m−2 s−1 over the waveband 400–700 nm at plant level for 16 h d−1. Supplementary light was supplied by six mercury vapour lamps (400 W) giving 200 µmol m−2 s−1 (400–700 nm) at plant level during early morning and evening. PAR was measured twice weekly throughout the day at plant level using a light meter (SKD210; Skye Instruments Ltd, Powys, UK). In addition, an automatic weather station (Delta-T Devices Ltd, Cambridge, UK) located next to the glasshouse recorded continuous measurements of exterior PAR flux (µmol m−2 s−1 (400–700 nm)).
Soil temperature at 3 cm depth in both compartments of four boxes was recorded every 10 min (averaged 30-s readings) at 1, 3 and 5 cm distance from the mesh with thermistor probes (Grant Instruments Ltd, Shepreth, UK), and both compartments were watered daily with deionized water as required. From week 3 onwards a half-strength Rorison's nutrient solution (Hewitt, 1966) (containing 1/10 phosphate) was given twice weekly (10 ml) to compartment A, except within 1 cm of the mesh; all compartments were subsequently watered with deionized water to remove any nutrient solution from the leaves.
Experimental set-up and soil heating
There were six treatments with three replicates each: 13C-labelled mycorrhizal and nonmycorrhizal treatments and an unlabelled mycorrhizal treatment, all with and without soil heating. Unlabelled mycorrhizal treatments were needed to assess background 13C contamination during the two pulse labels, to monitor mycorrhizal effects on plant growth, and to provide a blank for calculating extra-radical mycelium (ERM) 13C respiration. Two spare boxes were used to determine when sufficient ERM had been produced in the hyphal compartment to start treatments. Soil in the hyphal compartment had a soil warming cable (Macpenny® Cameron; East Riding Horticulture, Sutton-on-Derwent, UK) inserted with four vertical bends at a distance of 7.5 cm from the mesh (Fig. 1). In control compartments, the cable was not connected to a power supply. To avoid solar heating of hyphal compartments and algal growth, reflecting shields were attached to their outer surface. Soil warming started directly after the first pulse label (see the next section). Air temperatures in the glasshouse during the experimental phase were 12 : 23 ± 3°C (day:night). Heating flux across the mesh from the hyphal compartment had no significant effect on the plant compartment (less than 0.5°C directly behind the mesh) as shown previously (Heinemeyer & Fitter, 2004), but raised soil temperature at the gas sample tube position (see the next section) by approx. 6 ± 3°C. Any soil moisture differences induced by soil warming were avoided by frequently watering all compartments to field capacity.
Pulse label of δ13C and gas sampling technique
Two weeks before the first pulse label, one gas sample tube (diameter 2.1 cm and length 7.5 cm; Universal Container; Sterilin, Stone, UK) was inserted to a depth of 2.0 cm into each hyphal compartment, 4.5 cm from the mesh (Fig. 1), giving a total air volume of approx. 19 ml. On the day of the first pulse (61 dap; 20 April 2001), two fitted fans (Turbo-Fan; Roof Units Group, Dubley, UK) were switched on until the end of the experiment, ensuring a constant replacement of internal air with fresh air through a chimney from the top of the glasshouse. During the pulse label, each plant was covered with an acrylic chamber (diameter 10 cm; height 19.5 cm; volume 1530 cm3 (i.e. π × 5 cm2 × 19.5 cm) resting on two halves of an acrylic dish which were fitted around the stem of each plant (Fig. 1), with the joins between chamber and dish, and between dish halves, sealed with petroleum jelly (Philip Harris, Hyde, UK). Each chamber had an inflow (with a fitted disperser at the bottom) and an outflow tube at the top. A stable isotope delivery system (SID; Ostle et al., 2000) was situated 20 m from the glasshouse and 12 lines (13-mm-diameter garden hose) introduced air to the labelling chambers. Each line was linked to a 360-l gas mix tank with individual flow gauges. The gas was continuously mixed with CO2-scrubbed air to approx. 360 ppm CO2 with 50%δ13CO2 (Sigma Aldrich, Gillingham, UK) from a gas-tight 2-l bag. The six control plants received air from outside via two pumps (3 l min−1 for each channel) placed next to the fresh air inflow of the two fans. Outflow tubes were vented externally. All connections were sealed with gas-sealing teflon PTFE tape (B&Q, Eastleigh, UK). Labelling was performed from 12:00 to 15:30 h at a flow rate of approx. 2.5 l min−1, avoiding condensation inside the chambers. Outside air was scrubbed through the system for an additional 2 h before the chambers were removed. The second pulse was applied 2 wk after the first (75 dap; 4 May 2001), following the same procedure.
Suba Seal stoppers (Scientific Laboratory Supplies Ltd, Nottingham, UK) were fitted to the gas sample tubes 4 h before any sampling. Gas samples (14 ml) from the tubes were taken with syringes (25 ml) at 2 h before and 9.5, 21.5, 28.5 and 41 h after each pulse and used to over-pressure a 10-ml sealed Na-glass Exetainer® (cat #438 W; Labco Ltd, Wycombe, UK). For each sample, a new syringe and needle were used and each syringe was purged with 1 ml of the sample air before sampling. Stoppers were removed once the sample had been taken. Analytical determinations of CO2 concentration and 13C:12C ratios were made at the Natural Environment Research Council, UK (NERC) stable isotope facility (CEH-Merlewood). For details of the δ13C analysis, see Ostle et al. (2000). Weather conditions were nearly identical at the times of the two pulses, giving a mean air temperature of 25 ± 2°C and a mean peak daytime PAR of 500 ± 50 µmol m−2 s−1.
Plant and fungal measurements
One day before the first pulse label for each plant, the maximum width and length of all leaves were recorded and later referenced against a calibration curve (obtained from scanned leaves of the second pulse) to calculate an estimated leaf area, from the known fresh:dry weight ratio (see the end of this section), in order to estimate individual shoot biomass for the first pulse. Soon after the last gas sample of each pulse and also shortly before the second pulse label, leaf cores (each 0.2 cm2) were taken from five healthy, fully expanded leaves of each plant, then pooled and weighed both fresh and dry. These shoot samples and a subsample of dried roots were ground to a fine homogenous powder for determination of carbon content and 13C:12C ratios (see the previous section).
After gas sampling of each pulse, a soil core (diameter 1.9 cm; length 7.5 cm) was taken from the same position in each plant compartment. Root samples (< 1 mm diameter) from this soil core were stained and investigated for colonization (Staddon et al., 1998). Roots were cleared in 10% KOH (10 min) and stained (twice) in 0.1% acid fuchsin (35 min), in both cases in a waterbath (85°C). Percentages of total root length colonized (LRC), arbuscules (LRCarb) and vesicles (LRCves) were scored separately.
Soil for extraction of ERM was taken from directly beneath the inserted gas-collecting tubes, by using these same tubes to take soil cores (26 cm3). The sample area was then refilled with the same soil mix. The gas-collecting tubes of each compartment were then inserted next to this sample area as before. From each soil sample, two subsamples were taken and their fresh weights obtained, one for oven-drying to calculate moisture content and one for ERM extraction (Staddon et al., 1999b). ERM length could then be expressed as ERM length (m) per gram dry weight soil.
At 78 dap (7 May 2001), all plants were harvested and divided into root and shoot. After removal of dead leaves, total leaf and root fresh weights were recorded. Roots were then chopped into c. 1-cm pieces and mixed with a glass rod in 600 ml of water. Two subsamples were taken at random from this pool for measurement of root length and root fresh:dry weight ratio. Shoot dry weight (WS) and root dry weight (WR) were recorded after oven-drying at 65°C for 4 d. Leaf area and root length were measured from scanned images (WinRhizo®; Régent Instruments, Quebec, Canada). Specific leaf area (SLA) and specific root length (SRL) were both calculated. The relative growth rate (RGR) of shoot dry weight was estimated as the slope of ln WS over time.
Net photosynthesis of P. lanceolata plants was determined by continuously measuring the difference between inflow and outflow CO2 concentrations of each chamber during the first 2 h of each pulse label by taking chamber subsamples via a 6.5-mm-diameter PTFE tube fed back from the outflow lines to the SID-Infra-red gas analyzer (IRGA) (see Ostle et al., 2000).
Calculation of carbon costs to the plant
We calculated the contribution of ERM growth and respiration to plant carbon uptake, with the following assumptions.
1Rates of ERM growth were calculated as the difference in ERM length between the start and the end of the 2-wk period in both the unheated and the heated hyphal compartments and were assumed to have been constant during this period. No allowance was made for hyphal turnover.
[Rsample and Rref, 13C:12C ratios of sample and Vienna-Pee Dee Belemnite (V-PDB) reference (0.0111797; the reference used for the ‰13C calculation; see http://deuterium.nist.gov/standards.html); δsample, the δ13C of the sample; ppmsample, the concentration of CO2 (µl l−1) in the gas sample; Csample, the total carbon content of the sample (µg); Vsample, the volume of the sample tube (0.019 l); VM, the molar volume under standard conditions (22.4 l mol−1); MC, the molar mass of carbon (12.01 g mol−1).]
Respiration in each fungal compartment (µg h−1) was calculated per compartment, firstly as total carbon (ERMresp) (Eqn 5) and secondly as 13C (Eqn 6):
(Individual gas sample carbon contents (µg) were calculated for mycorrhizal (Cmyc) and nonmycorrhizal controls (Ccontrol), using total Csample (Eqn 4) and 13C content (13Cmyc; 13Ccontrol), respectively; mean values for nonmycorrhizal controls were subtracted from those for individual mycorrhizal compartments. Differences are expressed per hyphal compartment using substrate weight under the gas sample tube (Wsample= 31.43 g) and in the hyphal compartment (WFcomp = 1200 g) and corresponding ERM length density (m g−1). Substrate weights could be calculated by assuming a uniform substrate density of 1.21 g cm−3.)
Total 13C (mg) for shoot and root material could be calculated by multiplication of 13Cx (mg mg−1) per shoot or root mass (Eqn 7) by total Cx (mg) in individual shoot or root samples (Eqn 8):
13Cx = Wsample × atom% 13Csample/Wsample(Eqn 7)
Total Cx = Wx × %Csubsample/100(Eqn 8)
[atom% 13Csample, the calculated individual atom%13C (Eqn 3); Wsample (mg) and Wx (mg), sample and total shoot or root weight, respectively; %Csubsample, the percentage of total carbon in the corresponding shoot or root subsample material used for analysis.]
Statistical analysis was performed using SPSS v10.0 (Norusis, 1999). All data were checked and transformed appropriately to normalize skewed distributions before statistical analysis (i.e. log transformation for all but the percentage data, for which an arcsin transformation was chosen). Data for CO2 uptake, net photosynthesis, and plant and fungal growth were tested for differences between labelling treatments with a one-way analysis of variance (ANOVA) with labelling as the factor. Because unlabelled controls differed from labelled treatments only at the first pulse and only in CO2 uptake, all other parameters of unlabelled controls were no longer treated separately in further analyses. Effects on fungal growth were tested with a repeated measurement design of the general linear model (GLM) with time as the within-subject and temperature as the between-subject factor. Means of plant growth, plant total 13C, CO2 uptake and net photosynthesis were tested for differences with a two-way ANOVA, with temperature and mycorrhizal treatment as factors. Data for δ13C signals, CO2 concentrations in the gas samples, the proportion of second to first pulse label of 13C:12C in the gas samples and ERM respiration were tested for any treatment effects with the repeated measurement design (as above) with temperature and mycorrhizal treatments as the between-subject factors. For 13C ERM respiration, a one-way ANOVA with temperature as the factor was used to detect significant differences at particular harvests. Total respiration was tested with mean PAR received during a 12-h period before the sampling time as a covariate in a repeated measures design (as above) with temperature and label as factors, omitting the first sample of the first pulse without soil warming.
Temperature impact on ERM respiration
Extra-radical mycelium respiration was calculated from two variables measured in the gas samples, the concentration of CO2 and the δ13C signal. After the first pulse label, CO2 concentration showed diurnal fluctuation (Fig. 2a), being highest in the evening and lowest in the morning. CO2 concentration was higher in mycorrhizal treatments at both pulses. The δ13C of labelled treatments did not differ between temperature treatments in either pulse but was highest in labelled mycorrhizal treatments: it increased from a natural background of c. −13.0‰ to c. +1300‰ within the first 10 h and thereafter declined rapidly (Fig. 2b for the first pulse). Significant effects were the same at the second pulse but the overall δ13C signal in the second pulse was much higher (c. +1900‰) and the difference between heating treatments in the mycorrhizal compartments was smaller. Control treatments in both labelling periods had a δ13C of up to −10.0‰.
In both labelling periods, mean ERM respiration per compartment was c. 10.0 µg h−1 in ambient treatments but covaried with PAR received before each gas sampling (Fig. 3). Soil warming had no significant effect on respiration in either pulse (Fig. 3). Mean respiration per unit length of extra-radical hyphae (LERM) (data not shown) in the ambient treatments was 2.4 ng C m−1 h−1; in the soil warming treatment it increased slightly after the first pulse to 3.8 ng C m−1 h−1 but after the second pulse decreased to 1.5 ng C m−1 h−1 (F1,7 = 19.99, P = 0.003). Respiratory 13C loss per compartment was less than 1.5 µg h−1 and peaked in both pulses within 10 h after the start of labelling (Fig. 4); values were higher and declined more rapidly after the second pulse. Soil warming significantly increased respiration at the second sampling after the first pulse (F1,4 = 7.76, P = 0.049) (Fig. 4a) but there was no warming effect after the second pulse. The percentage of 13C relative to total carbon respired was similar in the two pulses in the ambient treatments and declined from 10% initially to less than 3% (Fig. 5). However, the 13C:12C ratio was always (but not significantly) higher in the soil warming treatment after the second pulse (Fig. 5b), and a repeated measures analysis of the proportion of the respiration ratios of 13C:12C of the second normalized by those of the first pulse label per temperature treatment (combination of Fig. 5a and b) showed a weakly significant temperature effect (F1,4 = 4.73, P = 0.095).
Impacts on fungal growth
Soil moisture content did not differ amongst pulse labels or treatments (data not shown). In nonmycorrhizal treatments there was no colonization of roots and LERM was < 0.04 m g−1. In mycorrhizal treatments, LERM increased between the two pulses but when the hyphal compartment was heated the increase in LERM was greater, and was 75% higher in heated than unheated compartments at the second sampling (Table 1). In contrast, total (LRC) and arbuscular (LRCarb) root colonizations did not change between the two pulses and were unaffected by soil warming (Table 1).
Table 1. Effects of soil warming on (a) length of colonized roots (LRC) and extra-radical mycelium (ERM) length density and (b) plant growth for the first and second labelling periods
(a) Significant F-ratios in the repeated measures design for the mean values of LRC, arbuscular LRC (LRCarb) and ERM length density (m per g dry weight of soil) in labelled mycorrhizal treatments for ambient and heated treatments are presented for the first and second pulse labels; neither nonmycorrhizal nor unlabelled treatment means differed significantly from these means. (b) Significant F-ratios for the two-way analysis of variance (ANOVA) for shoot dry weight (WS), specific leaf area (SLA), root dry weight (WR) and specific root length (SRL) at the final harvest (second pulse only) are presented for labelled mycorrhizal and nonmycorrhizal treatments, either ambient or heated, respectively; mean values of unlabelled plants did not differ from these means. There were no significant heating or interaction effects on plant growth in the second pulse.
ERM length density increased during the 2-wk period by 0.60 and 2.77 m g−1 for ambient and heated treatments, respectively (Table 1). These values equate to a minimum hyphal biomass production in the entire box of 0.19 and 0.54 mg C d−1, assuming constant ERM growth and no death in either treatment during the 2-wk period. For the ambient treatment in the first pulse and for both temperature treatments in the second pulse, mean ERM respiration was c. 10.0 µg C h−1 per compartment (Fig. 3) providing an additional carbon demand for the entire box of 0.48 mg C d−1. For the heated treatment in the first pulse, the figures were 13.6 µg C h−1 per compartment and 0.57 mg C d−1 per box. Therefore, total ERM carbon demand was constant at 0.67 mg C d−1 for the ambient treatment in both labelling periods but decreased in heated treatments from 1.11 at the first to 1.02 mg C d−1 at the second pulse. Total carbon demand by the ERM corresponded to less than 1% of net photosynthesis (126 and 181 mg C d−1 for the first and second pulses, respectively), assuming constant mean net photosynthesis of mycorrhizal plants of 6.6 and 8.4 µmol m−2 s−1 over a 16-h light period, respectively (data not shown).
Plant 13C:12C analyses
The mean natural abundance of 13C was 4.6 mg 13C g−1 in shoot and root material of all control plants. Four days after the first pulse, the 13C content of shoots was enriched by an average of 2.8 mg 13C g−1 (data not shown) but there were no differences among mycorrhizal or heating treatments. At the second harvest, after the second pulse, labelled plants contained significantly more 13C than unlabelled controls in both shoots (30.02 mg vs 11.7 mg; F1,14 = 7.00, P = 0.019) and roots (18.9 mg vs 10.2 mg; F1,14 = 11.73, P = 0.004; Fig. 6b); further, 13C biomass content in shoots (12.1 mg) and roots (10.0 mg) was unaffected by temperature in unlabelled plants (Fig. 6b). At the second pulse, labelled mycorrhizal and nonmycorrhizal plants had similar 13C shoot contents both before (30.3 mg) and after (27.2 mg) the gas sampling period. Consequently, in all treatments, the shoot 13C content declined similarly, by an average of 3.1 mg over 2 d (Fig. 6a). However, in labelled plants (Fig. 6b), mycorrhizal roots had 35% less 13C content than the roots of nonmycorrhizal plants. Further, mycorrhizal roots had the lowest 13C (14 mg) under ambient temperature, although there was no significant interaction with soil warming (P = 0.176).
Impacts on plant growth and photosynthesis
Plant growth was unaffected by warming of the hyphal compartment. Mycorrhizal and nonmycorrhizal plants had similar shoot and root dry weights (Table 1, Fig. 7) and shoot RGR (0.059 d−1) but mycorrhizal plants had significantly greater leaf area and SLA.
The mean plant CO2 uptake during the 3.5-h period in the first pulse was higher for outside air controls (1.63 ± 0.10 mmol CO2) than in labelled treatments (1.20 ± 0.07 mmol CO2) (F1,16 = 14.13, P = 0.002). During the second pulse, mycorrhizal plants had a 20% higher mean CO2 uptake (1.85 ± 0.06 mmol CO2) in 3.5 h than nonmycorrhizal plants (F1,14 = 6.34, P = 0.025). However, the mean value of net photosynthesis, 6.1 and 8.0 µmol m−2 s−1 for the first and second pulses, respectively, did not differ between pulses or treatments.
We report here the first quantitative estimate of the respiration of the ERM of an AM fungus and its response to soil warming, excluding background respiration. However, there is also surprisingly little information on ectomycorrhizal respiration. Ettema et al. (1999) and Hedlund & Augustsson (1995) reported basidiomycete (which are generally thicker as they tend to form rhizomorphs) hyphal respiration of 40 µg C g−1 d−1 and 0.055 µg C m−1 h−1, respectively, in comparison with which our findings of 0.2 µg C g−1 d−1 and 0.003 µg C m−1 h−1 seem to be very low. However, their results were obtained from field soil with the addition of glucose and antibiotics or in axenic culture, respectively.
We used gas sample tubes, which enabled us to calculate fungal compartment ERM respiration, and demonstrated an initial increase in ERM respiration under soil warming (Figs 3a and 4a), which disappeared after 2 wk of warming. This response resembles the acclimation to temperature shown by roots (Atkin et al., 2000). However, in a warmer environment, longevity of hyphae might decline as reported for roots (Fitter, 1996; Norby & Jackson, 2000), leading to accumulation of dead hyphae, which were not measured in this study; the significantly lower respiration rate per unit hyphal length (LERM) after 2 wk of soil warming is consistent with the accumulation of dead hyphae, whereas the increased ratio of 13C to total carbon respired (Fig. 5b) is not, as it indicates more rapid carbon allocation to the ERM. Both observations could be explained by an increase in younger, and therefore thinner but also more active, mycelium under soil warming; such an increase is likely to have occurred. Nonetheless, there was a residue of 13C label in the ERM by the time the second label was applied. Although values in 13C-labelled treatments had returned to near starting values (c. +13 and +30‰ for nonmycorrhizal and mycorrhizal treatments, respectively), the enrichment in the second pulse was very high (c. +1900‰), and consequently any residue effect was negligible.
We also determined ERM carbon costs for the host plant, which were very similar for the two pulse periods and were < 1% of net photosynthesis, similar to the value of 0.8% given by Jakobsen & Rosendahl (1990). However, we were only able to estimate the ERM in the plant compartment and did not account for root internal colonization with dense colonization of thicker hyphae and storage vesicles; total fungal carbon costs might then increase to over 5%, as reported by Snellgrove et al. (1982), Koch & Johnson (1984) and Cooper (1984). Further, we had to assume a similar hyphal mass for our G. mossea strain to that used by Harley & Smith (1983), which is somewhat uncertain. In our study, the carbon cost for ERM respiration was slightly higher than for growth, as estimated by Harris & Paul (1987). Further, although the ratio of 13C:12C respired (Fig. 5) was the same in ambient and heated treatments in the first pulse, it increased slightly under soil warming in the second pulse. This shows (i) that there was rapid carbon transport to the ERM in < 10 h, (ii) that only c. 10% of the carbon respired came from recently fixed carbon, and (iii) that there was faster carbon allocation inside the ERM under higher temperature, possibly as a result of increased production of young hyphae.
The finding that control plants did not show any 13C enrichment demonstrates that we successfully prevented contamination of unlabelled plants via 13C leakage from labelling chambers; roots of labelled plants increased 13C content by c. 18 mg (Fig. 6b), sufficient to follow a 13C signal into the hyphal compartment. Mycorrhizal plants had less labelled carbon than nonmycorrhizal plants in the roots, but not in the shoots. Although the difference in root 13C was only marginally significant (P = 0.089), it may demonstrate allocation of fixed carbon from mycorrhizal roots to the AM fungus (Jakobsen & Rosendahl, 1990) without affecting the shoot carbon pool. However, warming of the ERM did not result in a further decrease.
Diurnal differences in measured CO2 concentrations (Fig. 2a) reflected different ERM activities, corresponding to the carbon fixation of the plants: the concentration of root sucrose, which is believed to be one of the main forms of carbon transported to the fungus (Bago et al., 2003), increases during the day, consistent with peak activity of the AM fungus during and shortly after the photoperiod. Further, as ERM respiration clearly depended on short-term PAR effects, our findings could be explained both by a transfer mechanism based on leakage of carbon compounds into the apoplast followed by fungal capture, and by a transfer mechanism with greater regulation by the plant. There are two technical problems in measuring hyphal respiration as influenced by the plant compartment. First, mycorrhizal roots may have higher respiration rates than nonmycorrhizal roots (Solaiman & Saito, 1997); because we assumed that the 24% greater root length in the nonmycorrhizal plants (Fig. 7b) should have counteracted such an effect, we might have underestimated ERM respiration. Secondly, δ13C signals in the first pulse were highly variable, as is often the case (Figs 2b and 4a), which made it difficult to test statistically for differences. The variation might reflect either individual differences in the growth of the hyphal front or differences in ERM activity among replicates.
Plants showed typical mycorrhizal responses – increased leaf area, increased SLA and reduced root length (Harris & Paul, 1987; Gavito et al., 2001) – and mean net photosynthesis ranging between 4.5 and 8.8 µmol m−2 s−1 was very similar to data for this species obtained by Staddon et al. (1999a) under saturated light. Hyphal length densities were between 3 and 7 m g−1, within the range reported for glomalean fungi in other compartment experiments (Jakobsen et al., 1992; Hodge et al., 2001) but lower than values obtained in pot experiments or field data. The latter sometimes exceed 100 m g−1 (Sylvia, 1990), although much of this may be either dead (Sylvia, 1988) or nonmycorrhizal (Sylvia, 1986) material.
Photosynthesis is expected to increase under higher temperature and elevated CO2 (Ceulemans et al., 1999) if water supply is not limited. Plants will then need an increased nutrient supply to maintain higher growth rates (Olesniewicz & Thomas, 1999). The greater amount of ERM in the warmer environment found here will increase the effectiveness with which the host plant explores soil and if, as suggested by Staddon & Fitter (1998) for elevated CO2, more carbon were available to the fungus, the positive growth response of the ERM of AM fungi might well supply this increased nutrient demand in a warmer climate.
In this study, we have obtained evidence for faster carbon allocation to and increased respiration of the ERM under higher soil temperature, but also rapid acclimation and an immediate response to changes in available carbon from a previously mixed root carbon pool. The positive growth response of the ERM to higher temperature together with acclimation of its respiration might lead to a significant increase in carbon accumulation in soils when the climate becomes warmer. However, light seems to be the overall controlling factor in carbon allocation to the fungus. This study underlines the important role of the AM fungal symbionts in global carbon cycling and suggests that a more mycocentric view in ecological studies should be considered in future climate modelling.
The authors are very grateful to C. Abbott, who looked after the plants. We would also like to thank P. Wilson, who provided us with the AM inoculum. This work was partly funded by the NERC (UK), the Evangelische Studienwerk (Germany) and the University of York.