Growth depression of mycorrhizal Citrus seedlings grown at high phosphorus supply is mitigated by elevated CO2


Author for correspondence: J. H. Graham Tel: +1 863 956 1151 Fax: +1 863 956 4631 Email:


  •  Gas exchange and growth responses of pot-grown sour orange (Citrus aurantium) and sweet orange (C. sinensis) were studied at high soil-P where growth depression is predicted.
  •  Seedlings were either inoculated (AM) with the arbuscular mycorrhizal fungus Glomus intraradices or not inoculated (NM), and grown at elevated (eCO2) or ambient CO2 (aCO2) for 11 wk.
  •  At aCO2, growth of AM sour orange was depressed (18%) compared with NM seedlings, but at eCO2, AM sour orange plants were 15% larger than NM plants. Growth depression coincided with increased rhizosphere respiration, reduced root starch concentration, and lower relative growth rates. Net photosynthesis (Pn) of both genotypes was enhanced by eCO2. For sour orange, the stimulation was greater in AM than in NM seedlings and this may have compensated for the carbon (C) expenditure on mycorrhizas. Gas exchange and growth of sweet orange were unresponsive to colonization by G. intraradices.
  •  Differential responses to treatments suggest that C expenditure on mycorrhizas is more tightly regulated in citrus genotypes of low mycorrhizal dependency (MD) such as sweet orange than in high MD genotypes (e.g. sour orange).


In soils with low phosphorus (P) content, arbuscular mycorrhizal fungi (AMF) can greatly enhance P acquisition and improve plant growth (Graham, 2000), however, this symbiosis occurs at a cost to the autotroph. As obligate symbionts, AMF rely solely on their hosts for organic compounds. Estimates of the amounts of carbon (C) allocated to the fungus vary from 4 to 20% of the plant’s total carbon budget (Graham, 2000). The extent of root colonization and benefits from mycorrhizas are generally highest when soil P is limiting, however, substantial root colonization still occurs at P saturation both in the field and in soil-less potting media (Graham et al., 1991; Peng et al., 1993). Mycorrhizal colonization in nutrient-rich environments neither benefits plant P status nor increases photosynthesis (Syvertsen & Graham, 1999), but frequently induces growth depression in a wide range of plant species (Valentine et al., 2001; Buwalda & Goh, 1982; Peng et al., 1993; Graham & Eissenstat, 1994). The physiology and significance of AMF-induced growth depression in managed and natural ecosystems are not well understood. By acting as facultative parasites during colonization at high soil-P, AMF can potentially reduce yield in intensively managed agroecosystems (Graham & Eissenstat, 1998). In naturally fertile ecosystems, growth reduction in mycorrhizae-dependent species could alter fitness and competitive outcomes.

Growth depression is attributed to increased diversion of host C to the mycobiont and is accompanied by increased rhizosphere respiration and reduced pools of nonstructural carbohydrates in the host (Peng et al., 1993; Buwalda & Goh, 1982). For instance, P-sufficient, mycorrhizal Citrus plants expended 37% more C on soil/root respiration, which led to a 10–20% reduction in the specific C gain of AM compared to NM plants (Peng et al., 1993). The total nonstructural carbohydrate (TNC) content in host tissues, particularly root TNC, is a good indicator of the amount of C allocated to the AMF (Graham et al., 1996, 1997).

Stimulation of CO2 assimilation by elevated CO2 (eCO2) and the concomitant improvement in plant carbohydrate status could reduce the C cost of mycorrhizal colonization under high soil-P conditions. In a study conducted under nonlimiting conditions, the interaction of eCO2 and mycorrhizal colonization differentially altered the growth of two citrus genotypes with contrasting mycorrhizal dependency (MD; defined as the ratio of AM plant dry mass to NM plant dry mass at low P supply) (Syvertsen & Graham, 1999). Mycorrhizal colonization at eCO2 enhanced the growth of sour orange (the high MD genotype) but had no effect on growth of sweet orange (the low MD genotype). Although the basis for this differential growth response was presumed to be due to a mycorrhizal-mediated response of photosynthesis (Pn) at eCO2, this was not confirmed by the ambient CO2 (aCO2) measurements taken at 12 wk after root inoculation with the AMF Glomus intraradices. To better understand the effects of mycorrhizal colonization on host C economy, direct colonization effects must be distinguished from effects related to improved P nutrition. One approach is to grow AM and NM plants with saturating supply of P and other nutrients, thereby producing plants with similar growth rates and tissue P levels (Peng et al., 1993; Syvertsen & Graham, 1999). In this study, we have compared the response of citrus genotypes to air pCO2 and root colonization by G. intraradices under conditions of P saturation.

The first objective was to determine whether CO2 enrichment would compensate for potential growth depression at high-P supply in closely related citrus genotypes with contrasting MD. The second objective was to determine whether Pn of AM seedlings would be stimulated as a result of increased C sink strength in mycorrhizal roots even at high-P supply. We hypothesized that growth depression of AM seedlings at high-P will be greater in the high MD genotype due to the increased C cost of mycorrhizal colonization, and that enhanced C supply at eCO2 would alleviate this negative effect.

Materials and Methods

Plant culture

Seeds of sour orange (Citrus aurantium L.) and ‘Ridge Pineapple’ sweet orange (Citrus sinensis L. Osbeck), were germinated and grown in a commercial peat/perlite/pine bark medium (1 : 1 : 1 by volume; Metro-Mix 500; The Scott’s Co., Marysville, OH, USA) with starter nutrients for 30 d in a glasshouse. After the seedlings had developed two to three true leaves, roots of uniform seedlings of each genotype were gently cleaned of the germination medium and transplanted into 150-cm3 plastic containers (Stuewe and Sons, Inc., Corvalis, OR, USA) filled with autoclaved Candler fine sandy soil (pH 6.8) with 3.8 mg kg−1 of available P as determined by double-acid extraction (Mehlich, 1953). Mycorrhizal treatments (AM) were established by placing a pad (c. 0.2 g) of air-dried maize (Zea mays) roots and soil containing approx. 560 propagules of the AMF Glomus intraradices (INVAM FL208) in the middle of the soil volume in each container during transplanting. Nonmycorrhizal (NM) soil received an extract of inoculum that had been passed through a 38-µm sieve to establish the same microflora associated with maize roots. G. intraradices is widespread in native and orchard soils and is considered an aggressive colonizer (defined as the rate of root colonization in the absence of competing AMF; Graham et al., 1996) of citrus roots both at low and high soil-P. G. intraradices is also less sensitive to agricultural management practices such as fertilization and tillage compared to members of other genera such as Gigaspora (Miller & Jastrow, 1992; Graham & Abbott, 2000).

After transplanting and inoculation, seedlings were placed in either of two identical unshaded, temperature-regulated glasshouses located at the University of Florida’s Citrus Research and Education Center (Lake Alfred, FL, 28.09° N, 81.73° W; elevation 51 m). The eCO2 glasshouse was maintained at a CO2 partial pressure of 70 Pa 24 h d−1, while the aCO2 glasshouse was maintained at 36 Pa. Within each glasshouse (CO2 level), there were eight replicate plants for each genotype × mycorrhizal combination and these were randomized every second week. The CO2 treatment was switched between glasshouses twice at 7 d, and at 14 d after transplanting, however, this procedure was discontinued to prevent potential cross-contamination of NM plants and because after each switch, it took several hours for pCO2 to come to a stable set level. However, after equilibration, growth conditions were identical between glasshouses except for pCO2. Maximum photosynthetic photon flux (PPF) at the canopy level was about 1500 µmol m−2 s−1 using natural summer photoperiods. Average integrated daily PPF during the study period (June to August) was approx. 28 mol m−2. Average day/night temperatures were 32/22°C and air relative humidity varied from 60 to 100%. Seedlings were well-fertilized daily using standard Hoagland’s solution (Hoagland & Arnon, 1939) containing 2 mM P (2 mM KH2PO4, 5 mM KNO3, 5 mM Ca(NO3)2, 2 mM MgSO4, and micronutrients) and watered to excess every other day with tap water to prevent buildup of nutrient salts. Seedlings were grown for 11 wk (15 June to 31 August) in 1999 and the experiment was repeated during 14 wk (25 May to 30 August) in 2000.

Gas exchange measurements

Whole-plant gas exchange was measured at 4, 6, 8, and 11 wk after transplanting on eight replicate seedlings per treatment. Net photosynthetic (Pn) and dark respiration (Rd) rates of the entire shoot were measured with a portable photosynthesis system (LI-6200; Li-Cor, Inc., Lincoln, NE, USA) operated in closed-mode. During each measurement, the shoot was sealed in a 4-l teflon-coated polycarbonate chamber (16 × 18 × 14 cm, LI-6000–10, Li-Cor). At 8 and 11 wk, when seedlings became too large for the 4-l chamber, a 10-l clear cylindrical acrylic chamber (21 cm diameter, 29 cm tall) similar to the design of Eissenstat et al. (1993) was used. Three fans, clamped to the chamber walls, were used to continuously stir the air inside the chamber. This chamber was not temperature-controlled, but it was allowed to equilibrate with CO2, temperature, and vapor pressure conditions in the glasshouse between measurements. During all measurements of Pn, PPF was > 800 µmol m−2 s−1, and air temperature inside the chamber was 29 ± 4°C. Shoot respiration was measured in the same fashion as Pn except that plants were dark-adapted in their growth environment for 2 h before measurements and the gas exchange chamber was darkened with aluminium foil. The entire shoot of each seedling was enclosed within the chamber and allowed to reach a steady rate of CO2 depletion (Pn) or efflux (Rd) for 30–45 s before the first of four observations was taken. Gas exchange measurements were made at pCO2 of 36 ± 5 Pa for aCO2-grown plants and at 70 ± 6 Pa for eCO2-grown plants. At 11 wk, reciprocal transfer Pn measurements were conducted to determine if treatments had altered the Pn capacity of plants. CO2 treatments were switched by transferring plants from the aCO2 glasshouse to the eCO2 glasshouse and vise versa. Pn was then measured after plants had equilibrated for at least 1.5–2 h in the new pCO2. By comparing Pn measured at the same pCO2, the confounding effects of differing treatment pCO2 are removed (Gunderson & Wullschleger, 1994; Sage, 1994). These comparisons (especially those made at low pCO2) are similar to the CO2 assimilation vs intercellular pCO2 (A/Ci) analysis in that Pn is compared at the same Ci, and from the comparisons, inferences about CO2 and AM effects on Pn capacity can be made although no mechanistic interpretations are offered.

Combined root and soil (rhizosphere) respiration was measured with a portable soil respiration chamber (LI-6000–09; Li-Cor), connected to the LI-6200 console, and operated in the closed-mode. A cylindrical PVC collar (12 cm diameter, 16 cm in length) was attached to the LI-6000–09 chamber body (in an upside-down position) using coil springs and an airtight foam gasket. The entire root system in its plastic container was then inserted into the PVC collar and sealed around the base of the stem with two semicircular plexiglass plates lined with foam gaskets to make an airtight seal. Air circulating in the space between the root container and the collar was thoroughly mixed by a fan attached to the inner wall. Typically, a root system was sealed in the root chamber and the CO2 concentration in the closed system drawn down about 5 Pa CO2 below the ambient level by scrubbing the flow with soda lime for 10 s. The system was then allowed to attain a steady rate of CO2 efflux (20–30 s) before recording CO2 efflux rates. Four observations per plant were then taken to span the growth CO2 level and averaged to calculate a specific rhizosphere respiration rate.

Root colonization and plant nutrient status

Following gas exchange measurements, seedlings were harvested and separated into leaves, stems, taproot (> 2 mm), and fibrous roots (< 2 mm). Total leaf area per plant was measured with a portable leaf area meter (LI-3000, Li-Cor) and roots were gently washed free of soil with tap water. All plant components were dried at 70°C for 48 h and weighed. Leaf P and K were determined on 100 mg subsamples of finely ground tissue using an inductively coupled plasma atomic emission spectrometer (ICP-AES), after the tissue had been ashed (500°C, 5 h) and resuspended in 1 mM HCl. Leaf N was determined on another sample of leaf tissue using an NA1500 C-N analyser (Fisson Institute Inc., Dearborn, MI, USA).

The incidence of intraradical (IR) colonization by G. intraradices was determined as the percentage of the fibrous root length segments that contained vesicles, arbuscules or hyphae (Graham et al., 1991). Twenty, 1-cm fibrous root segments were subsampled from each seedling, cleared in KOH followed by NaOCl, stained with trypan blue, and then individually mounted on microscope slides for examination at 250×.

Total nonstructural carbohydrates in leaves and fibrous roots were determined enzymatically using 30 mg subsamples of dried, ground tissue. Samples were boiled in water for 2 min and centrifuged at 800 g for 2 min. The supernatant was analysed for soluble sugars and the pellet was analysed for starch after a 48-h digestion with amyloglucosidase (Haissig & Dickson, 1979). Glucose standards were used throughout the analyses.

Treatment effects on plant growth and biomass allocation were examined by calculating relative growth rates (RGR, the slope of the regression of loge (plant dry mass) and time), root fraction (RF; root dry mass : total plant dry mass), and root-to-shoot ratios (root dry mass : shoot dry mass). Allometric relationships between roots and shoots were calculated to further separate potential ontogenetic effects from size effects as plants grew (Farrar & Gunn, 1998). The amount of biomass allocated to roots was described by the allometric equation:

inline image

(a and k, constants.) The slope, k, of this relationship is the allometric coefficient such that when k > 1, root allocation rises with ontogeny and if k < 1 root allocation declines. For plants at steady-state nutrition, the RGR of roots, shoots, and whole plant are similar and, thus, k= 1 (Ingestad & Ågren, 1991; Farrar & Gunn, 1998).

Data analyses

The experimental design was a split-plot in a 2 × 2 × 2 factorial arrangement with two CO2 levels as the main plot and with two genotype and two mycorrhizal treatments as subplots. This study was first conducted in 1999 (15 June to 31 August) and repeated in 2000 (25 May to 30 August) in order to increase the scope of inference of the results since there were only two glasshouses. Lower growth temperatures in 2000 delayed overall plant development by 3 wk after transplanting. Although seedling responses were identical to those in 1999, the protracted experiment in 2000 complicated time course evaluations. Thus, most of the time-course data reported here are from the 1999 experiment. Where appropriate, data were analysed using the general linear model (GLM) procedures of the Statistical Analysis System (SAS Institute, Cary, NC, USA) and treatment means were compared by the Protected Least Significant Difference (LSD) method.


Plant growth

Leaf area of both genotypes was significantly increased by eCO2 at all harvest dates (range 7–42% at final harvest; Table 1). Leaf area of sour orange was greater than that of sweet orange. Mycorrhizal colonization did not affect leaf area of either genotype, however, for sour orange, leaf area was increased slightly by mycorrhizas at eCO2 but reduced at aCO2 relative to NM seedlings. Specific leaf area (SLA; leaf area per unit leaf mass) was consistently reduced by growth at eCO2. Sour orange SLA was increased by mycorrhizas only at aCO2 while sweet orange SLA was increased by mycorrhizas at eCO2 (P < 0.01 for the CO2 × genotype × AMF interaction; Table 1).

Table 1.  Leaf area (LA), specific leaf area (SLA), total plant biomass (TBM) and root biomass fraction (RF) of sour orange (SO) and sweet orange (SwO) seedlings grown at ambient (35 Pa) or elevated pCO2 (70 Pa) with high phosphorus supply and inoculated (AM) or not inoculated (NM) with the arbuscular mycorrhizal fungus (AMF) Glomus intraradices. Means (n = 8) followed by the same letters within a column are not significantly different at P < 0.05 (LSD)
GenotypepCO2 PaAMFLA cm2SLA cm2 g−1TBM gRF g g−1
SO35NM201.5bcd110.9b 3.9c0.32a
  AM192.5cd121.1a 3.2d0.31ab
 70NM229.6b 97.4cd 5.2b0.29b
  AM273.7a 96.0cd 6.0a0.30ab
SwO35NM177.9d110.1b 3.4cd0.31ab
  AM182.1d106.5b 3.5cd0.30ab
 70NM213.7bc 92.9d 5.2ab0.29ab
  AM214.4bc 98.8c 4.8b0.29b
 dfF Values   
  1. * , **Probability value ≤ 0.05, 0.01, respectively.

CO2 (C)1 28.46**143.71*85.37*5.55*
Genotype (G)1 10.26**  4.96* 1.580.82
AMF (M)1  1.07  3.31 0.390.89
C×G1  0.75  8.01* 0.520.07
C×M1  2.32  0.04 3.362.37
G×M1  0.73  0.57 0.181.00
C×G×M1  3.06 14.09* 7.52**0.23

Total plant dry mass was increased by eCO2 at all harvests dates for both genotypes (Table 1 for final harvest). At 11 wk, total dry mass of AM sour orange was 15% greater than that of NM seedlings at eCO2; however, at aCO2 the dry mass of AM sour orange seedlings was 18% less than that of NM seedlings (P < 0.001 for AMF × CO2 × genotype interaction; Fig. 1a; also Table 1). The dry mass of sweet orange seedlings was significantly increased by growth in eCO2 but unaffected by mycorrhizal colonization.

Figure 1.

The total dry mass, plotted on a loge scale, of mycorrhizal (AM) and nonmycorrhizal (NM) (a) sour orange (Citrus aurantium) and (b) sweet orange (Citrus sinensis) seedlings grown at ambient (35 Pa) or elevated pCO2 (70 Pa) with high phosphorus supply. 35 NM, open circles; 35 AM, closed circles; 70 NM, open squares; 70 AM closed squares. The slope of the line for each treatment (relative growth rates (RGR)) was calculated by least-square regression analyses.

Between 29% and 32% of total plant biomass was allocated to roots (Table 1). Root fraction (RF) of eCO2-grown plants tended to be lower than that of aCO2-grown plants. Root : shoot allometric coefficients (k) were not significantly different from one (0.89–1.01, Table 2; Fig. 2). For sour orange grown in eCO2, k for AM seedlings was greater than k for NM seedlings.

Table 2.  The effects of growth pCO2 (35 or 70 Pa) on the root : shoot biomass allometric coefficient, k, and whole plant relative growth rate (RGR) of sour orange (SO) and sweet orange (SwO) seedlings grown at high soil-P and inoculated (AM) or not inoculated (NM) with the arbuscular mycorrhizal fungus (AMF) Glomus intraradices. Values followed by the same letters within a column are not significantly different at P < 0.05
GenotypepCO2 PaAMFkRGR
Week 4–6Week 8–11
Figure 2.

The allometric relationships between loge (root dry mass) and loge (shoot dry mass) of mycorrhizal (AM) and nonmycorrhizal (NM) (a) sour orange (Citrus aurantium) and (b) sweet orange (Citrus sinensis) seedlings grown at ambient (35 Pa) or elevated pCO2 (70 Pa) with high phosphorus supply. 35 NM, open circles; 35 AM, closed circles; 70 NM, open squares; 70 AM closed squares.

Whole plant relative growth rates (RGR; slope of lines in the loge plots, Fig. 1) varied among treatments and with time (Table 2). For sour orange at both CO2 concentrations, RGR was highest between 4 and 6 wk (range 0.25–0.39 g g−1 wk−1) and declined thereafter. Between 8 and 11 wk, the RGR of AM sour orange at aCO2 was approx. 30% lower than that of NM sour orange at aCO2. At eCO2, however, the average RGR of AM seedlings was 30% higher than that of NM seedlings. The RGR of sweet orange was largely unresponsive to CO2 or mycorrhizal treatments.

Mycorrhizal colonization and mineral nutrition

Between 9 and 12% of root lengths were colonized by mycorrhizas 4 wk after seedlings were transplanted and inoculated (Table 3). Roots of nonmycorrhizal (NM) seedlings remained free of mycorrhizal colonization throughout the study. Root colonization was not significantly altered by the CO2 concentration during growth but the incidence of root colonization was relatively constant throughout the growth period.

Table 3.  Effect of pCO2 (35 or 70 Pa) on the percentage of root length colonized by the arbuscular mycorrhizal fungus Glomus intraradices in sour orange (SO) and sweet orange (SwO) seedlings grown at high phosphorus supply (2 mM) for 11 wk. Each point is a mean of 8 replicate plants. Within a time point, means followed by the same letters are not significantly different at P < 0.05 (LSD)
GenotypepCO24 wk6 wk8 wk11 wk
SO35 9.1 ± 4.1a 9.0 ± 4.5a 9.0 ± 4.5a11.5 ± 3.4a
 70 9.2 ± 3.5a11.1 ± 4.5a 7.9 ± 3.9a 9.2 ± 3.5a
SwO3512.3 ± 4.7a15.9 ± 5.6a14.1 ± 5.4a17.7 ± 4.3a
 7012.5 ± 3.8a12.2 ± 4.3a12.5 ± 3.8a12.5 ± 3.8a

Mycorrhizal colonization at high P had no effect on leaf P concentration (Table 4). Leaf P, N and K concentrations were significantly reduced by growth at eCO2 especially in sour orange. Sweet orange had higher leaf nutrient concentrations than sour orange (P < 0.01 for genotype effect). Leaf P, N, and K concentrations were stable throughout the 11-wk study period (data not shown), indicating that a condition of steady-state plant nutrition was sustained.

Table 4.  Leaf nitrogen (N), phosphorus (P), potassium (K), starch, and fibrous root starch at wk 11 of sour orange (SO) and sweet orange (SwO) seedlings grown at ambient (35 Pa) or elevated pCO2 (70 Pa) with high phosphorus supply and inoculated (AM) or not inoculated (NM) with the arbuscular mycorrhizal fungus (AMF) Glomus intraradices. Means (n = 8) followed by the same letters within a column are not significantly different at P < 0.05 (LSD)
GenotypepCO2AMFNPK (mg g−1)Leaf starchFibrous root starch
SO35NM 29.1c  2.0c13.5bc 7.9bc11.1ab
  AM 30.8b  2.1c14.9ab 4.3cd 6.9d
 70NM 21.9d  1.6d10.5de20.9a13.0a
  AM 23.3d  1.5d10.2e21.4a13.1a
SwO35NM 33.4a  2.8a15.7a 2.7d 6.2d
  AM 35.1a  2.8a15.9a 3.1d 7.85cd
 70NM 28.2c  2.2bc12.5c11.5b10.3b
  AM 31.1b  2.3b12.1cd10.8b 9.6bc
 dfF Values    
  1. * , **Probability value ≤ 0.05, 0.01, respectively.

CO2 (C)1136.5**112.10**75.82**63.29**42.03**
Genotype (G)1120.1**208.27**17.24**17.07**21.23**
AMF (M)1 13.3**  0.09 0.22 1.18 1.79
C×M1  9.9**  0.13 0.52 0.99 0.52
C×M1  0.52  0.34 1.91 0.04 0.91
G×M1  0.40  0.84 0.60 0.75 6.15*
C×G x M1  0.55  1.77 0.45 0.62 9.93**

Growth in eCO2 significantly increased starch content in leaves and fibrous roots of both genotypes (P < 0.01; Table 4). Leaf and fibrous root starch were higher in sour than in sweet orange. Fibrous root starch of AM sour orange grown in aCO2 was significantly lower (37%) than that of NM sour orange at week 8 and 11 (P < 0.01 for CO2 × genotype × AMF interaction).

Gas exchange characteristics

Net shoot photosynthetic rates (Pn) of plants grown in eCO2 were significantly higher than that of plants grown at aCO2 at all measurement dates (P < 0.045; Fig. 3a,b). There were significant two–way interactions between CO2 and mycorrhiza on Pn of sour orange at 6, 8, and 11 wk (P < 0.05; Fig. 3a). At eCO2, shoot Pn rates of AM sour orange were significantly higher than those of NM seedlings. At aCO2, however, mycorrhizal colonization had no effect on shoot Pn. Shoot Pn of sweet orange was largely unaffected by mycorrhizal colonization, except at week 8 where mycorrhizal colonization increased Pn at eCO2, but reduced it at aCO2 (Fig. 3b).

Figure 3.

Whole plant CO2 uptake (Pn, a, b), specific shoot respiration (SSR) (c, d) and specific rhizosphere respiration (SRR) (e, f) of sour orange (Citrus aurantium) (a, c, d) and sweet orange (Citrus sinensis) (b, d, f) seedlings inoculated (AM) or not inoculated (NM) with the arbuscular mycorrhizal fungus Glomus intraradices, and grown at ambient (35 Pa) or elevated pCO2 (70 Pa) with high phosphorus supply for 11 wk. 35 NM, open circles; 35 AM, closed circles; 70 NM, open squares; 70 AM closed squares.

For both genotypes, the Pn of plants grown in eCO2 and measured at aCO2 (Fig. 4) were lower than the corresponding Pn of plants grown and measured at aCO2 (Fig. 3a,b), indicating down-regulation of Pn following long-term growth in eCO2. For sour orange grown in eCO2, the reciprocal Pn (measured at aCO2) of AM plants was higher than that of NM plants (Fig. 4a). Within a measurement pCO2, there were no differences in stomatal conductance and intercellular pCO2 between NM and AM plants (data not shown).

Figure 4.

Light-saturated rates of photosynthesis of sour orange (Citrus aurantium) (a) and sweet orange (Citrus sinensis) (b) seedlings grown for 11 wk at ambient (Amb.) or elevated (Elev.) pCO2 and measured after transferring from one pCO2 environment to the other. Seedlings were either inoculated (AM, open columns) or not inoculated (NM, closed columns) with the arbuscular mycorrhizal fungus Glomus intraradices. Each bar is a mean ± SE of 8 plants. Astertisk, statistically significant at the 95% probability level.

At 6 and 8 wk, shoot respiration rates (Sr) were higher in plants grown at aCO2 than in eCO2-grown plants. Between 6 and 11 wk, Sr of NM sour orange seedlings grown at eCO2 was higher than that of AM seedlings (Fig. 3c). For sour orange grown at aCO2, and sweet orange grown at both CO2 levels, there were no consistent mycorrhizal effects on Sr.

At aCO2, rhizosphere (root + soil) respiration rates (Rr) of AM sour orange seedlings were higher than Rr of NM seedlings between 6 and 11 wk but this was significant only at week 8 (Fig. 3e). Between week 6 and 11, Rr of NM, sour orange seedlings grown in eCO2 were consistently higher than that of AM sour orange seedlings. The Rr of sweet orange was not affected by mycorrhizas, but the Rr of both genotypes was generally lower in aCO2 than in eCO2.


Growth at high-P supply (2 mM) produced AM and NM citrus plants with similar foliar-P status and enabled an evaluation of mycorrhizal effects on the autotroph under conditions of P sufficiency. Significant CO2, AMF and genotype effects on Pn and growth confirmed differences in host MD. The positive growth responses to AMF at eCO2 contrasted with growth depression at aCO2 for sour orange while sweet orange remained unresponsive to root colonization by G. intraradices. The differential responses of Pn in these genotypes provide the basis for their growth responses.

Sour orange is more mycorrhizal dependent and its growth was more responsive to colonization than that of sweet orange, a relatively low MD citrus genotype (Graham et al., 1991). Based on the relative MDs of different species and closely related genotypes, Graham et al. (1991) predicted that more host C would be partitioned to and utilized by the mycorrhizal root system of high MD genotypes compared to low MD genotypes. The pattern of nonstructural carbohydrate depletion in AM sour orange and sweet orange in this study is consistent with that prediction. At aCO2, leaf and root starch were reduced by AMF colonization in sour orange but not in sweet orange, and the Pn of neither genotype responded positively to colonization by G. intraradices as previously reported for other citrus genotypes (Peng et al., 1993; Syvertsen & Graham, 1999). Growth depression of AM sour orange occurred between 8 and 11 wk when there were no treatment effects on biomass allocation between roots and shoots. Onset of growth depression coincided with increased rhizosphere respiration rates, leaf and fibrous root starch depletion and reduced RGR of AM, compared to NM plants at aCO2. The observed depression can, therefore, be attributed to increased below-ground C expenditure coupled with a reduction in whole plant RGR between 8 and 11 wk rather than to changes in biomass allocation. At eCO2, the differences in starch and growth depression disappeared and positive growth responses appeared because Pn apparently more than compensated for the increased root C demand.

In a previous study, colonization by G. intraradices at high P and eCO2 significantly increased growth of sour orange (high MD genotype) but not of sweet orange (low MD) (Syvertsen & Graham, 1999). Growth depression was not observed in that study but fibrous root starch was lowered by mycorrhizal colonization, particularly in sour orange at low P and eCO2, conditions under which Pn was most stimulated. This suggested that increased C expenditure on AM root colonization has the potential to deplete host plant’s C reserves and result in growth depression. These relationships were confirmed in the present study by the differential responses of these genotypes to CO2 and root colonization by G. intraradices.

Our finding that sweet orange is less responsive than sour orange to colonization at eCO2 is significant from the standpoint of MD and supports the earlier prediction that genotypes of low MD regulate their C more tightly than genotypes of high MD (Graham & Eissenstat, 1994). Sweet orange had a much higher leaf-P concentration which is indicative of its lower MD and lower rates of mycorrhizal colonization in the field compared to sour orange (Graham & Syvertsen, 1985; Graham et al., 1991). Root colonization at P saturation, where there is no benefit from symbiosis, varies substantially among closely related citrus genotypes, with low MD genotypes having lower rates of colonization than high MD genotypes. The rate rather than the extent of colonization seems to be limiting in low MD genotypes (Graham et al., 1991). By restricting the rate of C expenditure on the mycobiont during conditions of little or no benefit from symbiosis, the low MD genotypes could, presumably, allocate more C to competing requirements that are more likely to yield immediate returns on the investment such as leaf area development for carbon assimilation (Graham & Eissenstat, 1994). This reduces the potential for mycorrhizal-induced growth depression. In the present study, there were no significant effects of CO2 or genotype on the extent of root length colonized by G. intraradices, and any effects on the rate of colonization may have occurred prior to our first assessment at week 4. Although total root colonization was low, AMF colonization levels were similar to those observed in high-fertility orchard and nursery soils (c. 17% for sour orange; Graham et al., 1991; Graham & Eissenstat, 1998) and did not preclude mycorrhizal treatment effects on host physiology and growth.

Mycorrhizal colonization and growth at eCO2 had a more than additive effect on Pn and biomass accumulation of sour orange (Figs 1a, 3a). This result contrasts with that of Staddon et al. (1999) where increased Pn as a result of mycorrhizal colonization and eCO2 did not translate into increased plant biomass of Plantago lanceolata. These differences highlight the importance of considering other traits such as mycorrhizal dependency when evaluating the impact of root colonization and eCO2 on the C economy of host plants. Many herbaceous plants are, by and large, adapted to fertile environments and are less dependent on mycorrhizas than many woody plants, and this may explain why their RGR is less responsive to mycorrhizal colonization at eCO2. Reich et al. (2001) recently found substantial variation in the biomass response to CO2 and N availability among grassland species within four trait-based functional groups and concluded that functional classifications alone may not be sufficient in understanding plant responses to CO2 and N availability. Interspecific variation in Citrus response to mycorrhizal colonization at high-P and eCO2 in the present and previous studies (Syvertsen & Graham, 1999) indicate that mycorrhizal dependency is a functional trait of potential predictive value for understanding responses to CO2.

Shoot Pn was stimulated by AMF colonization only in sour orange grown at eCO2. Increased Pn by AMF colonization at aCO2 has previously been reported in sour orange (Eissenstat et al., 1993), clover (Wright et al., 1998a,b), faba bean (Kucey & Paul, 1982), leek (Snellgrove et al., 1982), cucumber (Black et al., 2000; Valentine et al., 2001), soybean (Harris et al., 1985; Brown & Bethlenfalvay, 1988), and tropical tree seedlings (Lovelock et al., 1997), especially when the P status of AM and NM plants differed. Where soil P supply is limiting, enhanced Pn of AM, relative to NM plants, is probably due to a mycorrhizal-dependent improved P nutrition (Black et al., 2000). Where AM and NM plants have similar tissue P status, Pn enhancement in AM plants has been attributed to increased C sink strength from the fungal symbiont upon colonizing roots of the autotroph (Brown & Bethlenfalvay, 1988; Eissenstat et al., 1993; Wright et al., 1998a,b; Valentine et al., 2001). In the present study, shoot Pn rates, RGR, and biomass of AM sour orange seedlings grown in eCO2 with optimum P nutrition were consistently higher than those of their NM counterparts. This coincided with overall lower specific shoot and root respiration rates of AM than NM plants. Since AM and NM plants had similar leaf P concentrations, Pn stimulation at eCO2 could be attributed more clearly to nonnutritional factors such as increased fungal C sink strength. Wright et al. (1998a,b) attributed increased Pn of AM Trifolium repens plants (compared to NM plants) to increased fungal sink strength for host carbohydrates following root colonization. This enhanced sink activity was correlated with increased activities of sugar metabolizing enzymes (sucrose synthase and invertase) in roots (Wright et al., 1998a). Carbohydrate abundance in eCO2 is also expected to enhance gene expression for enzymes involved in C storage and utilization (Koch, 1996) and this sink effect may account for higher Pn of AM compared to NM plants. Absence of an AMF effect on sour orange Pn at aCO2, compared to eCO2, may reflect differences in the metabolic activity of AMF, with a higher activity at eCO2 than at aCO2.

It is not unusual for Pn capacity to be down-regulated after long-term growth in eCO2 as resources are reallocated from the Pn apparatus to nonphotosynthetic processes (Bowes, 1993). Growing Citrus seedlings in eCO2 for 11 wk resulted in down-regulation of Pn capacity, as determined from the reciprocal transfer measurements in the present study (Figs 3, 4). For sour orange, again down-regulation of Pn was greater in NM than in AM plants, supporting the notion that mycorrhizas provided a strong sink for host carbohydrates and this may have lessened the magnitude of down-regulation of Pn at eCO2.

The effects of eCO2 on AMF colonization of woody plant roots are poorly documented and inconsistent compared to effects on ectomycorrhizal colonization (Lewis & Strain, 1996; Fitter et al., 2000; Treseder & Allen, 2000). Increased AMF colonization in eCO2 has been reported in Populus tremuloides (Klironomos et al., 1997) and Beilschmiedia pendula (Lovelock et al., 1996). In the present study, growth pCO2 did not alter the percentage of root length colonized by G. intraradices. However, total fine root mass and probably length (not measured) were increased by eCO2, resulting in more root tissue per plant available for colonization as has been reported earlier for Pascopyrum smithii and Bouteloua gracilis (Morgan et al., 1994; Monz et al., 1994). Growth pCO2 also did not alter the extend of AMF colonization in Populus × euramericana saplings (Lussenhop et al., 1998) or in P. lanceolata and Trifolium repens (Staddon et al., 1998), but the total colonized root length was greater in eCO2 than in aCO2. Mycorrhizal establishment appears to be regulated by root exudates, surface-bound recognition molecules, and soil/plant nutrient status (Koske & Gemma, 1992). The overall low levels of colonization and lack of a CO2 effect on AMF colonization in the present study could be due to soil and plant P saturation and its effect on exudation. High-P supply limits colonization by reducing membrane permeability and preventing leakage of exudates that stimulate colonization (Graham et al., 1981; Graham, 2001).

In conclusion, root colonization at high-P supply by G. intraradices reduced plant biomass of sour orange (the high MD genotype) at aCO2. The onset of growth depression coincided with increased rhizosphere respiration, root starch depletion and reduced RGR of AM, compared to NM plants. Enhanced Pn and C supply at eCO2 more than compensated for the C cost of mycorrhizal colonization, resulting in greater biomass gain in AM compared to NM sour orange. Sweet orange was largely unresponsive to root colonization as predicted from its low MD. Our results are consistent with the idea that mycorrhizal-related growth depression at high-P supply is linked to the MD of citrus genotypes and supports the hypothesis that increased C supply at eCO2 can mitigate this growth depression.


We thank Diane Bright, Jessica Cook, Jill Dunlop, Steve Sax, and Shannon Sloan for their technical assistance. We also thank Dr. D. M. Sylvia for supplying Glomus intraradices inoculum. This research was supported by the Florida Agricultural Station and approved for publication as Journal Series No. R-08196.