The influence of vesicular–arbuscular mycorrhizal (M) colonization on biomass production and photosynthesis of Trifolium repens L. was investigated in two experiments in which the foliar nitrogen and phosphorus contents of non-mycorrhizal (NM) plants were manipulated to be no lower than that of M plants. Throughout both experiments there was a stimulation in the rate of CO2 assimilation of the youngest, fully expanded leaf of M compared with NM plants. In addition, M plants exhibited a higher specific leaf area compared with NM plants, a response that maximized the area available for CO2 assimilation per unit of carbon (C) invested. Despite the increased rate of photosynthesis in M plants there was no evidence that the additional C gained was converted to biomass production of M plants. It is suggested that this additional C gained by colonized plants was allocated to the mycorrhizal fungus and that it is the fungus, by acting as a sink for assimilates, that facilitated the stimulation in the rate of photosynthesis of the plant partner.
Using conventional approaches to analyse the plant response to mycorrhizal colonization it is difficult to determine whether impacts upon its growth and C assimilation rate arise from fungal occupancy of the root, or from the improved access to nutrients, particularly phosphorus (P), normally provided by the fungus. It is therefore a prerequisite for the separation of these effects that the phosphorus status of uncolonized ‘control’ plants is not lower than that of those which are mycorrhizal.
Some earlier studies have been designed specifically to remove the confounding effects of mineral nutrition but have produced conflicting results. Brown & Bethlenfalvay (1988) found higher rates of photosynthesis in Glycine max mycorrhizal with Glomus mosseae when their leaf P and nitrogen (N) contents were lower than those of leaves of fertilized NM plants. However, Fredeen & Terry (1988) found no effect of colonization by Glomus fasciculatum on photosynthesis of this species when its leaf P contents were the same as or higher than those of NM plants. Recently, even at very low levels of colonization by Glomus mosseae, enhancement of maximum photosynthetic rate of the youngest, fully expanded leaf of Hordeum vulgare was observed relative to those seen in NM plants of similar foliar P status (Fay, Mitchell & Osborne 1996). In addition, several studies using citrus species have produced contradictory results (Johnson 1984; Graham & Syvertsen 1985; Syvertsen & Graham 1990; Peng et al. 1993).
In the present study we have tested the hypothesis that VA colonization can stimulate photosynthesis and provide biomass increase independently of enhanced nutrient acquisition. The tests involve the production of two types (experiments 1 and 2) of NM plants in which the foliar N and P contents have been manipulated to be similar to, or higher than, those of VA mycorrhizal plants, thus largely eliminating effects of these elements on photosynthesis. In experiment 1 we used clover (Trifolium repens L.) in which NM plants, while having the same foliar P content as their M counterparts, were still of significantly smaller initial biomass. In terms of size, this situation resembles that seen in classical ‘big plant, little plant’ studies of the response to mycorrhizal colonization which abound in the literature. In experiment 2 we again produced clover plants in which both foliar P content and initial biomass were similar in M and NM plants. Pre-infected mycorrhizal plants were used so that high levels of colonization were present from the start of both experiments. In both experiments we have measured the rate of photosynthesis of individual leaves of M and NM plants on a weekly basis and followed the growth responses of such plants using five harvests during a 60–80 d growth period.
MATERIALS AND METHODS
Plant material and growth conditions
Clover (Trifolium repens L.) seeds were germinated on moist filter paper for 4 d. Mycorrhizal colonization of germinated seedlings was induced by transplantation into seed trays containing autoclaved dune sand mixed with chopped, infected root material. Infected root material was obtained from pot cultures of clover colonized by VA mycorrhizal fungi isolated from grassland turfs collected from Aber, North Wales. As previously demonstrated this procedure ensures high levels of colonization by a natural population of VA mycorrhizal fungi (Francis & Read 1994). The aim was to obtain M plants that were colonized by a population of VA mycorrhizal fungi representative of those seen in a natural community. Seedlings were maintained in these trays for 5 weeks to allow colonization of the clover roots during which time the plants were watered with distilled water only. NM plants were grown in trays of dune sand from which root material was omitted.
Two types of NM plants were produced. In experiment 1 NM plants were watered with distilled water only throughout the 5 week colonization period. In experiment 2 NM seedlings were supplemented with 50% Long Ashton nutrient solution to produce seedlings of similar biomass to M seedlings at the end of the colonization stage. A rhizobial inoculant (Nodulaid, Elsoms Seeds Ltd, Spalding, Lincolnshire, UK) was added to the dune sand before use to ensure uniform nodulation of M and NM plants. Seedlings were maintained in a Fisons Fitotron growth cabinet at 20 °C with a 16 h photoperiod (360 μmol m–2 s–1 irradiance). After the 5 week colonization stage M and NM plants were transferred into individual pots of autoclaved dune sand. Pot volumes were 216 cm3 and 1105 cm3 in experiments 1 and 2, respectively. Each pot received 20 cm3 of 50% Long Ashton nutrient solution (all as mmol m–3: NH4NO3 50; CaCl2 40; K2SO4 20; MgSO4·7H2O 15; Na2HPO4·12H2O 13·3; Fe-EDTA 1; H3BO3 0·5; MnSO4·4H2O 0·1; ZnSO4·7H2O 0·001; CuSO4·5H2O 0·001; Na2MoO4·2H2O 0·0005; CoCl2·6H2O 0·0002) once weekly and were watered with distilled water on all other days. Nutrient analysis of the dune sand showed it to contain only trace quantities of macronutrients.
Colonization of clover roots
Roots of M and NM plants were cleared in 100 kg m–3 KOH overnight. Roots were rinsed in 1 volume HCl: 9 volumes water for 20 min and then placed in 500 g m–3 trypan blue in lactoglycerol (1 volume lactic acid, 1 volume glycerol, 1 volume water) for 1 week to stain fungal structures. Roots were destained using several changes of 1 volume glycerol: 1 volume water and mounted in glycerol on a microscope slide. Percentage colonization of roots by M fungi (number of intercepts at which colonization was observed expressed as a percentage of the total number of observations) was determined using the grid line intersect method (Giovannetti & Mosse 1980).
Mycorrhizal and NM plants were harvested on the day of transplant (day 0) and on days 14, 35, 49 and 63 in experiment 1 and on days 18, 37, 51 and 80 in experiment 2. Five M and five NM plants were selected at random at each harvest, and clover shoots were separated into leaves, stems and stolons. Roots were separated from the sand by careful washing under running water. Nodules were carefully excised from the roots of M and NM plants under a dissecting microscope. All these plant components were then dried at 70 °C until constant weight was achieved. Leaf area was determined using a Δ-T Mk2 leaf area meter (Delta-T Devices, Cambridge, UK). Specific leaf area (SLA, leaf area per unit leaf dry weight) was calculated according to Hunt & Parsons (1974). Allometric growth coefficient k (dimensionless) (Pearsall 1927) represents the ratio of the logarithmic growth rate of the shoot to that of the root. The coefficient k was determined for M and NM clover plants using the equation:
where, Sw = shoot dry weight, Rw = root dry weight and k and b are constants (Hunt 1990).
Phosphorus and nitrogen determinations
The foliar concentration of P and N was determined on dried material of M and NM plants from each harvest. A subsample of finely ground leaf material was suspended in concentrated sulphuric acid (≈20 mg dry weight leaf cm–3 H2SO4) and heated to 340 °C in a digestion block. Carbon was removed from the sample during heating by the careful addition of 100 mm3 aliquots of hydrogen peroxide (100 volumes, Fisons Scientific, Loughborough, UK) until a colourless liquid was obtained (van Lierop 1976). The samples were heated for a further 15 min, allowed to cool and then made up to a volume of 25 cm3 using nanopure distilled water. Total N and total P were determined by flow injection analysis (Anonymous 1990) using a Tecator flow injection analyser (Perstorp Analytical, Maidenhead, Berkshire, UK). Recovery of a certified reference hay powder (Office of Reference Materials, Laboratory of the Government Chemist, Teddington, Middlesex, UK) was 94% for both N and P.
The rate of photosynthesis of clover plants
Because a primary objective of the present study was to determine whether rates of photosynthesis differed in M and NM plants of similar nutrient content a standard unit of measure, the youngest fully expanded leaf, was used for analysis. The rate of photosynthesis (330 cm3 m–3 CO2, 360μmol m–2 s–1 irradiance) of the youngest fully expanded leaf of five M and five NM clover plants was determined weekly using an LCA4 infrared gas analyser and LCA4 broad-leaf cuvette (Analytical Development Company, Hoddesdon, UK). The leaf of a randomly selected plant was placed into the leaf chamber. Actinic light (300–700 nm) was then supplied to the leaf by a Schott KL1500T lamp and the leaf allowed to reach steady-state photosynthesis (≈25 min), at which point the net rate of photosynthesis was recorded each minute for 8 min and the average calculated.
The data were subjected to analysis of variance (ANOVA) using the statistical package Minitab version 9·2.
Mycorrhizal colonization and nodulation of clover roots
As mycorrhizas were synthesized on clover seedlings before the beginning of experiments 1 and 2, colonization was high from the beginning of each experiment. More than 85% of the root length was colonized by mycorrhizal fungi throughout experiment 1. At least 79% of the root length was infected during experiment 2. Non-mycorrhizal controls remained free of infection. The possibility that other fungal colonists were transferred to the M plants in the original inoculum was checked but in the course of analysis of M root systems no such fungal infections were observed.
Both M and NM plants were treated with a rhizobial inoculant during the period when mycorrhizas were synthesized to ensure that plants in both categories were nodulated to a similar extent. In experiment 2 the dry weight of nodules present on the roots of M plants was not significantly different from that on the roots of NM plants on day18 (2·4 ± 0·3 mg compared with 2·0 ± 0·2 mg for M and NM plants, respectively) and day 51 (31·7 ± 5·9 mg compared to 29·2 ± 2·3 mg for M and NM plants, respectively).
The effect of mycorrhizal colonization on the nutrient composition and biomass production of clover
The nutrient applications enabled the production of M and NM plants in which the foliar P concentration was not significantly different, with the exception of day 49 when the NM plants had a greater amount of P per unit dry weight (P < 0·05) (Fig. 1). In general, the foliar N concentration was higher in NM plants compared with M plants (Fig. 1).
At each of the first three harvests (days 0, 14 and 35), the total dry weight of M plants was significantly greater than that of NM plants (P < 0·01). The percentage increase in the total dry weight of M compared with NM plants was 147% (day 0), 238% (day 14) and 255% (day 35) (Fig. 2 and Table 1). However, there was no significant difference in the total dry weight of both sets of plants at the final two harvests (days 49 and 63).
Table 1. . The total dry weight (DW) (mg), the dry weight of the leaves, stems, roots and stolons and the leaf area (cm2) of mycorrhizal and non-mycorrhizal clover plants in experiment 1. The percentage change (%) in each parameter for the mycorrhizal compared with the non-mycorrhizal plants is indicated. Values represent the mean of five replicates
There were no significant differences in the percentage of the total dry weight allocated to the different plant organs (leaf, stem, root and stolon) of M compared with NM plants on days 0, 14 and 35 (data not shown). However, on days 49 and 63, ≈10% more of the total dry weight of M plants was allocated to the roots of such plants at the expense of leaf dry weight in comparison with NM plants.
The dry weights of the leaves and stems of M plants were significantly greater than NM plants at days 0, 14 and 35 (Table 1). However, the dry weight of the roots of M plants was significantly greater compared with NM plants at all harvests (Table 1). The leaf area of M plants was significantly greater than that of NM plants on days 0, 14 and 35 but significantly lower than NM plants thereafter (Table 1). The dry weight of the stolons of M compared with NM plants was similar throughout the experiment (Table 1).
The root/shoot ratio of M compared with NM clover plants was not significantly different throughout this experiment (Fig. 3). A plot of loge (shoot dry weight) against loge (root dry weight) allows elucidation of whether mycorrhizal colonization merely slows down the growth of the host plant, or influences partitioning between above- and below-ground parts. In the present experiment the allometric coefficient, k, of M plants (k = 0·98) was not significantly different from that of NM plants (k = 1·12) (Fig. 4). Mycorrhizal plants exhibited a significantly higher (P < 0·05) SLA compared with NM plants on days 0, 14 and 35 harvest suggesting that the leaves of M plants were thinner. Thereafter, the SLA of both sets of plants was similar (days 49 and 63) (Fig. 5).
In experiment 2, the foliar P concentration of NM plants was significantly greater than that of M plants from day 37 (P < 0·05) (Fig. 1). With the exception of the initial time point, the foliar N concentration of M plants was not significantly different from that of NM plants (Fig. 1). The total dry weight of M and NM plants was similar through most of the experiment (days 0, 18, 51, 80) with the exception of day 37 when the percentage increase in the total dry weight of M compared with NM plants was 141% (Fig. 2 and Table 2). There were no significant differences in the proportion of the total dry weight allocated to the different plant organs of M and NM plants during each of the first three harvests (days 0, 18 and 37) (data not shown). However, compared with NM plants, a greater proportion of the total dry weight of M plants was allocated to stolons (11%) at the expense of leaf (7%) and stem (3%) dry weight on day 51 while, on day 80, M plants partitioned more dry weight into stolons (14%) at the expense of root dry weight (13%). In general, there were no significant differences in the dry weight of the leaves, stems, roots or stolons or in the leaf area of M compared with NM plants throughout most of the experiment (Table 2). However, on day 37, the dry weight of the leaves, stems and roots and the leaf area of M clover plants were significantly higher than NM plants (Table 2). The root/shoot ratio of M compared with NM clover was not significantly different throughout this experiment (Fig. 3). In this experiment the allometric coefficient of M plants (k = 1·05) was not significantly different from that of NM plants (k = 0·95) (Fig. 4). There was, again, a trend for SLA to be higher in M plants after the initial harvest although the differences between M and NM plants were significant only at day 37 (P < 0·05) (Fig. 5).
Table 2. . The total dry weight (DW) (mg), the dry weight of the leaves, stems, roots and stolons and the leaf area (cm2) of mycorrhizal and non-mycorrhizal clover plants in experiment 2. The percentage change (%) in each parameter for the mycorrhizal compared with the non-mycorrhizal plants is indicated. Values are means of five replicates
The effect of mycorrhizal colonization on the rate of photosynthesis of clover
The rate of photosynthesis of the youngest, fully expanded leaf of M plants was significantly higher than that of NM plants until day 56 after which the rate of photosynthesis was similar in both sets of plants (P<0·05) (Fig. 6). In this experiment the percentage increase in the rate of photosynthesis of M compared with NM plants was 393% (day7), 101% (day 14) and 15 ± 6% (average for days 21–63). There was no significant difference in the rate of transpiration or the stomatal conductance of M plants compared with NM plants (data not shown).
Throughout experiment 2 the rate of photosynthesis of the youngest, fully expanded leaf of M plants was again substantially higher than that of NM plants, with the exception of day 32 (Fig. 6). The percentage increase in the rate of photosynthesis of M compared with NM plants was 131% (day 8), 106% (day 17), 49% (day 24) and 19 ± 6% (average for days 32–53). Again, there was no significant difference in the rate of transpiration or the stomatal conductance of M plants compared with NM plants (data not shown).
The mechanisms by which mycorrhizal colonization may influence the biomass partitioning and photosynthetic metabolism of the host plant independently of mineral nutrient effects resulting from the presence of the mycobiont are poorly understood. To enable the study of such processes our main concern was to produce NM plants whose mineral status was not limiting with respect to that of M plants. In two separate experiments we manipulated the mineral nutrient, in particular P, status of NM leaves to be similar to or higher than that of M plants so that the well documented phosphorus growth response of M plants in comparison with unmatched NM plants could be avoided (Daft & Nicolson 1966; Gerdemann 1968; Hayman & Mosse 1971; Sanders 1975; Bowen, Bevege & Mosse 1975).
In experiment 1 the plants were initially of unequal size, a normal situation in conventional comparative studies of M and NM plants. After an initial lag, the rate of dry-weight accumulation of NM plants was greater than that of M plants for a substantial period of this experiment, although their absolute dry weight was much smaller than that of their M counterparts until the end of the experiment. The small pot size employed in experiment 1 may have led to some constraint on root growth of M plants towards the end of this experiment, allowing NM plants to attain a similar final whole plant biomass. However, as the dry weight of M roots was significantly higher than that of NM roots throughout the whole time course it would appear that these constraints were not severe. In experiment 2, where the initial biomass of both sets of plants was similar at the start of the experiment, the total dry weight of M and NM plants remained similar throughout the time course, suggesting that the C requirements of the mycorrhizal fungus had little overall impact on the growth of the plant partner.
Although the pattern of dry-weight accumulation differed between M and NM plants in each experiment there was little evidence that mycorrhizal colonization substantially altered the partitioning of biomass between different organs of M compared with NM plants. This is reflected in the similarity of the root/shoot ratios and allometric coefficients of M and NM plants in both experiments, which indicate that mycorrhizal colonization did not proceed at the expense of root dry-matter production. Previous studies reveal a range of effects of mycorrhizal colonization on partitioning to root biomass. Several have reported a decline in the root/shoot ratio upon mycorrhizal colonization, particularly in nonmatched M and NM plants, a situation that may indicate partitioning of C to the fungus at the expense of root production while maintaining an overall neutral effect on C allocation within the plant (Hayman & Mosse 1971; Sanders 1975; Smith 1982; Hall, Johnstone & Dolby 1984; Thomson, Robson & Abbott 1986; Bass & Lambers 1988; Fredeen & Terry 1988; Smith & Gianinazzi-Pearson 1990). However, in citrus species mycorrhizas tend to increase root biomass allocation, especially under conditions of high P supply (Graham & Syvertsen 1984; Douds et al. 1988; Peng et al. 1993; Eissenstat et al. 1993).
One major effect of M colonization on the growth of the plant partner was the substantial increase in the SLA of the autotroph. The increase in SLA indicates that the leaves of M plants were thinner than those of their NM counterparts. Given the substantially higher rates of photosynthesis observed in leaves of M plants in both experiments this suggests that the rate of photosynthesis, and hence the amount of C fixed per unit investment in photosynthetic machinery, was much greater in M plants, even in the latter stages of both experiments when the difference in the rate of photosynthesis between M and NM plants was reduced. Similar increases in SLA have been reported upon mycorrhizal colonization of Allium porrum and Glycine max (Snellgrove et al. 1982; Fredeen & Terry 1988).
In both experiments sustained, elevated rates of photosynthesis of the youngest, fully expanded leaf of M clover plants were observed in comparison with NM plants. It appears likely, in the absence of lower tissue P concentrations in NM compared with M plants, that the stimulation of photosynthetic rate was attributable to a non-nutritional impact of fungal colonization upon C assimilation. Stitt (1990) emphasizes the importance of the cytosolic pool of inorganic P as a major determinant of the rate of photosynthesis, but as it is difficult to measure this pool with accuracy we have attempted, by application of exogenous P, simply to ensure that P does not limit the rate of photosynthesis in either NM or M plants.
The increase in photosynthesis observed in this study was similar to that observed by Johnson (1984) and Brown & Bethlenfalvay (1988) in Citrus aurantium mycorrhizal with Glomus intraradices, and Glycine max colonized by Glomus mosseae, respectively, in comparison with NM plants of similar foliar N and P status. These authors attributed the increase in the rate of photosynthesis to increased sink strength resulting from the presence of the mycobiont. Recently, even at very low levels of colonization by G. mosseae, enhancement of the maximum rate of photosynthesis of the youngest, fully expanded leaf of Hordeum vulgare, grown at a low level of P supply, was observed relative to that seen in NM plants of similar foliar P status (Fay et al. 1996). In contrast, in other studies there was no stimulation of the rate of photosynthesis in response to VA mycorrhizal colonization when compared with NM plants even though foliar P content, size and growth rate were similar (Fredeen & Terry 1988; Syvertsen & Graham 1990; Eissenstat et al. 1993; Peng et al. 1993). Clearly these apparent differences require further investigation.
There is increasing evidence to support the view that the process of C assimilation can be influenced by the strength of the sinks to which photosynthates are allocated (Herold 1980; Gifford & Evans 1981; Farrar 1992). If this is the case, mycorrhizal colonization of the root, by increasing its sink strength, may be expected to increase the rate of photosynthesis of the plant such that the ‘cost’ imposed upon the plant's C economy by the growth of the fungus is reduced or even eliminated. In both of the present experiments the sustained, elevated rates of photosynthesis in conjunction with the increased SLA observed in M plants should result in a substantial additional C pool, from which allocation to growth or maintenance of either symbiont could take place. There appears, however, to be no evidence, in the form of biomass gain, for additional allocation of C to the autotroph. Therefore, it is suggested that the additional C was allocated to the fungus, so eliminating any ‘cost’ in terms of growth to the autotroph.
In conclusion, the results of the two experiments suggest that the stimulation in the rate of photosynthesis of M plants was attributable to the additional fungal sink arising through mycorrhizal colonization of the roots of the autotroph. This hypothesis is being investigated by quantifying whole-plant CO2 exchange, the distribution of carbohydrates in leaves and roots and the activity of sucrolytic enzymes in the roots of M and NM plants.
This research was funded by a NERC grant (GR3/9258).