Plant and Soil

, Volume 294, Issue 1, pp 125–136

Mycorrhizas in agroforestry: spread and sharing of arbuscular mycorrhizal fungi between trees and crops: complementary use of molecular and microscopic approaches

Authors

    • Centre for Ecology and Hydrology
  • J. Wilson
    • Centre for Ecology and Hydrology
  • R. C. Munro
    • Centre for Ecology and Hydrology
  • S. Cavers
    • Centre for Ecology and Hydrology
Original Article

DOI: 10.1007/s11104-007-9239-z

Cite this article as:
Ingleby, K., Wilson, J., Munro, R.C. et al. Plant Soil (2007) 294: 125. doi:10.1007/s11104-007-9239-z

Abstract

The spread of arbuscular mycorrhizal (AM) fungi from tree to crop roots was examined by molecular and microscopic methods in a glasshouse study. Growth of Calliandra calothyrsus Meissner trees inoculated with isolates of the AM fungi Glomus etunicatum Becker and Gerdemann and Gigaspora albida Schenck and Smith was monitored over an 18-month period. Three successive ‘intercrops’ of beans or maize were sown at 25, 50 and 75 cm distances from the tree and harvested during this period. At each crop harvest, the distribution of tree and crop roots and the spread of the inoculant fungi were determined using traditional microscopic methods and fungal specific primers. Both inoculants greatly improved the growth of the trees and colonization spread to the crops once the trees were 6 months old. However, benefits of inoculation to crop growth were not observed due to increased competition from the larger inoculated trees growing in a restricted soil volume. Of the two inoculant fungi, Glomus etunicatum appeared to be more mobile as it spread more rapidly, formed higher levels of colonization at increasing distances from the tree and was responsible for most of the mycorrhizal cross-contamination. In contrast, colonization of tree and crop roots by Gigaspora albida was higher nearest the tree. This work demonstrated the benefits of mycorrhizal fungus inoculation for tree growth and confirmed that trees and crops share the same AM fungi. Trees may therefore act as reservoirs of mycorrhizal fungi, either inoculant or indigenous, for surrounding crops or other annual vegetation. It was also shown that tree pruning, the normal practice in agroforestry systems, did not reduce mycorrhizal colonization or prevent spread to crops. However, the slow rates of inoculant spread found here suggest that it may take years before inoculants benefit the growth of crops sown several metres from the tree. The work also demonstrated that microscopic quantification of mycorrhizal colonization and the use of molecular probes to identify specific fungi within roots can complement each other effectively. Molecular probes were more sensitive at detecting mycorrhizal fungi than microscopic methods, but did not discriminate between full mycorrhizal structures and traces of hyphae.

Keywords

Calliandra calothyrsusGigaspora albidaGlomus etunicatumMolecular probesTree-crop linkages

Abbreviations

AM

Arbuscular mycorrhizal

RLD

Root length density

PCR

Polymerase chain reaction

BEG

International bank for the glomeromycota

Introduction

Fast-growing, multi-purpose tree species are widely planted on farms in semi-arid Africa as they perform a key role in stabilizing and improving farm soils while providing many additional and varied products such as timber, fodder and fruit and increasing total farm productivity through exploitation of different niches, above and below ground (Sanchez et al. 1997). Many of the tree species employed are leguminous and form symbiotic associations with N2-fixing bacteria (rhizobia) and arbuscular mycorrhizal (AM) fungi, which enable them to sustain growth in the phosphorus and nitrogen deficient soils typical of the region. These soils are often degraded through over-cultivation and erosion, and such intensification of land-use may lead to insufficient or ineffective populations of microsymbionts (Alvarez-Solis and Anzueto-Martinez 2004). In these cases, inoculation with effective rhizobia and AM fungi may be needed for the re-establishment of trees, while long-term improvements in soil fertility and growth of the crops will require land management regimes which sustain and promote mycorrhizal populations (Sieverding 1991).

As AM fungi are the predominant mycorrhizal type in dry tropical soils and associate with a wide range of plant species, they have the potential to benefit the growth of both tree and crop species in agroforestry systems. Tree legumes such as Senna siamea, Gliricidia sepium and Calliandra calothyrsus have shown high-mycorrhizal dependency and respond to inoculation (Habte and Turk 1991; Ingleby et al. 2001). Similarly, field crops such as cassava are known to be obligately dependent on AM fungi, and inoculation using several AM fungus inoculants has been highly beneficial to crop yields in a range of soils (Howeler et al. 1987). However, these responses vary widely according to the host species, the AM fungus inoculants used, soil fertility and the levels of indigenous populations of AM fungi, and these factors should be investigated before AM fungal inoculants are selected (Sieverding 1991).

The importance of maintaining active populations of AM fungi in agroforestry soils in order to sustain crop productivity has also been demonstrated (Sieverding and Leihner 1984; Dodd et al. 1990). More recently, Arihara and Karasawa (2000) have shown that maize yields were better and mycorrhizal fungus colonization higher in maize crops cultivated after other mycorrhizal crops, than in maize cultivated after non-mycorrhizal crops. AM fungus inoculum in the soil normally occurs as spores, mycorrhizal roots and mycelial networks, and Miller (2000) attributed early infection of maize seedlings and increased final grain yield to the key role AM mycelial networks play in enhancing phosphorus absorption in young plants. Although most sensitive to disturbance, AM mycelial networks are primarily responsible for the rapid colonization of new roots, and have been shown to retain their capacity to colonize roots even after long periods of drought typical of tropical regions (Brundrett and Abbott 1994).

It is now widely accepted that AM mycelial networks form links between plant species in ecosystems, and that they are responsible for the transfer of nutrients between different plant species (Read 1991). Haselwandter and Bowen (1996) proposed that AM fungi associated with agroforestry tree species may serve an additional role by maintaining active AM propagules in the soil, which could then rapidly colonize roots of emerging crop seedlings. Subsequent studies have supported this view: Leakey et al. (1999) reported that maize grown in soil taken from close to S. siamea formed more mycorrhizas than when it was grown in soil collected at 2 m distance, while Diagne et al. (2001) examined soils from agroforestry systems in Senegal and found beneficial effects of Acacia tortilis trees on mycorrhizal fungus colonization and growth of millet seedlings. The role of perennial trees in maintaining AM fungus inoculum and in sustaining mycelial networks for short-lived crops may therefore be an unintended benefit of agroforestry systems and provide an alternative approach to the use of cover crops to build up soil inoculum.

This paper reports the results of a glasshouse study, which examined the spread of AM fungi from tree to crop roots, and the resulting effects on plant growth. The experiment used C. calothyrsus, a widely planted, multi-purpose, leguminous agroforestry tree species as the host tree, inoculated with two AM fungus inoculants and co-planted with maize or beans in sequence to simulate the cropping patterns in Kenya. ‘Traditional’ assessments of mycorrhizal colonization by staining and light microscopy were combined with molecular methods in order to accurately monitor the spread and distribution of the inoculant fungi.

Materials and methods

Design and set up of glasshouse experiment

On 6 February 2004, 75 cm3 pots containing a sterilized loam/grit-sand mixture and 20 g of root/soil inoculum from either Glomus etunicatum Becker and Gerdemann (BEG 176) or Gigaspora albida Schenck and Smith (BEG 173) pot cultures, or an autoclaved mixture of these inoculants, were sown with C. calothyrsus Meissner (Flores, ex. Maseno) seeds. These mycorrhizal fungus isolates originated from soil samples collected in proximity to C. calothyrsus in Honduras (soil pH 5.5) and Kenya (soil pH 5.9), respectively. Prior to registration with the International Bank for the Glomeromycota (BEG), they were known by their isolate numbers ‘Glomus etunicatum 1’ and ‘Gigaspora albida 2’ and had been shown to form mycorrhizas abundantly and promote the growth, shoot phosphorus and nodule dry mass of C. calothyrsus (Lesueur et al. 2001). After 11 days, germinating seedlings were thinned to one per pot and all seedlings were inoculated with 2 ml of a Rhizobium suspension comprising of two isolates also known to be effective with C. calothyrsus (isolates KWN35 and KCC6; Lesueur et al. 2001). About 6 weeks after inoculation, three seedlings were sampled from each treatment to examine their mycorrhizal status and confirm the effectiveness of the inoculation procedure. About 9 weeks after inoculation, the seedlings were transplanted, one per trough, to 100 × 20 × 20 cm3 troughs filled with a sterilized loam/grit-sand/coir mixture (3:3:1), pH 5.8, containing 63, 4.2 and 36 mg kg−1 of extractable NPK respectively, intended to simulate a P-deficient tropical soil. To improve drainage, the troughs were first lined with a 2–3 cm layer of coarse pebbles so that the actual depth of soil mixture in the troughs was ∼15 cm. Seedlings were planted 7.5 cm from one end of the trough. The three inoculation treatments were replicated in eight randomized blocks, with each treatment represented once within each block. The troughs were located in a glasshouse set to provide a day/night temperature regime of 28/20°C with high-pressure mercury vapour lamps to supplement natural sunlight and produce a day length of 14 h.

During the course of the study, crops were sown and harvested three times. On 2 June 2004 (15 weeks after AM fungus inoculation), Phaseolus vulgaris L. (seedlot Mwezi Moja GLP 1127 ex. Kenya 25/3/03) seeds were sown in the troughs 25 and 50 cm from the tree. After 1 week, emerging seedlings were thinned to one per distance. For this first cropping period, plants were harvested 6 weeks after sowing so that primary mycorrhizal colonization could be related to crop growth and the effects of the inoculation treatments. Subsequent cropping periods were extended to allow the crop plants to reach maturity before harvest, thus following cropping patterns in the field. On 6 September 2004 (28 weeks after AM fungus inoculation), Zea mays L. (seedlot H614D ex. Kenya 25/3/03) seeds were sown in the troughs 25 and 50 cm from the tree and thinned to one per distance as before. Plants were harvested 10 weeks after sowing. Finally, on 13 May 2005 (64 weeks after AM fungus inoculation), the trees were pruned to 30 cm height, removing most of the leaves and above-ground biomass of the inoculated plants, and the same seedlot of Z. mays was sown 25, 50 and 75 cm from the tree. Plants were harvested 12 weeks after sowing. Shoot pruning is regularly carried out in tropical agroforestry, and was done to evaluate its effects on mycorrhizal colonization and to reduce the intense tree-crop competition observed in the troughs in 2004.

Sampling and assessment

Growth of C. calothyrsus seedlings was monitored during the experiment by measuring stem diameter. Measurements were made every 2 weeks in 2004 and then every 4 weeks during 2005. Crop growth was assessed by taking weekly height measurements of the plants and measuring shoot dry weight at harvest. At harvest, crop shoots were severed at ground level, not uprooted. At the time of each crop harvest, two soil cores (1.6 cm diameter × 10 cm depth: ∼20 cm3 soil) were removed at each distance, and tree and crop roots were extracted for molecular and microscopic assessment of mycorrhizal fungus colonization. This coring depth focussed on the lateral and fine root development, which was concentrated in the upper soil layers, with only coarse tap roots developing through the pebbles at the base of the troughs. The root distribution in the troughs was confirmed after 28 weeks with the destructive harvest of troughs from block two, in which the uninoculated tree had become contaminated by Glomus etunicatum. In 2004, cores were removed at 0, 25 and 50 cm from the tree, and at 0, 25, 50 and 75 cm in 2005. Four, seven and six blocks were assessed in July 2004, November 2004 and August 2005 respectively. Coring holes were re-filled with the same soil mixture and care was taken to avoid re-filled holes on subsequent sampling occasions. Root sampling for molecular work demanded a rigorous approach in order to ensure that hyphal fragments did not cross-contaminate the samples: corers and all other implements used were surface sterilized between each sample. Soil from the two cores was bulked for each distance and spread in sterile 14 cm Petri dishes. Roots were first removed aseptically, washed in sterile water and separated into tree and crop fractions. These fractions were then cut into 1 cm root fragments and mixed, before ten fragments were randomly sampled and transferred to Eppendorf tubes for DNA extraction. The remaining roots were stained in Trypan blue (Koske and Gemma 1989) prior to assessment of root length and the proportion that was mycorrhizal, using the gridline intersect method (Tennant 1975). Root samples were used preferentially for molecular analysis and, in a few instances, insufficient roots remained for assessment of mycorrhizal colonization. As mycorrhizal colonization in C. calothyrsus roots was often difficult to observe under the dissecting microscope due to strongly staining epidermal cells, sub-samples of these roots were mounted on glass slides to confirm the presence of colonization under the compound microscope. Root length density (RLD) (cm root 100 cm−3 soil) was calculated.

Data analysis

For tree growth, a one-way analysis of variance (ANOVA) was used, with inoculation as the treatment factor. For all other parameters, differences between treatments were examined by two-way ANOVA using inoculation and distance from the tree as treatment factors. Data were examined for normality (Anderson-Darling, Cramer-von Mises and Watson tests; Stephens 1974), homogeneity of variances (Bartlett’s test; Sokal and Rohlf 1995) and transformed where necessary to conform with the requirements of ANOVA. Differences between means were compared using Fisher’s LSD test when the F-test from ANOVA was significant at P ≤ 0.05.

Use of molecular probes

DNA was extracted from the roots using a Qiagen DNeasy plant mini kit after grinding for 30 s at 30 Hz in a Retsch MM300 grinder. DNA extracts were quantified by eye after electrophoresis in 1% agarose gel and either retained as neat extracts or diluted 1:20 with deionized water. Extracts were used as template DNA for amplification by polymerase chain reaction (PCR): for each sample, PCR was carried out in triplicate and, at each stage, water samples and DNA extracts from spores of the two inoculant fungi were included as negative and positive controls respectively. To test for the presence of the inoculant fungi, nested PCRs were performed using the universal primers ITS1 (White et al. 1990) and NDL22 (van Tuinen et al. 1998) at the first stage, and primers developed for Glomus etunicatum BEG 176 and Gigaspora albida BEG 173 (Walters and MacDonald unpublished) at the second stage. Template DNA for the second stage PCR consisted of 2.0 μl of pure PCR product from stage 1. At both first and second stages, 25 μl PCR reactions contained 2.0 μl template DNA, 2.5 μl of 10 mM dNTPs (Promega, Mannheim, Germany), 1.0 μl of each 25 μM primer (MWG Biotech, Ebersberg, Germany), 0.5 μl of 0.4 μg μl−1 bovine serum albumin, 2.5 μl 10 ×  PCR buffer (New England Biolabs, Beverly, MA), 1 U Taq DNA polymerase (New England Biolabs) and 15.3 μl deionized water. Reactions were covered with foil seals and run on a ThermoHybaid MBS 0.2 G Thermal Cycler for one denaturing step of 94°C for 5 min then 30 cycles of 94°C for 60 s, 58°C for 60 s, 72°C for 60 s and a final extension step of 72 C for 10 min. PCR products were visualized by electrophoresis on 1% agarose gels. A successful amplification in any one of the three triplicate PCRs was considered to indicate presence of the target fungus: we considered triplication as representing sampling power rather than PCR verification, which was provided by successful positive control amplification.

Results

Tree growth

From the time of transplanting to the troughs in 2004 until the end of the experiment in 2005, C. calothyrsus seedlings inoculated with G. etunicatum and G. albida were significantly (P < 0.001) greater in stem diameter than the uninoculated control tree seedlings (Fig. 1). No significant differences were observed between trees inoculated with G. albida and those inoculated with G. etunicatum. Figure 1 also indicates a reduction in the growth rate of the inoculated trees after about 40 weeks.
https://static-content.springer.com/image/art%3A10.1007%2Fs11104-007-9239-z/MediaObjects/11104_2007_9239_Fig1_HTML.gif
Fig. 1

Stem diameter of inoculated and uninoculated Calliandra calothyrsus trees in a glasshouse trough experiment during 2004–2005 [error bars indicate ± SE (n = 7); horizontal bars indicate cropping periods C1–3]

Crop growth

In July 2004, shoot dry weight of P. vulgaris harvested after 6 weeks was not significantly affected by inoculation treatment or distance from the tree (Table 1). In November 2004, shoot dry weight of Z. mays plants after 10 weeks was significantly (P < 0.001) higher at 50 cm distance from the tree than at 25 cm, indicating that crops growing closest to the trees were suffering from competition, especially with the larger inoculated trees. In August 2005, shoot dry weight of Z. mays after 12 weeks was significantly (P < 0.001) higher in the uninoculated troughs where trees were smaller. However, growth was much better across all treatments, suggesting that shoot pruning of the trees in May 2005 had reduced competition, especially from the larger inoculated trees.
Table 1

Shoot dry weight (g) of crop plants grown at different distances from Calliandra calothyrsus trees inoculated with two different AM fungal isolates or left uninoculated

Inoculation treatment

Glomus etunicatum

Gigaspora albida

Uninoculated

P-value

Distance from tree (cm)

25

50

75

25

50

75

25

50

75

Inoculation

Distance

Inoculation × Distance

P. vulgaris harvest July 2004

2.60

2.46

n.a.a

2.58

2.72

n.a.

2.66

3.33

n.a.

0.403

0.445

0.513

Z. mays harvest November 2004

0.99

3.75

n.a.

0.75

3.92

n.a.

1.37

4.05

n.a.

0.833

<0.001

0.928

Z. mays harvest August 2005

7.5

5.0

16.6

3.7

6.0

5.9

11.2

18.2

17.4

<0.001

0.075

0.229

Values are means of 4 (July 2004), 7 (November 2004) and 6 (August 2005) replicates

aSamples were not assessed (n.a.) at 75 cm distance in July and November 2004

Root growth

In July 2004, tree RLD was greatest on inoculated trees (P < 0.001) and nearest the tree (P < 0.001), whereas crop (P. vulgaris) RLD was greatest further away from the tree (P < 0.001) (Table 2). Similar differences were observed in November 2004 and August 2005, when Z. mays plants were harvested. However, after tree pruning in 2005, concentrations of crop roots found near the tree were much higher than those found in 2004. The results also show that, by 2005, roots of inoculated trees had extended throughout the trough.
Table 2

Root length density (cm 100 cm3 soil) and mycorrhizal infection (% root length) of tree and crops at the time of three crop harvests made during 2004–2005

Inoculation treatment

Glomus etunicatum

Gigaspora albida

Uninoculated

P-valuea

Distance from tree (cm)

0

25

50

75

0

25

50

75

0

25

50

75

Inoculation

Distance

Inocualtion × Distance

Tree root length density

July 2004

188

86

76

n.a.b

301

73

40

n.a.

69

4

0

n.a.

<0.001

<0.001

0.500

November 2004

362

156

147

n.a.

841

261

81

n.a.

102

4

0

n.a.

<0.001

<0.001

0.113

August 2005

582

481

180

104

767

417

115

87

191

64

34

0

<0.001

<0.001

0.331

Crop root length density

July 2004

3

68

229

n.a.

4

63

186

n.a.

33

75

231

n.a.

0.142

<0.001

0.710

November 2004

5.2d

105b

141a,b

n.a.

6.5d

61.2c

151a

n.a.

23.4d,c

136a,b

125a,b

n.a.

0.128

<0.001

0.007

August 2005

208

372

412

415

242

214

323

427

453

573

676

582

<0.001

0.016

0.703

Tree mycorrhizal colonization

July 2004

31.0

8.3

2.6

n.a.

40.6

4.7

0.4

n.a.

17.3

3.6

0

n.a.

0.057

<0.001

0.549

November 2004

28.4b

6.1c

0.9d

n.a.

48.6a

24.8b

1.1d

n.a.

2.5c,d

0d

0d

n.a.

<0.001

<0.001

<0.001

August 2005

29.1

23.3

6.1

2.5

42.6

40.7

10.5

0.1

7.1

3.1

0

0

<0.001

<0.001

0.159

Crop mycorrhizal colonization

July 2004

n.r.

0

0

n.a.

n.r.d

0

0

n.a.

0

0

0

n.a.

November 2004

n.r.

20.5a

1.6b

n.a.

n.r.

19.3a

0.4b

n.a.

11.0b

0b

0.6b

n.a.

0.199

<0.001

<0.001

August 2005

45.6b

48.8b

44.7b

27.9c

75.0a

71.9a

35.8b,c

10.7d

4.4d,e

6.3d,e

0e

0e

<0.001

<0.001

<0.001

Values are means of 4 (July 2004), 7 (November 2004) and 6 (August 2005) replicates

aSquare root and angular transformations were performed on root length density and mycorrhizal colonization for statistical analysis; untransformed means are shown in this Table

bRoot samples were not assessed (n.a.) at 75 cm distance in July and November 2004

cLetters indicate significant differences within each row for the inoculation x distance interaction as determined by Fisher’s LSD test, when P < 0.05 as determined by ANOVA

dNo roots (n.r.) were present in these samples

Mycorrhizal colonization

About 6 weeks after inoculation, and prior to transplanting into the troughs, both inoculants had formed mycorrhizas on the C. calothyrsus seedlings: those inoculated with G. etunicatum had 12% of their root length colonized, while those inoculated with G. albida had 40%.

Subsequently, in July 2004, mycorrhizal colonization of tree roots was greatest nearest the tree (P < 0.001), but was not found in any crop roots, although very few crop roots were found near the tree where most tree root mycorrhizal colonization occurred (Table 2). Although significant differences between inoculation treatments were absent (P = 0.057), colonization of G. albida inoculated trees close to the stem remained at 40%, while that of G. etunicatum inoculated trees was 31%. In November 2004, a significant inoculation × distance interaction was found for mycorrhizal colonization of tree roots, with colonization of G. albida inoculated trees greater than that of G. etunicatum inoculated trees at 0 and 25 cm from the tree. Although both inoculants had colonized roots at 50 cm from the tree, colonization was greatest nearest the tree and decreased at 25 and 50 cm from the tree. By this time, mycorrhizal colonization by both inoculant fungi was well established on crop roots at 25 cm from the tree.

In August 2005, mycorrhizal colonization of tree roots followed a similar pattern to the previous November. Mycorrhizal colonization of crop roots was also greatest in inoculated troughs and nearest the tree. However, a significant inoculation × distance interaction (P < 0.001) showed that although levels of colonization of crop roots by G. albida remained higher than those of G. etunicatum, and those growing with uninoculated trees remained the lowest, colonization of G. albida crop roots decreased at 50 and 75 cm from the tree whereas colonization by G. etunicatum was more consistent and only decreased at 75 cm from the tree. The results in August 2005 also showed that high levels of colonization were present on both tree and crop roots despite the heavy pruning of the trees prior to sowing this crop. Although some mycorrhizal colonization was found in tree and crop roots from uninoculated troughs, the more detailed data presented in Figs. 24 shows that this was sporadic colonization of individual plants rather than widespread contamination.
https://static-content.springer.com/image/art%3A10.1007%2Fs11104-007-9239-z/MediaObjects/11104_2007_9239_Fig2_HTML.gif
Fig. 2

ac Extent of mycorrhizal colonization (% root length) determined microscopically, and origin (G. etunicatum, G. albida or other) of mycorrhizal fungus determined by molecular methods, on roots of trees (C. calothyrsus) and crops (P. vulgaris) growing together in troughs. Samples collected in July 2004 from cropping period C1 (see Fig. 1). Trees were previously inoculated with (a) G. etunicatum, (b) G. albida or (c) not inoculated. Samples were taken at different distances from the tree (0, 25 and 50 cm) in four replicate troughs. X-axis shows block numbers of samples taken at different distances. Data for trees and crops taken from the same soil cores are presented in adjacent columns. The presence of a small coded section at the top of a bar indicates molecular confirmation of one of the two inoculant fungi

Rate of spread and molecular identification of the inoculant fungi

In order to compare the results from microscopic and molecular assessments, this section presents data from the individual troughs rather than treatment means. Figures 24 show the % of colonization as determined by conventional staining, and the identity of the causal AM fungi as determined by the molecular probes. These assessments were made on parallel sub-samples of roots, so that the figures indicate the level of mycorrhizal fungus colonization in each sample and the presence or absence of the two inoculant fungi.

At the time of the P. vulgaris crop harvest, 21 weeks after inoculation of the trees and 6 weeks after crop sowing, levels of mycorrhizal fungus colonization in C. calothyrsus roots sampled at 0 cm varied from 21 to 51% for those inoculated with G. etunicatum and from 20 to 52% for those inoculated with G. albida (Fig. 2a–c). Although roots of the large, inoculated C. calothyrsus seedlings had extended beyond 50 cm (Table 2), only sporadic colonization was detected beyond 0 cm, and colonization of crop roots was negligible. The molecular probes indicated that G. albida had not yet extended to 25 cm from the tree, whereas G. etunicatum was present in one trough at 50 cm. The molecular probes also indicated that the mycorrhizal fungus colonization observed at 0 cm on the uninoculated C. calothyrsus seedling in block two was attributable to G. etunicatum (Fig. 2c).

Assessments of tree and crop root samples from the harvest of the second crop in November 2004 showed that both inoculant fungi had colonized tree roots at 25 cm and had spread to the crop roots (Fig. 3a, b). Tree roots had now extended more than 75 cm from the tree, but neither inoculant fungus had established a significant presence on the tree roots at 50 cm, although G. etunicatum was present in three of the root samples. Mycorrhizal fungus colonization was recorded in two uninoculated troughs (Fig. 3c), but the fungal specific primers did not detect either of the inoculant fungi on the roots, indicating that other AM fungi present in the glasshouse may have been responsible.
https://static-content.springer.com/image/art%3A10.1007%2Fs11104-007-9239-z/MediaObjects/11104_2007_9239_Fig3_HTML.gif
Fig. 3

ac Extent of mycorrhizal infection (% root length) determined microscopically, and origin (G. etunicatum, G. albida or other) of mycorrhizal fungus determined by molecular methods, on roots of trees (C. calothyrsus) and crops (Z. mays) growing together in troughs. Samples collected in November 2004 from cropping period C2 (see Fig. 1). Trees were previously inoculated with (a) G. etunicatum, (b) G. albida or (c) not inoculated. Samples were taken at different distances from the tree (0, 25 and 50 cm) in seven replicate troughs. X-axis shows block numbers of samples taken at different distances. Data for trees and crops taken from the same soil cores are presented in adjacent columns. The presence of a small coded section at the top of a bar indicates molecular confirmation of one of the two inoculant fungi

By August 2005, both inoculant fungi had colonized tree roots at 50 cm and to a lesser extent at 75 cm and, when present, had successfully spread to the crop roots at these distances (Fig. 4a, b). Crops had higher mycorrhizal colonization at 75 cm from the trees with G. etunicatum inoculation than with G. albida, and the spread of G. etunicatum from the tree to the crop appears to have been more consistent than that of G. albida (Table 2, Fig. 4a, b), even though differences in tree RLD were not found between the two inoculation treatments at these distances. As it was more than 16 months since the inoculated C. calothyrsus seedlings were transplanted to the troughs, it was perhaps not surprising that mycorrhizal cross-contamination had occurred in several troughs by this time. Of the two inoculant fungi, most cross-contamination was attributable to G. etunicatum. This inoculant was responsible for the contamination of two troughs inoculated with G. albida, whereas only one trough inoculated with G. etunicatum was contaminated by G. albida. At this time, three uninoculated troughs were contaminated by either G. etunicatum or G. albida.
https://static-content.springer.com/image/art%3A10.1007%2Fs11104-007-9239-z/MediaObjects/11104_2007_9239_Fig4_HTML.gif
Fig. 4

ac Extent of mycorrhizal infection (% root length) determined microscopically, and origin (G. etunicatum, G. albida or other) of mycorrhizal fungus determined by molecular methods, on roots of trees (C. calothyrsus) and crops (Z. mays) growing together in troughs. Samples collected in August 2005 from cropping period C3 (see Fig. 1). Trees were previously inoculated with (a) G. etunicatum, (b) G. albida or (c) not inoculated. Samples were taken at different distances from the tree (0, 25, 50 and 75 cm) in six replicate troughs. X-axis shows block numbers of samples taken at different distances. Data for trees and crops taken from the same soil cores are presented in adjacent columns. The presence of a small coded section at the top of a bar indicates molecular confirmation of one of the two inoculant fungi

Rates of spread were calculated for the inoculant fungi from the starting position of the transplanted tree seedling to their positions as detected by molecular probes at different times during the study. For 2004, rates for G. etunicatum were 1.2–2.5 mm day−1 and for G. albida were 1.2 mm day−1. Over the whole experiment, rates of spread were between 1.1 and 1.6 mm day−1.

Over the course of the experiment, the molecular probes consistently differentiated between the two inoculant fungi and appeared to be more sensitive than microscopic assessment in the detection of the inoculant fungi in the roots. For all the root samples, the molecular probes detected the inoculant fungi in 22 samples in which no mycorrhizal fungus colonization was observed under the microscope. On the other hand, mycorrhizal colonization was observed in 19 samples in which the molecular probes failed to detect the inoculant fungi. Given the consistency of these results, the fact that the PCRs were performed in triplicate, and that DNA extracted from spores of the inoculant fungi was used as positive controls, it is most likely that mycorrhizal colonization in the absence of molecular detection indicated the presence of other, contaminating AM fungi.

Discussion

As previously reported (Lesueur et al. 2001), C. calothyrsus responded well to mycorrhizal fungus inoculation using these AM fungal isolates, and growth was poor in the controls despite rhizobial inoculation. Although Gigaspora albida formed more mycorrhizas on C. calothyrsus than Glomus etunicatum, both were similarly effective in promoting tree growth. As roots of the larger inoculated trees had already extended almost the length of the troughs after 30 weeks, the reduction in growth rate of the inoculated trees after 40 weeks is attributed to the trees becoming increasingly pot-bound.

Although mycorrhizal fungus inoculation had clearly stimulated growth of the trees, strong tree-crop competition was detrimental to the growth of the crops in the restricted soil volume of the troughs. Suppressed crop yields and strong below ground competition for nutrients and water are regular features of semi-arid agroforestry systems (Tilander and Ong 1999), although soil volume would not be restricted under field conditions. Following tree shoot pruning in 2005, the increase in crop RLD and shoot growth near the trees compared to 2004, suggests that this treatment successfully reduced tree-crop competition, and supports the work of Namirembe (1999). Contrary to the work by Whitcomb and Stutz (2001), who suggest that shoot pruning reduces tree root biomass and levels of AM colonization, the high levels of colonization found on both tree and crop roots indicate that shoot pruning had not impaired the viability of the AM inoculants. This is encouraging, as it suggests that normal tree management procedures will not damage the activity of AM fungus inoculum in agroforestry systems. However, it is not known whether tree shoot pruning, and the concomitant reduction of C supplied to roots, would slow the spread of AM hyphae in the soil. It is more likely that tree root pruning, also used to control below-ground competition and tillage, which destroys most tree roots in the top 10–15 cm of soil (Rao et al. 2004), will have adverse effects on the spread and transfer of AM fungi to crop plants. Both of these practices are subjects requiring longer-term studies in field plots.

The spread of both inoculant fungi on the tree roots was slower than expected, given that mycorrhizas were established on the tree roots at the time of transplanting to the troughs and could provide an immediate base from which AM hyphae could spread through the soil. Although this study did not involve in situ observations of fungal mycelia, rates of spread were determined by mycorrhizal colonization and presence of the inoculant fungi on the roots. These rates of spread (1–2.5 mm day−1) are low compared to the observations of Jakobsen et al. (1992) and Jansa et al. (2005) (determined by direct hyphal observation and indirectly through measurements of P acquisition respectively) where hyphal growth rates of 1.5–3.2 mm day−1 through soil from which roots were excluded were measured. In this trough experiment, we would expect faster rates of spread as roots were not excluded and colonized roots would have assisted the spread of the fungi. These rates of spread may overestimate that which would occur in the field, as factors such as seasonal stresses, competition from other AM fungi and soil microbes, lower root length densities (Odhiambo et al. 1999; Olsson and Wilhelmsson 2000) and disruption of mycelial networks through hand or machine tillage (Kabir 2005), might slow the spread of inoculants. On the other hand, random dispersal of AM propagules through wind, water, animal or human activity was strictly controlled in the glasshouse. Nevertheless, these results suggest that it may take years before AM fungus inoculants applied to trees will benefit the growth of crops sown several metres from the tree.

Of the two inoculant fungi, G. etunicatum appeared to be the more mobile, as it spread more rapidly through the troughs, established higher levels of colonization on the crops at increasing distance from the tree and was responsible for more cross-contamination of troughs. In contrast, G. albida formed higher levels of colonization on tree and crop roots nearest the tree. These observations support the work of Voets et al. (2006) who reported the contrasting behaviour of developing mycelial networks in Glomeraceae and Gigasporaceae, and their divergent strategies for the exploration and exploitation of new substrates.

This work has also demonstrated that microscopic quantification of colonization and the use of molecular probes to identify specific AM fungi within roots can complement each other effectively. The fungal specific primers we used as molecular probes consistently differentiated between the two inoculant fungi, and showed greater sensitivity in detection of the inoculant fungi in root samples compared with the traditional microscopic methods of assessment. Molecular methods were therefore more sensitive, detecting fungal fragments and enabling positive identification of the fungal isolates, whereas microscopy allows discrimination between functioning and non-functioning mycelia on the basis of mycorrhizal structures observed within the root. However, although the molecular primers were developed to be ‘isolate-specific’, it is possible that they may have amplified sequences from other isolates of the same species or even other species. As sequence length (number of base pairs) should differ between species, we would anticipate successful detection of non-specific amplification but, in the case of conspecific isolates, homologous fragments may be produced, particularly from field samples. We therefore recommend that fragment specificity should be confirmed by sequencing and that further primer development is undertaken to verify and improve the degree of isolate specificity.

This study has shown that trees can potentially act as reservoirs of either inoculated or indigenous AM inoculum, even though rates of spread of the inoculant fungi were slow. The experiment demonstrated the difficulty in promoting mycorrhizal activity on tree roots in order to obtain early mycorrhizal formation on crop plants, while avoiding competition for water and nutrients between tree and crop roots. However, strong competition also occurs under field conditions where soil volumes are not restricted, and further work is needed to develop land management methods which reduce tree-crop competition and promote the activity of mycorrhizal propagules in the soil.

Acknowledgements

This work was partly funded by the European Commission (EU-INCO-DC; Contract no ICA4-CT-2001-10093; SAFSYS project). We wish to thank our project partners: the Scottish Agricultural College (SAC) for the sequencing and development of fungal specific primers, and the Kenya Forestry Research Institute (KEFRI) for providing the rhizobial inoculants.

Copyright information

© Springer Science+Business Media B.V. 2007