OUP user menu

Temporal variation in the arbuscular mycorrhizal communities colonising seedlings in a tropical forest

Rebecca Husband, Edward Allen Herre, J.Peter W. Young
DOI: http://dx.doi.org/10.1111/j.1574-6941.2002.tb01002.x 131-136 First published online: 1 October 2002


In order to investigate temporal variation in the arbuscular mycorrhizal (AM) fungal community in a tropical forest in the Republic of Panama, seedlings of the canopy emergent Tetragastris panamensis were sampled three times over a period of 3 years. We used AM-specific primers to amplify and clone partial small subunit ribosomal RNA gene sequences. Over 550 clones were classified into 18 AM fungal types. As the seedlings matured, the fungal diversity decreased and there was a significant shift so that previously rare types replaced formerly dominant fungal types. Further, seedlings of different ages sampled at the same time point were colonised by significantly different fungal populations. Our results indicate that both time and host age may influence the mycorrhizal population.

  • Arbuscular mycorrhiza
  • Small subunit
  • AM1
  • Diversity
  • Succession
  • Seedling recruitment

1 Introduction

Arbuscular mycorrhizal (AM) associations are widespread in terrestrial ecosystems and represent the natural state for the majority of plant species. The AM fungi (order Glomales) receive carbon compounds from the host plant and in return can confer a number of benefits. The primary benefit to plants is an increased uptake of immobile nutrients, especially phosphorus. Other benefits include improved water relations and protection against pathogens [1]. Recent studies have shown that individual AM fungi may differ in their effects on specific host plants [27] and consequently affect the diversity and productivity of artificial plant communities [8]. However, this has yet to be demonstrated in the field.

Many community studies are constrained by an inability to identify which AM fungi are in active association with roots, which severely limits the conclusions that can be drawn. The main hindrance is the AM fungi themselves. They are obligate symbionts that can only be cultured in the presence of their host, and the identification of AM fungi either by spore characteristics or by morphology in planta is extremely difficult. The majority of our knowledge about AM fungi in the environment is based upon spore counts, however, the population of spores in the soil may not accurately reflect the species composition in roots [10].

Recent advances in molecular techniques now make it possible to directly identify the AM fungi colonising roots in the field. Helgason et al. [11] published one of the first studies that utilised the polymerase chain reaction (PCR) to analyse the diversity of AM fungi colonising roots in the field. Using the sequence data available at that time, the AM1 primer was designed to amplify all known glomalean small subunit (SSU) rDNA sequences whilst excluding plant DNA. Cloning and PCR-RFLP (restriction fragment length polymorphism) separate the AM fungi into classes and sequencing of examples of each class give an insight into their identity. If it is assumed each fungal type is amplified and cloned proportionally, then the numbers of each class can be perceived as an approximate estimate of their proportion in the root. In this manner it has been shown that there are significant differences between the fungal communities colonising hosts in a semi-natural woodland and adjacent arable sites [11]. Seasonal variation in the pattern of AM fungal colonisation has been demonstrated [12,13] and non-random associations with host species or dominant vegetation type have also been detected [12,14].

The sequence types recognised by these molecular methods cannot be equated directly with the formal species that are identified on the basis of spore morphology [14], but many of the types have been recovered repeatedly in different studies and appear to represent entities as widespread and stable as those defined by morphology. Whether the classification is molecular or morphological, there are likely to be important functional differences between the types, but there may also be significant hidden variation within them. Nevertheless, the molecular typing allows us to address questions of AM community composition and distribution in plant roots, which is the ecologically significant niche and is inaccessible to methods based on spore morphology.

The aim of this project is to characterise the molecular diversity of AM fungi colonising roots in a tropical forest on Barro Colorado Island (BCI), Republic of Panama. Since we had no prior knowledge of which AM fungi might be present or functionally important, we decided to target as many glomalean fungi as possible by using the AM1 primer. New sequence data have revealed that the AM1 primer site is not well conserved in certain divergent lineages of the AM fungi, the Archaeosporaceae and the Paraglomaceae [15]. Nevertheless, AM1 remains the most broadly applicable primer suitable for field studies, it reliably detects the three traditional AM families (Glomaceae, Acaulosporaceae, Gigasporaceae), and has been used in numerous studies that provide useful comparisons.

This strategy was successfully used in a preliminary investigation of the mycorrhiza colonising different hosts, sites and time points on BCI [30]. The overall AM fungal diversity was found to be high, there were large differences in the AM fungal population colonising different sites, and non-random associations with different hosts. Highly significant temporal shifts in the AM fungal community were also detected, but the design did not allow the possible effects of plant age to be separated from secular shifts in community composition. Thus the aim of this study is to explore and potentially to separate these two factors. Seedlings of Tetragastris panamensis (Engl.) Kuntze spanning different ages were sampled from a single site over a period of 3 years.

2 Materials and methods

2.1 Sampling

All samples were collected from the Miller site on BCI, a field station operated by the Smithsonian Tropical Research Institute in the Republic of Panama (9°10′N, 79°51′W), mean annual precipitation 2.6 m. Root samples were taken from six seedlings of T. panamensis in October 1998 and November 1999, and eight in November 2000. Seedlings sampled in 1998 were approximately 3 months old and still had their cotyledons attached. No new seedlings germinated the following year, therefore the 1-year-old seedlings sampled in 1999 survived from the 1998 cohort. Again in 2000 there were no new seedlings. Five seedlings from the 1998 cohort were sampled, plus three seedlings from an obviously older cohort (later referred to as 2-year-old and 5-year-old seedlings respectively). At each sampling the entire seedling was harvested and all fine roots used for analysis.

2.2 Molecular analysis of roots

DNA was extracted from plant roots using the potassium ethyl xanthate method [16]. DNA extracts were further purified using Concert™ spin columns (Gibco BRL). Partial ribosomal SSU DNA fragments were amplified using Taq DNA polymerase (Gibco BRL), a universal eukaryotic primer NS31 [17] and a primer AM1, designed to amplify fungal but not plant sequences [11]. Reactions were performed using 0.2 mM dNTPs, 10pmols of each primer and the supplied reaction buffer to a final volume of 50μl (PCR conditions: 94°C for 3 min, 58°C for 1 min, 72°C for 1 min 30 s, then 24 cycles at 94°C for 30 s, 58°C for 1 min, 72°C for 1 min 30 s). Products were cloned into pGEM-T Easy (Promega) and transformed into Escherichia coli (DH5α). Putative positive transformants were screened using a second NS31/AM1 amplification (PCR conditions: 24 cycles at 94°C for 30 s, 58°C for 1 min, 72°C for 1 min 30 s). At least 20 positive clones from each sample were tested for RFLP by digestion with HinfI and Hsp92II, according to the manufacturer's instructions (Promega). Examples of each RFLP type were reamplified using SP6 and T7, purified using Concert™ spin columns, and sequenced on an ABI 377 automated sequencer using the Dye terminator cycle sequencing kit with AmpliTaqFS DNA polymerase (Applied Biosystems). The sequencing primers were SP6 and T7.

2.3 Data analysis

Forward and reverse sequences were aligned and edited using Lasergene SeqMan (DNAStar Inc.). Clustal-X [18] was used for multiple alignment and neighbor-joining phylogenetic analyses [19], using Corallochytrium limacisporum, a putative choanozoan [20] as the outgroup. Diversity indices were computed using EstimatesS (Version 5, R.K. Colwell, http://viceroy.eeb.uconn.edu/estimates). Rank abundance plots for the three largest samples were corrected for sample size bias by taking 10 random subsamples of 65 clones each.

3 Results

3.1 Phylogeny of the AM fungal types

In all, 558 clones were classified into 18 AM fungal types according to their HinfI and Hsp92II RFLP patterns (Table 1). A phylogeny constructed from representative sequences of each fungal type is shown in Fig. 1. The overall topology is in agreement with previously published studies, with the sequences clearly resolving into the three traditional families [21,13]. Of the 18 AM fungal types, 16 belong to the Glomaceae and one each to the Acaulosporaceae and Gigasporaceae respectively. The AM fungal types Glo 1b and Glo 8 include sequences with over 99% identity to the sequences for Glomus mosseae (Z14007), and Glomus intraradices (X58725) respectively. The remaining types do not cluster closely with any sequence from AM fungi in culture.

View this table:

Number of clones of each RFLP type detected colonising Tetragastris seedlings at each time point

19983 months75524513319121321222126.81.3
19991 year2276186155411121171515.21.2
20002 years4819142955281303.80.8
20005 years532112426542.6
  • The mean and standard deviation of the number of AM fungal types detected per individual seedling is indicated in the last two columns.


Neighbor-joining phylogenetic tree showing examples (in bold) of the AM fungal types colonising Tetragastris seedlings at a single site on BCI. Bootstrap percentages greater than 70% are shown. Group identifiers (for example Glo8) relate to the classes defined in Helgason et al. [13] and Husband et al. [30]. Asterisks (*) identify new sequences from this study submitted to GenBank under accession numbers AY129634-AY129643. Scale bar represents 0.1 substitutions/site.

3.2 Temporal variation in AM abundance

Over the 3 years, there is a consistent trend whereby the dominant AM fungal type 1 year is replaced by a previously rare type the following year (Table 1,Fig. 2). In the newly emergent seedlings sampled in 1998, Glo 1b and Glo 8 were both co-dominants accounting for 26% and 24% of the clones detected respectively. In the surviving seedlings sampled the following year, only a single Glo 1b clone was detected and Glo 8 was not detected at all. Instead seedlings from the second time point were heavily dominated by the Glo 30 type (50%) with moderate colonisation by Glo 18 (14%). Finally, at the third time point Glo 18 became the most abundant AM type, accounting for 37% of the clones detected in the 2-year-old seedlings and completely dominating the 5-year-old seedlings (82%). These changes over time are significant (χ2=522, 27 df, P<0.001). Furthermore the AM fungal populations colonising the two different age classes of Tetragastris sampled at the same time are significantly different (χ2=43, 6 df, P<0.001). These results indicate that both time and host development influence the AM fungal population.


Proportion of the five most abundant AM fungal types colonising Tetragastris seedlings at each sampling period (white=Glo1b, diagonal striped=Glo8, horizontal striped=Glo30, grid=Glo35, black=Glo18).

3.3 Temporal variation in AM diversity

A rank abundance plot reveals how the mycorrhizal diversity decreases with time and seedling age (Fig. 3). The newly emergent seedlings have the most even distribution of AM fungal types, whilst the 5-year-old seedlings are almost exclusively dominated (82%) by a single type. Although the 1- and 2-year-old seedlings show an intermediate distribution, the 1-year-old seedlings are unusual because at the same time as being heavily dominated by a single AM fungal type, they have the highest ‘species’ richness of any of the samples.


Rank abundance curves for each sampling point. Plots are corrected for sampling bias by taking 10 random subsamples of 65 clones each (solid black=′98/3mo, short dashed=′99/1yr, solid grey=′00/2yr, long dashed=′00/5yr). The figure is discussed in the text.

These distributions are reflected in the diversity indices. Again the 1- and 2-year-old seedlings are intermediate between the highly diverse new emergents and the low-diversity 5-year-old seedlings. However, the high ‘species’ richness (R) of the 1-year-old seedlings and the atypical sparseness of rare types in the 2-year-old seedlings results in them having similar Shannon diversity indices H: ′98/3mo H=1.96, R=13; ′99/1yr H=1.75, R=15; ′00/2yr H=1.71, R=8; ′00/5yr H=0.79, R=7. To test for any bias due to uneven sample size we recalculated the diversity indices for the three largest samples by taking 1000 random subsamples of 65 clones each. The unbiased indices confirm an overall decline in diversity: ′98/3mo H=1.90, R=10; ′99/1yr H=1.68, R=11; ′00/2yr H=1.78, R=8; ′00/5yr H=0.79, R=7. In addition to a decline in diversity, the average number of AM fungal types colonising each root significantly decreases with time and seedling age (Table 1, F=4.942, df=3, P=0.13).

4 Discussion

These data show that the diversity of mycorrhiza colonising seedlings in a tropical forest is high compared to a temperate woodland or temperate arable sites (Shannon diversity index=1.44 and 0.398 respectively) [11]. If it is assumed there is a link between above- and below-ground community structure, the high AM fungal diversity may logically be expected in a tropical forest; however, the almost total dominance of Glomus types is not. In contrast, the AM fungal community of a temperate woodland was found to have an even distribution between the families, with Acaulopsora and Scutellospora types being much more abundant than on BCI [12]. Observing an overwhelming dominance of Glomus types colonising arable crops, Daniell et al. [13] hypothesised that Glomus species are better adapted to disturbed environments where their high sporulation rates may enable them to recover more readily. Yet tropical forests are deemed very stable environments, in general, so sporulation rates should be less critical. A possible factor could be tropical small mammals, which are known to consume and hence disperse AM fungal spores [22]. In an investigation of mycophagy on islands surrounding BCI it was demonstrated that the spores consumed were all Glomus species [23]. This enhanced dispersal of Glomus species may contribute to their dominance within roots.

There is much evidence for temporal variation in mycorrhizal populations. However, most of this information comes from within-year sampling of temperate annual hosts or spore populations, therefore such changes are usually attributed to seasonal variation or host phenology [2426,13]. However, in this study across years, we have demonstrated that the AM community colonising a single cohort of seedlings not only changes over time, but undergoes repeated replacement of dominant types by previously rare types. In a wider study [30] we detected the same pattern of replacement at two different sites on BCI. Likewise the same initially dominant type (Glo 1b) was replaced by the same previously rare type (Glo 30). The data from the present study are all the more striking because we have information for a third year where again the dominant types were replaced by previously rare types. Furthermore, although the sample size is small, we have detected a significant difference in the AM community colonising two different cohorts sampled at the same time point. In addition, we have shown that the number of AM types colonising each root, as well as the overall fungal diversity, decreases with time and seedling age.

One possible explanation for these data would be that of AM fungal succession. Yet, whilst accounts of ectomycorrhizal succession are common [1], there are few reports of succession in arbuscular mycorrhizas [27]. However, there is some experimental evidence for arbuscular mycorrhizas having different life history strategies that could drive fungal succession (see Hart et al. [28] for review). Thus it is possible that the newly emergent Tetragastris seedlings were heavily colonised by AM fungal types with superior colonising abilities, and these were then slowly replaced in the older seedlings by AM fungal types with superior persistence abilities.

Alternatively, it is increasingly recognised that individual AM fungi have differential effects upon different host species [25,7]. If these differential effects can affect seedling recruitment then the most effective host–fungal combinations would have a higher probability of survival, and subsequently would become enriched in the surviving population. For example, if newly emergent seedlings were more likely to survive when colonised by Glo 30 than by Glo 1b, then the surviving seedlings would be expected to be dominated by Glo 30. Kiers et al. [6] demonstrated that tropical mycorrhizas might have differential effects on seedling growth. Further, their results implied that seedlings inoculated with AM fungi derived from the roots of conspecific adults exhibited reduced growth compared to seedlings inoculated with AM fungi derived from heterospecific adults. If this truly is a mycorrhizal affect, it would indicate that seedlings and adults gain maximum benefit from associations with different AM fungi. The implications for forest ecosystem functioning are great because seeds dispersed farther away from their parent would be more likely to find mycorrhiza beneficial to survival. Hence AM fungi may influence the negative density dependence that is known to enhance diversity in tropical tree communities [29]. The logical consequence of this premise would be that the community composition of AM fungi would change as seedlings mature.

Until we know how general these patterns are in the rest of the forest, and gain a functional knowledge of the interactions, it is very difficult to reach any firm conclusions. Even so, we have successfully demonstrated a persistent pattern of replacement in the mycorrhizal community over time and have shown that host developmental stage is a significant factor. Our study indicates that the mycorrhizal population on BCI is characterised by a far greater complexity than previously thought, and highlights the importance of further study of the mycorrhizal communities in planta in the field.


The authors would like to thank Rachel Gallery and Tanja Roehrich for help collecting samples; Ahn-Heum Eom and Scott Mangan for their comments; Allen Mould for assistance with the sequencing and the Andrew W. Mellon Foundation for funding the project.


  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
  28. [28]
  29. [29]
  30. [30]
View Abstract