Ectomycorrhizal ecology under primary succession on coastal sand dunes: interactions involving Pinus contorta, suilloid fungi and deer

Study site

The study ecosystem is located in the Oregon Dunes National Recreation Area 43°5′ N, 124°1′ W in the Siuslaw National Forest, Oregon, USA. An average of 178 cm of precipitation falls per year, with the majority in winter (Wiedemann et al., 1999). Temperatures average 16°C in the summer months and 4°C in the winter months (Wiedemann et al., 1999). Although seasonally dry in the summer, the lack of precipitation in the summer is mitigated by ocean fog. We identified four sites that each contained three zones: mature forest to the east (approx. 100 yr old; D. Segotta, pers. comm.), centrally located sand dunes that we called isolated areas, and a younger deflation plain forest (approx. 40 yr old; D. Segotta, pers. comm.) neighboring the ocean on the west (Fig. 1) (n = 4 for each zone). The deflation plain forms where wind has blown loose sand off to expose the water table. Wind direction varies by season with north-north-west and south-south-west winds prevailing in the summer and winter, respectively (Wiedemann et al., 1999). The spring and autumn seasons are transitional and winter winds alternate with summer winds. Distances from the forest zones to isolated areas on the sand dunes average 334 m and range from 127 to 598 m. The four sites are along four positions on a north–south axis running along approx. 7 km of the coastline.

The mature forest edge and the deflation plain forest are dominated primarily by Pinus contorta var. contorta. Other EM hosts include Picea sitchensis Bong. Carr., Pseudotsuga menziesii Mirb. Franco, and Tsuga heterophylla Raf. Sarg. Shrub species occurring on the edges of the forests include the EM Salix hookeriana Barr., Arctostaphylos uva-ursi (L.) Spreng., Arctostaphylos columbiana Piper, and the arbuscular mycorrhizal (AM) Cytisus scoparius (L.) Link. By contrast, because of constantly shifting sands, the isolated dune zones either lack vegetation or are sparsely colonized by AM Ammophila arenaria, Lupinus spp. and C. scoparius. Pine seedlings establishing in the isolated areas are predominantly located in moderately stabilized locations partly created by A. arenaria. The soils consist entirely of sand. The mature forest and deflation plain forest have a shallow A-horizon of < 5 cm, except in seasonally flooded areas of the deflation plain forests where the A-horizon was up to 10 cm.

Mycorrhizas on field seedlings

One to three seedlings up to 5 yr old were collected per zone (mature forest edge, isolated areas and deflation plain forest edge) per sample site in three seasons following a randomized complete block design with the four sample sites representing blocks. Seedlings from the isolated areas were at least 120 m from any forest edges. An average of five seedlings were collected per zone in each site. Eighteen seedlings were collected in the mature forest and isolated areas, and 25 were collected in the deflation plain. Seedlings were carefully extracted from the soil, transported to laboratory within 24 h, and kept moist and refrigerated at 4°C until processed. We aged the seedlings by counting the rings on cross-sections of the stems at ground level under ×100 magnification with a compound microscope. Winter and summer seedlings were collected in 2000 and the autumn seedlings were collected in 2002. Sampling was opportunistic because of the limited number of seedlings in the isolated areas.

Within 1 wk, seedling roots were carefully washed of rhizospheric soil using a 0.5 mm mesh screen (No. 35 USA standard testing sieve) and cold tap water. All live root tips of each seedling were removed and sorted into morphological groups (morphotypes) using a dissecting microscope generally following Agerer (1987–2002). Root tips with ambiguous mycorrhizal characters were also sorted and processed, including those with root hairs. Morphotypes were differentiated by characteristics such as color, branching pattern, and distinct features of the extramatrical hyphae and rhizomorphs). No attempt was made to match morphotypes between seedlings. Morphotypes were freeze-dried and stored at −20°C.

Soil bioassay

Soils were collected from each zone in the four sample sites during the peak fruiting season in the autumn of 2002. Soils were brought back to SUNY-ESF and bioassays were set up within 3 wk of collection. Each soil sample consisted of three pooled soil cores taken at random locations within a zone in each site (5 cm diameter × 15 cm deep). The soil samples were mixed with autoclaved peat and vermiculite (1 : 1 : 1). Pinus contorta seeds were surface-sterilized with 30% hydrogen peroxide for 10 min, washed in distilled water for 1 min, and a single seed was planted into Ray Leach tubes (Stuewe & Sons, Inc., Corvallis, OR, USA) filled with a soil mix. Ten seedlings were grown for each soil sample. Three seedlings were grown individually in a 1 : 1 : 1 soil mix with autoclaved field soil, peat and vermiculite as controls. The seedlings were randomized in the racks and grown for 6 months at room temperature with fluorescent grow lights (14 h daylength; 90 µmol m⁻² s⁻¹). The root tips were harvested and processed as described above. Owing to problems with a soil pathogen and insect infestation, many of the seedlings died before harvest. Of the 40 seedlings planted in soil from each zone (4 zones × 10 seedlings per sample site), 25 survived in the mature forest soil (5–7 seedlings per site) 28 survived in the isolated area soil (6–8 seedlings per site), and 26 survived in the deflation plain soil (5–7 seedlings per site).

Deer fecal bioassays

Deer were rarely seen during the study, but tracks and fecal piles were commonly observed. Twenty-one fecal samples of black-tailed deer (Odocoileus hemionus Rafinesque) were collected opportunistically throughout the four sites in the autumn of 2000 during the peak fruiting season. Each fecal pellet in the sample was divided in two by halving individual pellets with a razor blade, and sorting them into two portions. One portion was used immediately as inoculant for a bioassay and the other portion was stored dry at room temperature for 1 yr to assess spore survival. Before use as inoculum, the portions were blended separately with 25 ml of deionized water to make an inoculant slurry. The majority of the spores were smooth and brown and morphologically identified as Suillus (asymmetrical) or Rhizopogon (symmetrical). A few spores of Laccaria, Russula and Cortinarius were observed in some fecal samples. Spore concentrations were quantified with a hemocytometer. We chose seven fresh fecal samples with the highest spore concentrations (portions that contained at least 10⁸ spores) to use for the paired fresh-dry inoculant bioassay. We made a dilution series of fresh fecal–water mixtures containing spore concentrations of 10⁸ to 10² spores ml⁻¹ of water following Lamb and Richards (1978).

Pinus contorta seeds were surface-sterilized with 30% hydrogen peroxide for 10 min, rinsed with sterile water, and planted into individual tubes containing a sterile mixture of sand, peat and vermiculite (1 : 1 : 1). We inoculated seedlings 3 wk after they had germinated, when they had developed their first lateral roots. Three seedlings were inoculated with 10 ml of an inoculation solution for each fecal sample (3 seedlings × 7 spore concentrations × 7 fecal samples = 147 seedlings). Three seedlings were inoculated with autoclaved fecal slurries at full strength and 20 seedlings were inoculated with only deionized water. These last two groups were used as controls. Two weeks later, an additional 10 ml of the same inoculation treatments were added to each seedling. The seedlings were grown and the root tips were harvested and processed as described earlier.

The same procedure was conducted 1 yr later for the dry fecal samples with one change. Because seedlings in the fresh fecal treatment did not become mycorrhizal at low spore concentrations, only the two dilutions with the highest spore concentrations were made for the dry treatment. To investigate possible effects of increased sample size on the diversity of EMF on root tips, eight extra dried fecal samples with high spore concentrations were used to inoculate seedlings.

Molecular identification of fungi

DNA extractions, polymerase chain reaction (PCR) amplifications and restriction fragment length polymorphism (RFLP) generation followed White et al. (1990) and Gardes & Bruns (1996). We used either internal transcribed spacer (ITS)-1f and ITS-4 or ITS-1f and ITS4b as primers to amplify the fungal nuclear rDNA ITS region. DNA from PCR products and RFLPs were run in 3% agarose gels, stained with ethidium bromide, and digitally photographed (EpiChemi II Darkroom; UVP Laboratory Products, Upland, CA, USA). RFLP patterns were generated with HinfI, AluI, and DpnII, and band sizes were measured using the labworks computer software (UVP Laboratory Products, Upland, California, USA). Band sizes were calibrated using a 100 base pair ladder (Promega; Fisher Scientific, Pittsburgh, PA, USA). We compared RFLP patterns from EM root tips with each other and with those from voucher sporocarps of over 100 species collected in the area (T. R. Horton unpublished). A species name was assigned to a mycorrhizal root tip sample when the ITS-RFLP patterns from the root tips matched the sporocarp patterns with all three enzymes when possible (in a few cases, a restriction pattern was not obtained with AluI). Studies from this research site and in Scandinavia suggest matching RFLP patterns with two to three enzymes is a good approximation for species matching (Kårén et al., 1997; Horton, 2002).

We generated sequences for the nuclear rDNA ITS region for the main EMF in the study, including Suillus neoalbidipes, whose sporocarps (but not mycorrhizas) were collected in the study area (Table 1). To determine the identity of unknown sequences, a minimum of 95% sequence overlap to an existing ITS sequence of at least 450 bp in the GenBank database was required. Those samples with 97–100% similarity with existing species were considered a match and were named to the species level. Those sequences with 96% or lower similarity to existing sequences were designated names to the genus, family, order, or in one case, class (putative ericoid ascomycete). Not all unknown patterns sufficiently amplified to allow sequencing. Most of these were rare types and we did not pursue their identity any further. We called unknown patterns by their sample number in the study (e.g. 5152 1f/4, RFLP S186A, RFLP Pat2, etc.).

Data analyses

We used the absolute frequency of each EMF species on the seedlings in each zone for the field and soil bioassay seedlings. For example, in the field seedling survey Rhizopogon fuscorubens A. H. Sm. was found on none of four seedlings in the mature forest site one, and one of six seedlings mature forest site two, two of six seedlings mature forest site three, and zero of three seedlings mature forest site four. The average absolute frequency for this fungus was then ((0/4) + (1/6) + (2/6) + (0/3))/4 = 12.5% for the mature forest seedlings.

We initially calculated the absolute frequency of seedlings with mycorrhizas in the deer fecal bioassays based on the three seedlings per dilution for each fecal sample. The average frequency of seedlings with mycorrhizal roots was 0, 0, 0, 0, 0.33, 0.857 and 0.857 when inoculated with 10², 10³, 10⁴, 10⁵, 10⁶, 10⁷ and 10⁸ spores from fresh fecal samples, respectively. Similar results have been reported elsewhere using spore slurries made from sporocarps (Lamb & Richards, 1978; Castellano et al., 1985). The data from the top three spore concentration treatments (i.e. those that yielded mycorrhizal seedlings) were analysed using a generalized linear model based on a binomial distribution and a logit link (proc genmod in sas version 8, SAS Institute, Cary, NC, USA). The model included variables representing the seven blocks (fecal samples). A statistically significant spore concentration effect was present (likelihood ratio test, 2 df, P < 0.0001), but this effect was clearly attributable to the difference between concentrations 10⁶ and 10⁷ and 10⁶ and 10⁸. Seedlings inoculated with 10²−10⁵ spore concentrations did not produce mycorrhizal seedlings and were therefore eliminated from further analyses. Concentrations of 10⁷ and 10⁸ spores yielded exactly the same proportion of successes over the seven blocks, so the test of whether a concentration of 10⁷ differed from a concentration of 10⁸ was clearly not significant. Because no significant differences were found between the top two spore concentrations, data from these spore concentrations were pooled to make a total of six seedlings for each fecal sample used in paired nonparametric t-tests.

We tested the null hypothesis that there was no difference in the frequency of mycorrhizal seedlings in the paired fresh and dry treatments using the Wilcoxon signed-rank test. The same test was used to test the null hypotheses that there was no difference between the frequency of seedlings colonized by hypogeous or epigeous fungi in the paired fresh and dry fecal treatments. These latter analyses considered only suilloid fungi (Bruns et al., 1998).

We calculated the average Simpson's diversity on field seedlings collected from each zone across the four sites and report species richness for each deer fecal sample. To test for differences in diversity on field seedlings between zones we used a randomized complete block model anova, using the average number of mycorrhizal species occurring on individual seedlings in each zone and the sample sites for blocks. We used Jackknife estimates (Burnham & Overton, 1979) to compare the expected species richness between zones or bioassay treatments. The estimates provided by the jackknife equations are not intended to indicate the true species richness. Rather, they provide a useful tool to compare zones and fecal pellets for their EMF species availability to seedlings. Burnham & Overton (1979) found a minimum mean square error was usually achieved in the first three orders of the jackknife estimator, and so these orders were used here. The most appropriate of the three orders for each part of the study was chosen by testing the null hypothesis that there is no difference between the expected values following Burnham & Overton (1979).

Mycorrhizas on field seedlings

All field seedlings were colonized by at least one EM fungus. We observed 25 and 21 RFLP types on the seedlings harvested from the edge of the mature forest and deflation plain forest, respectively (Table 3). The mature forest and deflation plain forest seedlings averaged 2.2 ± 0.24 (mean ± 1 SE) and 2.39 ± 0.42 RFLP types per seedling, respectively. Many typical EMF were found on the root tips of the seedlings growing in the proximity of mature pines, including Basidiomycetes in Boletaceae, Cortinariaceae, Russulaceae, Thelephoraceae, Tricholomataceae and Ascomycetes such as C. geophilum and Wilcoxina mikolae. Many RFLP types in the forest zones occurred with an average frequency of less than 5%. The relatively high beta diversity of EMF in the forest zones contrasts with the 10 RFLP types found on seedlings from in the isolated areas. Seven of the 10 RFLP types were suilloid fungi belonging to the closely related genera Suillus and Rhizopogon. Seedlings from the isolated areas were colonized by an average of 1.39 ± 0.12 RFLP types per seedling.

Mycorrhizas on bioassay seedlings

A total of 78 seedlings survived in the soil bioassay and no contamination was detected on the control seedlings. The mature forest soil yielded 24 mycorrhizal seedlings and one nonmycorrhizal seedling. The deflation plain soil yielded 26 mycorrhizal seedlings. The isolated area soils yielded 22 mycorrhizal seedlings and five nonmycorrhizal seedlings. Eighteen RFLP types were found in soils collected from the two forests, nine of which were Suillus or Rhizopogon spp. (Table 3). Ten RFLP types were found in the isolated dune soils, six of which were Suillus or Rhizopogon spp. Rhizopogon spp. were more commonly observed in the forest soil bioassays than in the isolated area bioassay.

Ten RFLP types were detected on the fecal bioassay seedlings (Table 3). Eight RFLP types occurred on the fresh treatment seedlings, and seven RFLP types occurred on the dry treatment seedlings. Five RFLP types occurred in both the fresh and dry treatments. Seven of the 10 RFLP types were Suillus or Rhizopogon species (pooled data from fresh and dry treatments). Other EMF were rare types with a frequency of less than 5% and included Thelephora americana Lloyd, another member of the Thelephoraceae, and one unidentified species. The additional eight dry fecal samples yielded one extra species, Inocybe lacera. Letting the fecal pellets dry for 1 yr reduced the total number of mycorrhizal seedlings from 90% in the fresh treatment to 62% in the dry treatment. This difference was marginally significant (df = 6, P = 0.09). The decrease in mycorrhizal seedlings in the dry treatment paralleled a drop in the frequency of occurrence of the epigeous genus Suillus (P < 0.05, n = 7, Fig. 2). At the same time, the difference in performance of hypogeous genus Rhizopogon in the fresh and dry treatments was not significant (P = 0.45, n = 7).

Diversity estimates

Although isolated seedlings tended to have a lower diversity of fungi on them compared with seedlings from the two forest zones, there were no significant differences between blocks (F = 0.38, P = 0.77) or zones (F = 1.47, P = 0.30). Simpson's diversity on seedlings inoculated with deer fecal pellets was similar to the diversity on field seedlings. According to Jackknife estimates, at least 30 RFLP types were available to field seedlings in the proximity of mature and deflation plain forests (Table 4). By contrast, the expected pool of EMF fungi available to the isolated dune seedlings was below 15 RFLP types based on the field or soil bioassay data. Similarly, the expected number of RFLP types available to seedlings inoculated with fecal pellets was also below 15.

Spore dispersal

Dispersal of spores by wind produces a scattered and diffuse spore rain over long distances (Allen et al., 1992). For example, Allen (1987) reported an average of one Thelephora spore was captured in unvegetated areas of the Mt St Helens blast zone per 24 trap hours. Even with a diffuse spore rain, viable spores can accumulate as a resistant spore bank. Despite over 100 epigeous (and presumably wind dispersed) species fruiting in the forest edges, most were not frequently encountered on field or bioassay seedlings. We did not quantify sporocarp production in our sampling sites. However, observational evidence suggests that species of Suillus and Rhizopogon were among the abundant fruiters in the forest zones, particularly S. brevipes, S. umbonatus, R. occidentalis, R. evadens and R. fuscorubens. The jackknife estimates suggest that our root tip sampling was fairly complete in the isolated areas for species richness. Epigeous basidiomycetes such as those found in the isolated areas on seedlings or in soils (Suillus, Thelephora and Laccaria) are thought to be wind dispersed, and were not observed fruiting in the isolated areas during the study. However, the presence of Rhizopogon in the isolated areas points to spore dispersal via animal vectors playing an important role in moving spores to the isolated areas in this case.

Rhizopogon spp. are truffle-like fungi whose sporocarps are produced below ground or erumpent at the soil surface and have lost the ability to propel their spores into the air. While many sporocarps may decay, leaving their spores below ground (Miller et al., 1994), long-distance dispersal is thought to be primarily via mammals (Luoma et al., 2003). Previous work on mycophagy and EMF has focused almost exclusively on hypogeous fungi and small mammals (Luoma et al., 2003). Hypogeous fungi produce odors that attract rodents and marsupials that dig up the sporocarps, consume them and disperse their spores when they defecate (Thiers, 1984; Johnson, 1996; Luoma et al., 2003). Spores in feces are viable and have been shown to successfully yield mycorrhizas on seedlings (Kotter & Farentinos, 1984; Claridge et al., 1992; Colgan & Claridge, 2002). Small mammals on the Oregon dunes ecosystem, particularly the Townsend chipmunk (Tamias townsendii) and the deer mouse (Peromyscus maniculatus), preferentially eat Rhizopogon sporocarps (Ashkannejhad, 2003). However, the distance out to the isolated seedling areas reaches up to 600 m from the established forest and is sparsely vegetated. Limited cover from predators and a lack of food in these areas makes dispersal across them risky for small mammals (T. Manning, pers. comm.). To investigate mammal dispersal from the forests to the isolated areas, we focused on deer.

Deer and elk are known to eat epigeous and hypogeous fungi (Launchbaugh & Urness, 1992), and many epigeous fungi produce odors that likely attract mammals. Mycophagous mule deer and mountain goats were observed traversing primary successional glacial forefronts where small mammals did not venture (Cázares & Trappe, 1994; Jumpponen et al., 1999), and elk were presumed to carry arbuscular mycorrhizal inoculum across the Mt St Helens blast zone (Allen, 1987). In our study, the same fungal species being dispersed by deer were found on the isolated seedlings. To our knowledge, this is the first demonstration that deer disperse viable spores of epigeous, wind-dispersed fungi. Our soil bioassays suggest that suilloid spores are abundant in soils, as others have also shown (Baar et al., 1999; Taylor & Bruns, 1999; Kjøller & Bruns, 2003). Here we provide evidence that spores can survive desiccation for at least 1 year, and that Rhizopogon survives desiccation better than Suillus. While our results suggest the spores can survive as a resistant spore bank in soils, the deer pellets would likely be subjected to repeated wetting and drying on the dunes, eventually disintegrating into the soil matrix. We do not know if spores are capable of withstanding such conditions, but Terwilliger and Pastor (1999) report that 8 wk of flooding did not completely eliminate the ectomycorrhizal potential of forest soils.

Suilloid fungi (Rhizopogon and Suillus spp.) were the principle EMF on pines establishing where mycelial networks were absent. This suggests that EMF are not functionally redundant and that suilloid fungi provide an excellent example of a group with specific ecological adaptations for establishment of pines in harsh or early successional habitats. Whether other species are absent on the isolated seedlings because they are not dispersing to the area or are incapable of surviving the harsh conditions requires further study.