Methods
Plant species selection
Plant species were selected for pair-wise comparisons based on phylogenetic relatedness and structural characteristics (e.g. lignin and thickness) and chemical composition, (nitrogen and phosphorus concentrations). Croton poecilanthus Urban and Sapium laurocerasus Desf. both belong to the Euphorbiaceae, but they differ in physical and chemical characteristics. Croton poecilanthus and Manilkara bidentata (A.DC.) Chev. leaves have similar physical and chemical traits including the presence of latex, though the latter belongs to a different family (Sapotaceae). Inga vera Willd. and I. fagifolia (L.) Willd. were selected to represent two closely related species in the Leguminoseae with high lignin and nitrogen concentrations, but differing slightly in concentrations of secondary plant compounds.
Recently fallen leaves of five tree species were collected at three sites in a non-seasonal, late secondary, subtropical wet forest. The sites were El Verde, EV (3500 mm ppt y-1; 350 masl), Sabana, Rte. 988 (3000-4000 mm ppt y-1; 250 masl) and Bisley (3000 mm ppt y-1; 150-250 masl); all sites are in the Caribbean National Forest, Puerto Rico. Leaves of M. bidentata and C. poecilanthus were all collected from EV, I. fagifolia and I. vera leaves were collected at EV and Sabana, and S. poecilanthus leaves were collected at EV and Bisley. Leaves of each species were taken from at least five trees that were growing at least 20 m apart at each site.
Decomposed leaves were collected and air dried for
three hours before microfungal species were isolated using the particle
filtration method as described by Bills and Polishook (1994, 1996) and Polishook
et al. (1996), and modified by Bills (2000). The decomposed leaves were placed
in a sterile blender and pulverized at high speed for one minute. Pulverized
samples were washed with a stream of sterile distilled water (10 min) through a
2 mm brass pre-screen, and two sterilized polypropylene mesh filters (210 and
105 µm) to remove spores. The particles trapped on the 105 µm mesh filter were
placed in a 50 ml sterile polystyrene centrifuge tube and 50 ml of sterile
distilled water was added to re-suspend them. After letting the particles
settle, the water was decanted to remove more spores and any remaining water
was extracted with a sterile pipette. Particles were washed again with a second
volume of water to obtain a 20:1 dilution; this procedure was repeated again
twice. Washed particles were diluted to 1:100, 1:5000 and 1:2,000,000 using a
serial dilution method. From each serial dilution, 0.1 ml was plated and evenly
distributed on the agar surface in 90 mm petri dishes using a flamed,
bent-glass rod (10 plates each of Malt-Cyclosporin Agar and Bandoni’s Medium,
as described below). This procedure was carried out in a sterile hood. The plates were incubated at room temperature (ca. 22 C)
with a 12 h photoperiod.
Two types of culture media were used for the initial
dilution plates, Malt-Cycloporin Agar (Malt Yeast Agar, as described below,
with 10 mg of Cyclosporin A added when the medium was cool; Polishook et al.
1996) and Bandoni's Medium (4g of L-sorbose, 0.5 g yeast extract and 20 agar L-1
distilled water). Fifty mg L-1 of Chlortetracycline and Streptomycin Sulfate were also added to
the Malt-Cyclosporin Agar and Bandoni's Medium when the agar was cool to
prevent bacterial growth. Fungi growing from the particles were transferred to
90 mm petri dishes with Malt Yeast Agar (MYA: 10 g of malt, 2g yeast extract
and 20 g agar per L distilled water). Oatmeal (OMA), Corn Meal (CMA), Malt (MA)
and Potato Dextrose Agar (PDA) were made according to Rossman et al. (1998),
and were used for separation and identification of strains.
All fungi growing from particles were isolated. Growing fungi were transferred to duplicate slants of MYA and morphologically similar species (morphospecies) were sorted after one month of growth. Sub-samples of similar species and species that did not sporulate were transferred to slants of OMA, CMA, PDA and MEA to verify that the morphospecies were correctly classified. Autoclaved banana leaves, and in some cases, autoclaved leaves of the species from which the fungi were originally isolated, were added to PDA and MEA media to promote fungal sporulation. These cultures were allowed to grow for another month and sorted. After two months, morphologically similar species were sorted again and their frequencies were recorded.
Determination
of maximum likehood of fungal diversity to adjust methods
A trial experiment to isolate fungal species from S. laurocerasus leaves was set up in
order to determine the dilution that yielded the maximum diversity of fungal
species and would therefore allow the best possible representation of the
fungal community. Sapium laurocerasus
was chosen for the trial because it is known to have high litter quality and be
easily decomposable (Harmon 2000), and can
therefore be expected to support an initially
high diversity of fungi.
Freshly fallen senescent S. laurocerasus leaves were collected at
El Verde, surface dried with paper towel, placed in litter bags, returned to
the same site, and decomposed for five weeks (May 13 to June 16, 2000). The particle filtration method described
above in section 2.3. was used except that only 3.75 g of air dry leaves were
available instead of the recommended five grams. The maximum number of strains
isolated using each medium was considered the maximum expected number of species
that we could potentially isolate. Previous research on two litter species, Manilkara bidentata and Guarea guidonea (L.) also provide
information on microfungal diversity in leaf litter species at El Verde, but
the diversity was higher in that study because litter at various stages of
decomposition was used (Polishook et al. 1996).
Determination of the dominant fungi in
each leaf species
In order to determine the ten vegetatively dominant fungal species among the early stage decomposers in each leaf litter species, a five-week decomposition experiment was established underneath five representative trees per plant species. Five grams of senescent freshly fallen leaves of each species were placed in separate 20 cm2 mesh bags. Five bags per species were placed under each of the five tree replicates of the same species (25 bags per tree species). Leaf litterbags were collected after five weeks of decomposition to isolate early stage fungal decomposers. One litterbag per tree replicate was randomly selected for fungal isolation among the original five and these were pooled for isolation of fungi. Isolations were made from a 5 g subsample from the pooled litter sample. The random selection of one bag from each of the five tree locations allowed us to adjust for heterogeneity within tree species.
Fungi were isolated using the particle filtration method and sorted into morphospecies as described above in section 2.3, except that only five of the ten plates were randomly selected from the 1:100 dilution as the source of the culture isolates. A total of 182-294 cultures were obtained from each of the leaf species (Table 1). Classification of the morphotypes was reviewed at two and four months, and the total number of isolates per morphospecies was recorded. The most common species were identified to genus when possible.
Selection
of fungal isolates for the microcosm experiment
For each of the leaf species, the
fungal morphotypes were arranged by their rank abundance based on frequencies
of occurrence, and selections were made from the five most frequent
morphotypes. In general, two dominant fungal species were selected from each
leaf species that did not occur among the dominants in the other leaf litter
species. While the frequency data from this study (Table 1) confirm previous
findings of differences in fungal community composition and dominance among
litter species (Holler and Cowley 1970, Cornejo et al. 1994, Polishook et al.
1996), there was some overlap in species composition among dominant fungi from
the five leaf types. Fusarium solani
the most frequent fungus from S.
laurocerasus was also frequent in I.
fagifolia. Fusarium was selected
to represent S. laurocerasus since
two other dominants (Trichoderma sp.
and Penicillium) were eliminated
based on their ability to produce airborne spores which makes them dangerous to health and difficult to control in terms
of cross-contamination (Table 1). Similarly,
Volutella sp. was the second most
frequent isolate from C. poecilanthus
leaves, but it also occurred among the top five dominants in S. lauracerasus (pre-trial), and also
among the five most frequent species in M.
bidentata (Table 1). The most frequent isolate from M. bidentata in a previous study (Polishook et al. 1996) as well
as in this study, Pestalotia sp., was
used although it was also among the five most frequent species in C. poecilanthus leaves. Pestalotia species are common but
difficult to identify, and the species from M.
bidentata and C. poecilanthus may
or may not be the same. Plant species were assigned a
code corresponding to the first letter of genus and species. Fungal species
were assigned the plant code from their source leaf followed by the number 1 or
2 according to their relative ranks. For example, the two isolates from M. bidentata were coded as MB1 and MB2.
The selected fungal isolates and the other eight most frequent isolates from
each leaf litter species appear in Table 1.
Microcosm design
Translucent
plastic containers (19 cm x 12 cm and 8.5 cm deep; 1892 ml volume, Glassware™)
were modified for use as microcosms. A ventilation tube made from a 7 cm length
of plastic pipette (1 ml) allowed gas exchange. The ventilation tube was sealed
to the microcosm with hot glue, and one cm extended inside the container. The
ends of the ventilation tubes were covered with two layers of parafin film
(Parafilm) to prevent entrance of fungal spores and mites. A wire mesh platform
(1 cm mesh) was inserted to form a shelf 4 cm from the bottom of each
microcosm. An autoclaved piece of cloth was placed over the wire platform to
prevent litter from falling through the mesh. A gap in the wire mesh in one
corner allowed insertion of vials containing a solution to trap CO2 .
The microcosms were placed in a laminar flow hood and misted with 70 ml
of sterile deionized water once per week.
Leaf sterilization
Senescent freshly fallen leaves were collected and sterilized with gamma irradiation (3640 rads of Cs137). Culturing from those leaves showed that viable propagules had survived the first sterilization. A second, stronger irradiation was therefore performed a month later (500 rads/min Cs137 exposure time 8.18 min total radiation 4090 rads). No colonies grew from the subsamples that were cultured after the second irradiation.
Leaf inoculation
Five grams
of leaf litter of a single plant species were placed in each microcosm. Leaves
in each microcosm were inoculated with only one fungal species. The five
leaf species were inoculated with two dominant decomposer fungi from each of
the five leaf species in all possible combinations. There were three replicates
for each of the 50 treatment combinations for a total of 150 microcosms.
Sterile leaves were placed in the microcosms, moistened with sterile distilled
water, and inoculated with 10 agar plugs (5 mm dia)
from clean cultures. In addition, 1 ml of liquid medium containing the fungus
was added to the leaves to ensure rapid fungal colonization.
The
numbers for the 150 microcosms were assigned randomly to randomize the
placement of treatment combinations during incubation. When measurements of CO2
were made, however, the microcosms were divided into two groups to allow better
management and to insure that treatment combinations that were critical for
testing the original hypotheses were measured on the same day. Group I
contained S. laurocerasus, C.
poecilanthus and M. bidentata for
phylogenetic relatedness versus litter quality contrasts. Group II was
inoculated on the following day, and was comprised of the two Inga species.
Re-isolation of fungi
Subsamples of litter from each microcosm were cultured on MEA to determine the fungi present at the end of the experiment. Some microcosms containing leaves of Croton poecilanthus were found to have agaric basidiomycetes with ligninolytic enzymes in addition to the microfungal isolate in the treatment. One microcosm had a Marasmius sp., while the others with basidiomycetes had Melanotus eccentricus (Murrill) Sing. This apparently resulted from incomplete gamma sterilization of Croton leaves. Comparison of weight loss in replicates with and without basidiomycetes indicated significant differences in weight loss (t-tests). Of the 150 microcosms, 16 were found to have microfungal contaminants at the end of the experiment. (Appendix Table 1). These were probably contaminated late in the experiment, and values for weight loss and CO2 evolution did not differ between contaminated and uncontaminated replicates (t-tests). Data from microcosms with microfungal contaminants were therefore treated according to the fungi in the original treatment, while those with delignifying basidiomycetes were treated separately in statistical analyses.
CO2 sampling
Traps that contained 15 ml 1 M NaOH were placed in the microcosms, which were then closed tightly for 24 h to measure the rates of CO2 evolution. Samples were collected during wks 1, 2, 3,4, 8 and 17. Group II was processed one day after Group I on each sample date so that the number of days of incubation was the same for both groups. After each collection, a 5 ml subsample from the NaOH-CO2 trap was combined with 2.5 ml of BaCl2 (1M) and 2 drops of Phenolphthalein. The subsample was then titrated with HCl (1N). The volume of HCl consumed was used to calculate the amount of CO2 in the trap solution. Mean daily CO2 evolution on each of the four sample dates was calculated from the replicates for each plant-fungus combination. Mean daily CO2 evolution g-1 air dried litter across sample dates was also calculated.