SUPPLEMENTAL
METHODS
Study site. This study was conducted at the Amazon Rainforest Microbial Observatory (ARMO) site, which was established in 2009 to quantify microbial community responses to deforestation in the Brazilian Amazon rainforest. ARMO is located on a cattle ranch, the Fazenda Nova Vida, within RondôniaState, and was chosen as a model site to represent the agricultural development occurring in the Amazon region. Cycles of agricultural conversion and subsequent abandonment have lead to a patchwork of land-use types at the ARMO site, including primary forest and pastures and secondary forests of various ages. In this region, pastures are established by selective logging of timber trees, cutting and burning of the remaining vegetation, and aerial seeding of fast growing African grasses (primarily Urochloa birzantha and Urochloa decumbens), with infrequent burning to control for weeds. No herbicides, tillage or chemical fertilizers are commonly used. When the soil becomes unproductive and is abandoned, pastures are colonized by plant species found in the primary forest, leading to the development of secondary forest, which is commonly of lower plant diversity than the original forest (Pires and Prance 1985).
Within the ARMO sites, semi-permanent plots were established within multiple land use types, including primary forest, pastures of various ages (ranging from 6 to 99 years old) and secondary forests of various ages (ranging from 11 to 17 years old). Classification of land use types and ages of pastures and secondary forests were based on verbal accounts from long-term land managers at the Fazenda Nova Vida. At each site, a nested sampling scheme was established, centered on a 100 m2 quadrat, with 10 m2, 1 m2. 0.1 m2, and 0.01 m2 quadrats nested within, for a total of 12 sampling points per 100 m2 quadrat (Rodrigues et al. 2013).
To examine changes in fungal diversity along this gradient of deforestation, we selected three of the established hectare plots within ARMO: a primary forest site, a pasture established in 1972 and a secondary forest site taken out of pasture in 1998. The soil sampling was conducted in April 2010, the period immediately following the conclusion of the wet season and before the onset of seasonal drought. Soil was sampled to 10cm using PVC cores, and stored onsite at 4°C. Soils were shipped on dry ice, sieved through 2mm mesh, and stored at -80°C.
Molecular analysis of fungal communities. Total soil DNA was extracted from 0.25g of soil using the MoBio PowerSoil PowerLyzer extraction kit (MoBio, Carlsbad CA). DNA was extracted according to the manufacturer’s instructions, but with a modified lysing time of eight minutes using a vortex adapter. Following DNA extraction, all soils were archived at -80°C and DNA was stored at -20°C prior to PCR amplification and sequencing.
To measure the diversity and community composition of soil fungi, we used a novel two-stage PCR approach designed at the University of Oregon to prepare samples for the high-throughput Illumina Hi-Seq200 sequencing platform. The ITS1 region of the internal transcribed spacer region, a universal barcode for fungi (Schoch et al. 2012), was targeted with PCR1 using the fungal-specific primers ITS1F (CTTGGTCATTTAGAGGAAGTAA) and ITS2 (GCTGCGTTCTTCATCGATGC) that had a six-nucleotide barcode and a partial Illumina adapter. The use of combinatorial primers for paired-end Illumina sequencing of amplicons allows for the use of fewer primers while maintaining the diversity of unique identifiers (Gloor et al. 2010). The forward primer sequence was 5’ TCTCGGCATTCCTGCTGAACCGCTCTTCGATCT-XXXXXX-CTTGGTCATTTAGAGGAAGTAA 3’, and the reverse primer was 5’ ACACTCTTTCCCTACACGACGCTCTTCCGATCT-XXXXXX-GCTGCGTTCTTCATCGATGC3’, where XXXXXX represents a unique six-nucleotide barcode sequence to facilitate multiplexing. The ITS1 region was amplified using Phusion High Fidelity Hot Start II polymerase (ThermoScientific) using 1 µl of undiluted template, and a final concentration of 0.4 mM dNTPs, 0.2 µM of each primer, 0.2 mM MgCl2, and 1 unit of polymerase in a 20 µl reaction. The reaction was run on a Eppendorff MasterCycler thermocycler with a 30 second initial denaturation step at 98°C, and 18 cycles of 98°C for 15 seconds, annealing at 65°C for 30 seconds, and extension at 72°C for 30 seconds, with a final extension at 72°C for five minutes. Following PCR1, products were cleaned using the MoBio UltraClean 96-well PCR Cleanup Kit according to the manufacturer’s instructions and eluted in 50 µl EB buffer.
Following amplification of the target gene, the remaining portion of the Illumina-specific sequence was added in a second PCR reaction (PCR2). For PCR2, the forward primer sequence was 5’ AAGCAGAAGACGGCATACGAGATCGGTCTCGGCATTCCTGC 3’ and the reverse sequence was 5’ ATGATACGGCGACCACCGAGATCTACACTCTTTCCCTACACGACG 3’. PCR2 was performed using Phusion High Fidelity Hot Start II polymerase using 10µl of PCR1 template, and a final concentration of 0.4 mM dNTPs, 0.2µM of HPCL purified primer, and 1 unit of polymerase in a 20µl reaction. Thermocycling conditions were identical to those used in PCR1, but the reaction was run for a total of 14 cycles. The entire PCR2 reaction was loaded onto a 1% TAE agarose gel, visualized using a UV transilluminator, and the band was excised and extracted using the MoBio UltraClean GelSpin DNA Extraction Kit according to the manufacturer’s instructions. The final product was eluted into 50µl of EB buffer and quantified using the Invitrogen Qubit fluorometer (Invitrogen, Carlsbad CA). Five ng from each sample were combined, concentrated using the Zymo Clean and Concentrator kit (Zymo Research, Irvine CA) and adjusted to 10nM for Illumina sequencing. The concentration was further verified by QPCR using Illumina-specific primers at the UO Genomics Core Facility (htsep.uoregon.edu).
ITS1 sequences were generated using 150 paired-end reads with the Illumina HiSeq2000 at the University of Oregon. Sequences were trimmed to 100bp to remove low quality ends, and quality filtering was done using the fastx toolkit ( to remove sequences that hada quality score lower than 20 over more than 99% of bases. Quality-filtered sequences were processed using the QIIME platform (Caporaso et al. 2010). Further quality filtering to remove any sequence with an ambiguous base or mismatch to the forward amplicon primer was done during de-multiplexing. Operational taxonomic units (OTUs) were delineated at 97% sequence similarity and chimeras were removed using USEARCH with a minimum cluster size of 2 (Edgar 2010). The OTU table was exported, and all statistical analyses were conducted using the statistical platform R using the packages vegan and ecodist(Goslee and Urban 2007).
Plant root analysis. During soil sieving, plant roots were separated and frozen at -20. Up to eight roots per sample were randomly selected for molecular analysis to identify the plant community. The high plant diversity in tropical forests makes complete sampling difficult, but given that the diameter of the soil cores measured five centimeters, sampling eight roots per core constitutes extensive sampling when scaled up to sampling efforts undertaken at the hectare scale, and represented exhaustive sampling for three of the primary forest soil cores. Plant DNA was extracted and the trnL-UAA intron region of the chloroplast was amplified using the trnL c and d primers described inTaberlet et al. (2007) using the Phire DirectPCR for Plants kit (ThermoScientific). The trnL region has been used as a plant barcode in Amazonia (Gonzalez et al. 2009), and was chosen because it was readily amplifiable across a broad range of plant species (Taberlet et al. 2007) and over 27 000sequences have been deposited in GenBank. While molecular analysis of roots has not been broadly used for plant identification, this method has been applied successfully in temperate ecosystems (Frank et al. 2010). Amplicons were sequenced from the reverse primer using Sanger sequencing withan ABI 3730 DNA Analyzer at the Functional Biosciences laboratory ( Sequences were manually edited to correct erroneous base calls and trimmed to remove poor quality regions using the CodonCode Aligner software. Quality filtered sequences were identified using BLASTn with the GenBank database (NCBI) and assigned to genera based on the top BLAST hit. Acommunity table was constructed by combining sequences from the same genera into a single group. Although this a conservative way to cluster the plant community, plant-fungal interactions may be driven by host traits that are apparent at higher levels of taxonomic classification (e.g.,Gilbert and Webb 2007). Plant genera detected within the three sites are listed in Table S1.
Soil chemistry. Measures of soil chemistry were conducted at the Universidade de São Paulo, Brazil. Analyses were conducted using protocols modified for tropical soils according the methods described in Raij et al. (2001). Each individual soil measure was compared using a one-way ANOVA with Tukey’s post-hoc tests. Mean values and results from ANOVA analysis for all measured soil variables are listed in Table S2.
Statistical analysis. The distance matrix for soil properties was constructed using Hellinger-transformed Euclidean distances. Community similarity for plants and fungi was based on the Bray-Curtis measure. Multiple regression on distance matrices (MRM) was undertaken using the full dataset of soil properties and rarefied plant and fungal community matrices. Although many studies that have used MRM have separated different environmental factors (e.g, Martiny et al. 2011), we were interested in comparing site-level changes in soil properties, particularly given that significant differences were found for all measures included in the analysis (Supplemental Table 2).
References
Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD, Costello EK, et al. (2010) QIIME allows analysis of high-throughput community sequencing data. Nat Meth 7:335–336. doi: 10.1038/nmeth.f.303
Edgar RC (2010) Search and clustering orders of magnitude faster than BLAST. Bioinformatics 26:2460–2461. doi: 10.1093/bioinformatics/btq461
Frank DA, Pontes AW, Maine EM, Caruana J, Raina R, Raina S, et al. (2010) Grassland root communities: species distributions and how they are linked to aboveground abundance. Ecology 91:3201–3209.
Gilbert GS, Webb CO (2007) Phylogenetic signal in plant pathogen-host range. Proceedings of the National Academy of Sciences 104:4979–4983. doi: 10.1073/pnas.0607968104
Gloor GB, Hummelen R, Macklaim JM, Dickson RJ, Fernandes AD, MacPhee R, et al. (2010) Microbiome profiling by Illumina sequencing of combinatorial sequence-tagged PCR products. PLoS ONE 5:e15406. doi: 10.1371/journal.pone.0015406.t002
Gonzalez MA, Baraloto C, Engel J, Mori SA (2009) Identification of Amazonian trees with DNA barcodes. PLoS ONE 4:e7483.
Goslee SC, Urban DL (2007) The ecodist package for dissimilarity-based analysis of ecological data. Journal of Statistical Software 22:1–19.
Martiny JBH, Eisen JA, Penn K, Allison SD, Horner-Devine MC. (2011) Drivers of bacterial beta-diversity depend on spatial scale. Prot Nat Acad Sci USA 108:7850–7854. doi: 10.1073/pnas.1016308108
Pires JM, Prance GT (1985) The vegetation types of the Brazilian Amazon. 109–145.
Rodrigues JLM, Pellizari VH, Mueller R, Baek K, Jesus EC, Paula FS, et al. (2013) Conversion of the Amazon rainforest to agriculture results in biotic homogenization of soil bacterial communities. Prot Nat Acad Sci USA 110:988–993. doi: 10.1073/pnas.1220608110
Schoch CL, Seifert KA, Huhndorf S, Robert V, Spouge JL, Levesque CA, et al. (2012) Nuclear ribosomal internal transcribed spacer (ITS) region as a universal DNA barcode marker for Fungi. Prot Nat Acad Sci USA 109:6241–6246.
Taberlet P, Coissac E, Pompanon F, Gielly L, Miquel C, Valentini A, et al. (2007) Power and limitations of the chloroplast trnL (UAA) intron for plant DNA barcoding. Nucleic Acids Res 35:e14. doi: 10.1093/nar/gkl938
1