DNA Extraction, PCR, Cloning and Sequencing
Genomic DNA was extracted from a maximum of 1.0 g of soil using the MoBio Powermax kit following manufacturer’s instructions. PCR of soil DNA extracts utilized Redtaq (0.25 units/rxn) (Sigma-Aldrich Corp. St. Louis, MO), and primers FGPS989ac (GGGGTCCGTAAGGGTC) (Bosco et al . 1992) and 23SFr3-ten (GGCAWGGGTGACAGGATTTA), the latter a degenerate version of a primer we developed for this study. This primer set targets the non-coding intergenic spacer region (IGS) between the 16S and 23S ribosomal RNA loci. Cycling conditions were: initial denaturation at 94°C for 4:00 min, 35 cycles of 94°C for 1:00 min, annealing for 0:30 secs (55°C for FGPS989ac/23SFr3-ten), 72°C for 1:30 min, and finally 72°C for 5:00 min. Amplicons from soil DNA extracts were cloned into the TOPO TA 4.0 vector with chemically competent Top10 cells following the manufacturer’s instructions (Invitrogen, Carlsbad, CA). Individual colonies (100 per library) were selected at random from each plate for PCR amplification using M13 forward and reverse primers, and sequencing of the insert, which was performed at Functional Biosciences (Madison, WI). We opted to clone our PCR products because the high GC content of the target locus interferes with sequencing. We were able to obtain much cleaner reads using the M13 vector primers. We hypothesize that the cloning and Sanger sequencing approach used here should result in less bias (particularly against high GC targets) than a direct NGS approach, and also provides longer contiguous sequences for analysis than the NGS methods available at the time of our benchwork.
DNA sequences were edited using CodonCode Aligner (CodonCode Corporation, Centerville, MA), and multiple sequence alignments were generated using the Muscle (Edgar, 2004) and Mafft (Katoh et al ., 2002) programs on the CIPRES (Miller et al . 2010) network. Sequence alignments were corrected by eye in Seaview (Gouy et al ., 2010) and trimmed to common length using Bioedit (Hall 1999). Likely chimeric sequences were identified using the program UCHIME (Edgar et al ., 2011) as implemented in the mothur package (Schloss et al ., 2009), and removed prior to subsequent analysis. Sequences derived from our mid-succession control plots were consistently shorter than those from any other clone library. We therefore generated two alignments for analysis: one that spanned the entire rIGS but excluded mid-succession control sequences (811 bp final alignment), and one that was truncated within the rIGS region and that only included sequences from mid-succession libraries (627 bp final alignment). We used the former as the basis for our phylogenetic analysis and comparison of early and late succession libraries and, within the latter, of alder and non-alder soils. The mid-succession-only alignment was used to compare differences among fertilization treatments in our mid-succession sites.