DNA fragments were scanned on an ABI 3730 automated DNA sequencer at Oklahoma AZD0530 manufacturer State University’s Recombinant DNA/Protein Core Facility. The T-RFLP data profiles were obtained and analyzed by using Tanespimycin solubility dmso GeneMapper Software version 4.0 (ABI). Data processing and statistical analysis In 16S-rDNA-T-RFLP profiles, a baseline threshold of 50 relative fluorescence units was used to distinguish ‘true peaks’ from background noise. Considering T-RF drift (improperly sized T-RFs due to differences in fragment migration and purine content), peaks were manually aligned using the method described by Culman et al. [22]. After background removal, raw peak height was normalized to balance the uncontrolled

differences in the amount of DNA between samples by dividing the peak height by the sum of all peak heights of each sample. Culman et al. [22] determined that relative peak heights are better than peak areas for comparisons in T-RFLP profile analysis, yielding greater signal to noise ratios. All the T-RFLP data were arranged into a matrix with each row as a community sample and each column as the relative abundance of each T-RF. The matrix was analyzed by partial Canonical Correspondence Analyses (pCCA) using Canoco for Windows 4.5 (Plant Research International) (32). We performed three kinds of pCCAs: using, as explanatory variables: sites, months, learn more and host species.

For each of these analyses, the other variables (e.g. for the third analysis,

months and sites) were used as covariables. This approach allowed us to isolate the independent effects of each factor. For each analysis, we performed a permutation test of significance with 9,999 permutations, conditioned on the covariables. Based on the complete T-RFLP data matrix, we calculated also the percentage of empty cells in the data matrix [23] as 100% x (total number of cells in the data matrix of T-RFs vs. samples – count of all cells with non-zero values)/(total number of cells in data matrix). Multivariate Analysis of Variance (MANOVA) was conducted using SAS v9.2 (SAS Institute Inc.) and Hierarchical Clustering Analysis was carried out with R (R development core team, 2003). The average proportion per SPTLC1 existence (APE) of all T-RFs found in five host species estimated the prevalence of T-RFs in diverse communities. APE is defined as the average proportion of one T-RF over those host samples which contain this T-RF in their T-RFLP profiles, and was calculated by the sum of the relative proportions divided by the number of the samples containing this T-RF, as in the following formula: where Pi is the relative proportion of the T-RF in ith sample, m is the total number of samples, and n is the number of these which have the T-RF. Results Mono-digestion T-RFLP In this study, we used T-RFLP profiles to study the features of the distribution of leaf endophytic bacterial communities.