PMC:5007713 / 7253-13806 JSONTXT

{"project":"2_test","denotations":[{"id":"27605333-16018312-69475450","span":{"begin":3205,"end":3207},"obj":"16018312"}],"text":"2. Results\nThe independent single-dye hybridization experiments (two labeled RNA samples hybridized at four different sample amounts to eight different subarrays) resulted in highly comparable signal intensities between the two corresponding samples. Although the strong signals could even be detected using a 125-fold diluted labeled RNA sample, the weaker signals were no longer recognizable. Therefore, probe signals obtained from the pair of hybridization experiments using undiluted labeled RNA were used for subsequent analysis (labeled as “Myk1” and “Myk2” in the NCBI Gene Expression Omnibus (GEO) accession GSE28018, respectively; see also experimental Section 4.6.). Based on a histogram of ascending signal intensities of the various probe groups (Figure 1), it was possible to distinguish the majority of unspecific signals close to the background value from the few specific probe signals. A threshold value of twice above the background (normalized values above 24 in the Cy3 channel) was selected to separate specific probe signals of enzyme transcripts from unspecific hybridization signals. A similar signal intensity profile was observed for precursor rRNA probes of Agaricales and Zygomycota, but not for Eurotiales probes. The majority of probe signals of the Eurotiales group exhibited significantly higher unspecific hybridization signals, which are most likely due to the comparatively smaller genetic distances among Eurotiales probe sequences compared to those among the Agaricales or Zygomycota. Therefore, a threshold value of four-fold above the background value (normalized values above 48 in the Cy3 channel) was applied to all ITS probe signals (including the Agaricales and Zygomycota probes). Correlation analysis of specific probe signals (above the threshold) revealed that signal intensities could be reproduced using undiluted labeled samples with an R2 value of 0.94 for precursor rRNAs.\nFigure 1 Signal intensity histogram of precursor-rRNA probes (A) and transcripts of genes encoding enzymes analyzed in this study (B); probes were grouped according to their original taxonomic assignment or according to their (putative) enzymatic activity and ordered by ascending signal intensity (taken from the Cy3 channel of the “Myk1” hybridization sample; see GEO accession GSE28018). Grey and red bars indicate probe signals below and above the threshold values of each target group, respectively.\n\n2.1. Identification of Precursor rRNA Transcripts as Phylogenetic Markers for Fungal Soil Communities\nDetailed post-experimental evaluation of ITS probe sequences exhibiting above-threshold hybridization signals (Figure 1A) revealed that only a fraction of these probe sequences really targeted members of the Agaricales, Eurotiales, or Zygomycota, respectively, indicating a serious misassignment of taxonomic information in public sequence databases (Table 2). Of the 48 strong ITS signals, only 13 signals indicated the presence of Agaricales (five signals) and Eurotiales (six signals). Two additional sequences indicated the presence of Paecilomyces (Eurotiales). However, these were similar to congeneric taxa that have recently been transferred to the Hypocreales [16]. The remaining 35 strong signals predominantly indicated the presence of Sordariomycetes, represented by Hypocreales (14 signals), Sordariales (two) and Xylariales (two), and the genus Arthrinum.\nmicroarrays-01-00025-t002_Table 2 Table 2 ITS rRNA probes with signal intensities above the threshold. The signal intensities in the two experiments (Myk1 and Myk2, cf. GEO GSE28018) are given for each probe. The fourth column indicates the taxonomic group actually targeted by the probe, as revealed by a BLAST search. The number of sequences among the perfect matches, deposited under matching (matches), contradicting (outliers), or lower level (ambiguities) taxonomic names is noted in parentheses. The ordinal affiliation of the target group is given in the last column.\n\n2.2. Identification of Microbial Community Function using Conserved Enzyme Domains\nMicroarray probes designed to integrate transcripts of different taxa via conserved protein domains displayed several hybridization signals above the threshold value (Figure 1B). Although these functional probes were designed to cover broad sequence variation at conserved protein sites, only a small fraction of probes was actually labeled by RNA from soil samples. Of the three domains selected for chitinase probe design (508, 499, and 502 probes, respectively), 10 hybridization signals were detected with probes targeting the first domain and seven signals were obtained with probes targeting the third domain. The second targeted sequence domain showed no hybridization signals above threshold values.\nTwo conserved regions were selected for the design of 501 and 490 endopeptidase probes, respectively. The hybridization results revealed that significant signals were found with three probes designed from the first of the two conserved endopeptidase regions.\nLaccase probes were designed from two conserved domains (254 and 280 probes, respectively); laccase transcripts were detected by three probes, two of which targeted site 1.\nThe amount of plant-derived RNAs within the total RNA samples from soil seemed to be comparably small, since only 5 and 17 probes encoding plant actin and invertase transcripts exhibited signals above the exclusion limit, respectively. Again, these sequences were designed to conserved domains in plant actin and invertase protein sequences and were used to estimate the amount of plant RNA within the total RNA preparation of the soil sample.\nIn summary, the functional enzyme signals allowed us to quantitatively describe transcript amounts of different functional enzymes (Figure 2). Reproducibility between these types of probes was also high between the two independent experiments, which is a prerequisite for comparative quantitative analyses. Interestingly, the amount of plant invertase transcripts was considerably higher and less reproducible. However, the spatial distribution and amount of plant tissues (e.g., roots) in soil samples is certainly not comparable to the distribution of fungal microorganisms, and one can expect that the presence or absence of such macroscopic tissues like roots will cause greater variability in hybridization signals.\nFigure 2 Expression profile of selected functional enzymes obtained from two independent hybridisation experiments; for detailed probe description, see the Experimental Section.\n\n3."}