A 16S rDNA PCR-based theoretical to actual delta approach on culturable mock communities revealed severe losses of diversity information

Abstract

Background: Subunits of ribosomal RNA genes (rDNAs) characterized by PCR-based protocols have been the proxy for studies in microbial taxonomy, phylogenetics, evolution and ecology. However, relevant factors have shown to interfere in the experimental outputs in a variety of systems. In this work, a 'theoretical' to 'actual' delta approach was applied to data on culturable mock bacterial communities (MBCs) to study the levels of losses in operational taxonomic units (OTUs) detectability. Computational and lab-bench strategies based on 16S rDNA amplification by 799F and U1492R primers were employed, using a fingerprinting method with highly improved detectability of fragments as a case-study tool. MBCs were of two major types: in silico MBCs, assembled with database-retrieved sequences, and in vitro MBCs, with AluI digestions of PCR data generated from culturable endophytes isolated from cacao trees.

Results: Interfering factors for the 16 s rDNA amplifications, such as the type of template, direct and nested PCR, proportion of chloroplast DNA from a tropical plant source (Virola officinalis), and biased-amplification by the primers resulted in altered bacterial 16S rDNA amplification, both on MBCs and V. officinalis leaf-extracted DNA. For the theoretical data, the maximum number of fragments for in silico and in vitro cuts were not significantly different from each other. Primers' preferences for certain sequences were detected, depending on the MBCs' composition prior to PCR. The results indicated overall losses from 2.3 up to 8.2 times in the number of OTUs detected from actual AluI digestions of MBCs when compared to in silico and in vitro theoretical data.

Conclusions: Due to all those effects, the final amplification profile of the bacterial community assembled was remarkably simplified when compared to the expected number of detectable fragments known to be present in the MBC. From these findings, the scope of hypotheses generation and conclusions from experiments based on PCR amplifications of bacterial communities was discussed.

Keywords: 16S rDNA metagenomics; ARDRA; CAPS; Chimeric sequences; Community structure; Fingerprinting methods; Hypervariable regions; Microbial ecology and diversity; PCR-RFLP; Plant-associated bacteria.

Fig. 1

Effects of chloroplast DNA from Virola officinalis and number of amplification reactions on PCR products from the 16S rRNA gene amplified with primers 799F and U1492R. a Amplification patterns from ‘nested’ and ‘direct’ PCR done on pooled DNA from a 5-OTUs’ mock bacterial community (MBC), with different percentages of amplified/purified V. officinalis cp DNA (see Methods) added to samples prior to the PCR. Agarose and polyacrylamide gels were stained with GelGreen™. ‘L’: size ladder; ‘100’, ‘65’, ‘30’, ‘0’: percentage of cpDNA; ‘C+’: positive control (DNA of a single bacterial isolate); ‘C–’: negative control. b Chloroplast and bacterial 16S rDNA amplification patterns in V. officinalis total leaf-DNA samples from nested PCR; ‘1 to 5’: distinct plant individuals from the same area. c Patterns of nested and direct PCR from the same samples of total DNA extracted from V. officinalis leaves; ‘6 to 9’: distinct plants from same area. d AluI restriction profiles from MBCs amplified with different percentages of cpDNA. ‘L’: size ladder (50 pb); ‘C5’: 5-OTUs’ MBC; ‘C10’: 10-OTUs’ MBC; ‘100’, ‘65’, ‘30’, ‘0’: percentages of cpDNA in the MBCs, prior to PCR, digestion and electrophoresis in 5–11% polyacrylamide gradient gel

Fig. 2

AluI restriction profiles of bacterial endophytes (OTUs) from cacao, amplified by 799F and U1492R primers. OTUs were chosen to illustrate the variety of banding patterns and counts observed from individual AluI digestions of the 72 culturable isolates. Amplicons from each OTU (lane) were digested by the 4-bp cutter restriction enzyme and electrophoresed in 5–11% polyacrylamide gradient gel. The ID numbers of isolates (Additional file 1: Table S1) appear at the top. Fragment sizes (in bp) are indicated on the left; ‘L’: 50-bp ladder. Arrows illustrate 5-bp size differences between fragments

Fig. 3

AluI restriction profile of mock bacterial communities (MBCs) assembled with DNA from cacao bacterial isolates (OTUs) prior to PCR. The increasing numbers of OTUs per MBC are indicated on top of the respective lanes. ‘L’: size ladder (50 bp); I, II and III: types of MBCs composed according to criteria described in Additional file 3: Table S3. The increasing number of OTUs in each MBC type was established by stepwise addition of 5 new OTUs to the existing ones (Additional file 3: Table S3). Equimolar amounts of DNA extracted from each isolate were used to compose each MBC, prior to direct PCR with primers 799F and U1492R; the total amount of template DNA used in the PCR for each MBC was 8 ng. The whole volumes of AluI digestions of the MBCs (8 μl) were applied on the 5–11% polyacrylamide gradient gels

Fig. 4

Theoretical and actual AluI restriction fragments of mock bacterial communities (MBCs). For all cases, the MBCs were assembled with 5-OTUs’ increments, according to the graphs. For both in silico and in vitro analyses, restriction fragments equal to, or less than 5-bp difference were considered as a single band and counted only once per MBC (see Methods). In silico analysis: the number of AluI fragments was defined from 16S rDNA sequences from various endophytic bacteria from database, followed by computational processing and restriction analysis by free online software (see Methods). ‘■’: sequences (OTUs) from the literature (Additional file 2: Table S2); ‘▲’: sequences from rice, Oryza sativa [54]; ‘♦’: sequences from bean, Phaseolus vulgaris [55]. The plotted data were the averages of maximum numbers of restriction fragments from 3000 random MBCs for the Additional file 2: Table S2’, rice’s and beans’ OTUs. In vitro analysis: the AluI fragments of 16S rDNA amplified by 799F/U1492R were obtained in two ways: ‘●’: pre MBC assembly ‘theoretical’ data; DNA extracted from cacao bacterial isolates (Additional file 1: Table S1) were individually subjected to PCR / AluI digestion (Fig. 2), computing the number of fragments per isolate prior to defining the MBCs; for each no. of OTUs on the horizontal axis, the total number of fragments for each of the 5 MBCs (Additional file 3: Table S3) were plotted. ‘○’: post MBC assembly ‘actual’ data; restriction fragments were obtained from PCR and digestion of previously pooled equimolar amounts of isolate DNAs, according to the no. of OTUs per MBC (Fig. 3). These experiments were repeated at least twice for all isolates and their MBCs. The respective regressions, equations and coefficients of determination are indicated in the graphs