Abstract

In the attempt to explore complex bacterial communities of environmental samples, primers hybridizing to phylogenetically highly conserved regions of 16S rRNA genes are widely used, but differential amplification is a recognized problem. The biases associated with preferential amplification of multitemplate PCR were investigated using ‘universal’ bacteria-specific primers, focusing on the effect of primer mismatch, annealing temperature and PCR cycle number. The distortion of the template-to-product ratio was measured using predefined template mixtures and environmental samples by terminal restriction fragment length polymorphism analysis. When a 1 : 1 genomic DNA template mixture of two strains was used, primer mismatches inherent in the 63F primer presented a serious bias, showing preferential amplification of the template containing the perfectly matching sequence. The extent of the preferential amplification showed an almost exponential relation with increasing annealing temperature from 47 to 61°C. No negative effect of the various annealing temperatures was observed with the 27F primer, with no mismatches with the target sequences. The number of PCR cycles had little influence on the template-to-product ratios. As a result of additional tests on environmental samples, the use of a low annealing temperature is recommended in order to significantly reduce preferential amplification while maintaining the specificity of PCR.

Introduction

Molecular approaches based on PCR amplification of 16S rRNA genes from directly extracted DNA have increased in popularity through the exploration of a tremendous prokaryotic diversity, which has been overlooked by the limitations and selectivity of traditional culture-dependent techniques (Hugenholtz et al., 1998). PCR-based microbial community analyses give for most environmental material a more realistic picture of the community structure than do classical techniques based on cultivation. The standard protocols for cultivation-independent analysis generally involve the isolation of nucleic acids from environmental samples, followed by PCR amplification with ‘universal’ primers, that is, primers binding to phylogenetically highly conserved regions of the 16S rRNA genes. The separation of heterogeneous products is carried out directly by cloning or by various electrophoretic separation techniques, such as denaturing gradient gel electrophoresis (DGGE) (Muyzer & Smalla, 1998) and single-strand-conformation polymorphism (SSCP) (Schwieger & Tebbe, 1998), or after restriction enzyme digestion (Weidner et al., 1996; Marsh, 1999). However, the implicit assumption that numerically important bacteria in environmental samples are represented by dominant clones in 16S rRNA gene libraries or strong bands in molecular fingerprinting patterns, such as DGGE or SSCP gel pictures, cannot be taken for granted.

In fact, almost each step of the molecular approach can introduce biases or errors (von Wintzingerode et al., 1997). While differences in cell lysis efficiencies may be overcome by the combination of different nucleic acid isolation techniques (Frostegard et al., 1999), the quantitative aspect of the multitemplate PCR amplification step is poorly understood. In many studies, the choice of PCR parameters seems to be driven by the aim of obtaining a high amount of specific PCR product instead of by a consideration of how parameters may distort the template-to-product ratio.

Different primer binding energies originating from the degeneracy of ‘universal’ primers are possible sources of bias (Polz & Cavanaugh, 1998; Lueders & Friedrich, 2003). Primer mismatch is an inherent characteristic of PCR with ‘universal’ primers, while, owing to single nucleotide variability even in the evolutionarily highly conserved regions of the rRNA genes, the designation of a perfectly matching ‘universal’ primer is simply not possible (Schmalenberger et al., 2001; Baker et al., 2003). Despite the observation of preferential amplification of certain 16S rRNA gene sequences (Reysenbach et al., 1992), only a few experiments have been carried out with defined template mixtures of 16S rRNA gene copies to assess how parameters influence the information obtained about community structure. The choices of annealing temperature (Ishii & Fukui, 2001), template concentration and cycle number have been thought to have a significant effect on biases caused by selective amplification (Chandler et al., 1997; Dohrmann & Tebbe, 2004). Moreover, genomic properties such as genome size, copy number of 16S rRNA genes, and G+C content influence the PCR product ratios (Farrelly et al., 1995; Crosby & Criddle, 2003). A cycle number-dependent bias was observed with the template-to-product ratio of 16S rRNA genes by Suzuki and Giovannoni (Suzuki & Giovannoni, 1996; Suzuki et al., 1998), and a mathematical model was developed to explain this bias by the reannealing of amplicons with increased concentrations, which inhibits the formation of the template–primer hybrids. On the other hand, subsequent studies did not detect any cycle-dependent template reannealing bias (Osborn et al., 2000; Lueders & Friedrich, 2003; Acinas et al., 2005), and suggested that the template reannealing was insignificant for diverse environmental samples.

In this study, the effect of primer mismatches was thoroughly investigated with defined genomic templates by terminal restriction fragment length polymorphism (T-RFLP) using two widely applied forward primers, namely 27F (Lane, 1991) and 63F (Marchesi et al., 1998), and one reverse primer, namely 1387R (Marchesi et al., 1998). A detailed, statistically supported analysis on the effects of annealing temperature and cycle number was carried out comparing template mixtures with and without primer mismatches. Furthermore, the study was extended to DNA from environmental rhizoplane samples in order to evaluate the PCR parameters that influence the outcome of environmental microbial community analyses.

Materials and methods

Samples and DNA extraction

The following strains were used in this study: Aeromonas hydrophila (ATCC 7966); Bacillus cereus (ATCC 14579); Bacillus subtilis (ATCC 6633); and Pseudomonas fluorescens (ATCC 13525). They were cultivated on solid DSMZ-1 medium (http://www.dsmz.de). Genomic DNA from 48-h cultures was extracted using a V-gene Bacterial Genomic DNA Mini-prep Kit (V-Gene Biotechnology Limited, Hangzhou, China) according to the manufacturer's instructions.

Typha latifolia (cattail) root samples were taken from a constructed wetland treating industrial wastewater at Demjén, Hungary in May 2003. The rhizoplane fraction was separated by washing and centrifugation as described in Nikolausz et al. (2004). DNA from the rhizoplane fraction was extracted as follows: a 200-mg sample was resuspended in 1 mL of CLS-TC buffer (FastDNA Kit, MP Biomedicals, Irvine, CA) in a 2-mL FastPrep tube containing lysing matrix A and 100 mg of polyvinylpolypyrrolidone (PVPP). The tubes were shaken in a bead beater for 2 min (Vibrogen-Zellmühle, Edmund Bühler, Hechingen, Germany). After centrifugation of the samples at 14 000 g for 5 min, the supernatant was transferred to a new tube, in which DNA was purified using the purification segment of the V-gene Bacterial Genomic DNA Mini-prep Kit. Both the pure cultures and the environmental DNA were eluted with 60 μL of 2.5 mM Tris-HCl, pH 8.5 buffer. DNA yield and purity were determined spectrophotometrically (Lambda 35 UV-VIS Spectrometer, PerkinElmer Life and Analytical Sciences, Boston, MA) by the A260 nm/A280 nm ratio in triplicate with an A320 nm baseline correction.

PCR amplification conditions

Forward primer 27F or 63F 5′-labelled with tetrachloro-fluorescein phosphoramidite (TET) and the 1387R primer were used in this study. Table 1 shows the alignment of the primers to the target sequence of the selected strains. PCR was performed in 0.2-mL tubes with a final volume of 25 μL on a Hybaid thermocycler (PCR Express, Thermo Electron, Middlesex, UK). The temperature protocol for PCR amplifications, if not otherwise stated, was as follows: initial denaturation at 96°C for 3 min, followed by 32 amplification cycles of 30 s at 94°C, 30 s at 52°C and 1 min at 70°C, followed by final extension at 70°C for 10 min. The annealing temperature for gradient PCR amplification ranged from 47 to 61°C. The PCR reaction mixture contained 1 × HotMaster Taq Buffer with 2.5 mM Mg2+ (Eppendorf AG, Hamburg, Germany), 200 μM of each deoxynucleoside triphosphate, 1.25 U of HotMaster Taq DNA Polymerase (Eppendorf), 0.325 μM of each primer, and 10 ng of genomic DNA template mixture in a total volume of 25 μL. In experiments with varying cycle numbers, reactions were stopped after 18, 24, 32 and 48 cycles, followed by incubation at 70°C for 10 min. PCR products were separated on a 1% agarose gel stained with ethidium bromide and visualized with UV excitation (Sambrook et al., 1989). Each PCR was performed in triplicate using templates obtained by independent mixing procedures.

1

Alignment of ‘universal’ 16S rRNA gene primers (27F, 63F and 1387R) with the target sequences of the four strains of this study: AH, Aeromonas hydrophila; BC, Bacillus cereus; BS, Bacillus subtilis; PF, Pseudomonas fluorescens

TET-27F-primer 5′ AGA GTT TGA TCM* TGG CTC AG 3′ 
AH 5′ AGA GTT TGA TCC TGG CTC AG 3′ 
BC 5′ AGA GTT TGA TCC TGG CTC AG 3′ 
BS 5′ AGA GTT TGA TCC TGG CTC AG 3′ 
PF 5′ AGA GTT TGA TCC TGG CTC AG 3′ 
TET-63F-primer 5′ CAG GCC TAA CAC ATG CAA GTC 3′ 
AH 5′ CAG GCC TAA CAC ATG CAA GTC 3′ 
BC 5′ CGT GCC TAA TAC ATG CAA GTC 3′ 
BS 5′ CGT GCC TAA TAC ATG CAA GTC 3′ 
PF 5′ CAG GCC TAA CAC ATG CAA GTC 3′ 
1387R primer 3′ CGG AAC ATG TGW* GGC GGG 5′  
AH 5′ GCC TTG TAC ACA CCG CCC 3′  
BC 5′ GCC TTG TAC ACA CCG CCC 3′  
BS 5′ GCC TTG TAC ACA CCG CCC 3′  
PF 5′ GCC TTG TAC ACA CCG CCC 3′  
TET-27F-primer 5′ AGA GTT TGA TCM* TGG CTC AG 3′ 
AH 5′ AGA GTT TGA TCC TGG CTC AG 3′ 
BC 5′ AGA GTT TGA TCC TGG CTC AG 3′ 
BS 5′ AGA GTT TGA TCC TGG CTC AG 3′ 
PF 5′ AGA GTT TGA TCC TGG CTC AG 3′ 
TET-63F-primer 5′ CAG GCC TAA CAC ATG CAA GTC 3′ 
AH 5′ CAG GCC TAA CAC ATG CAA GTC 3′ 
BC 5′ CGT GCC TAA TAC ATG CAA GTC 3′ 
BS 5′ CGT GCC TAA TAC ATG CAA GTC 3′ 
PF 5′ CAG GCC TAA CAC ATG CAA GTC 3′ 
1387R primer 3′ CGG AAC ATG TGW* GGC GGG 5′  
AH 5′ GCC TTG TAC ACA CCG CCC 3′  
BC 5′ GCC TTG TAC ACA CCG CCC 3′  
BS 5′ GCC TTG TAC ACA CCG CCC 3′  
PF 5′ GCC TTG TAC ACA CCG CCC 3′  
*

M, A/C; W, A/T.

Grey-highlighted letters indicate the three positions within primer 63F where mismatches with B. cereus and B. subtilis DNA occur. Primers 27F and 1387R match perfectly with all of the four strains.

Purification of PCR products and restriction enzyme digestion

The purification of PCR products was carried out using a Viogene PCR-M Clean Up System (Proteogenix S.A., Illkirch, France). In order to validate the model system, reliability measurements were carried out by mixing pure culture PCR products of known quantities prior to enzyme digestion. One hundred and fifty nanograms of purified PCR products mixed in a 1 : 1 ratio were digested in a 20-μL reaction volume with 3 U of restriction endonuclease for 3 h (three parallel reactions for each). Digestions were carried out separately at 37°C for enzymes Hin6I, MspI, Csp6I or Bsh1236I (Fermentas, Vilnius, Lithuania) or at 65°C for TasI (Fermentas) using only 0.4 U to prevent star activity (Nasri & Thomas, 1987). The same conditions were applied for the digestion of multi-template PCR products (model community and environmental samples). Prior to electrophoresis, restriction enzyme digestion mixtures were further purified with ethanol precipitation (Sambrook et al., 1989). Pellets were stored at −20°C and resuspended in 20 μL of distilled water prior to T-RFLP analysis.

Determination of nucleotide sequences

The almost complete 16S rRNA gene sequence of each strain was determined using a Big Dye Terminator Cycle Sequencing Kit V3.1 (Applied Biosystems, Foster City, CA), according to the manufacturer's protocol. Sequencing products were separated on a Model 310 Genetic Analyser (Applied Biosystems). Primers used for 16S rRNA gene sequencing reactions were 63F, 519R, 357F, 803F, 1114F, 1378R (Lane, 1991), 338R (Amann et al., 1990) and 1492R (Polz & Cavanaugh, 1998). The DNA sequences obtained from this study were submitted to GenBank (accession numbers DQ207728DQ207731).

T-RFLP analysis

Samples were prepared with 12 μL of deionized formamide, 0.6 μL of 500 TAMRA internal size standard (Applied Biosystems) and 0.5–1.5 μL enzymatic digest. Prior to T-RFLP analysis, the mixtures were denatured at 96°C for 5 min and immediately placed on ice. Fluorescently labelled terminal restriction fragments (T-RFs) were separated on a model ABI PRISM 310 Genetic Analyzer (Applied Biosystems) with POP-4™ Polymer, with three replicates for each digestion mixture to ensure reproducibility. The genemapper version 3.7 software (Applied Biosystems) was used to analyze the T-RFLP chromatograms. In the case of the environmental samples, the T-RFLP profiles were analysed using the method of Saikaly (2005) with the following modifications. Instead of peak height, the profiles were standardized for the cumulative peak areas, excluding the primer peaks. The range of the sum of peak areas was between 210 402 and 261 963 fluorescent units for the varying annealing-temperature experiment, and between 54 965 and 75 217 fluorescent units for the cycle-number experiment, where the total fluorescent intensity was adjusted to the weakest sample of 18 cycles.

In order to compare diversity between environmental samples, the Shannon–Weaver diversity index (H) was calculated as follows: H=−Σ(pi) (log2pi), where i represents all the unique fragments and pi is the relative abundance of fragment i (Shannon & Weaver, 1949). The abundance of each particular fragment was determined using the corresponding peak area in fluorescent units.

Results and discussion

Effect of capillary electrophoretic separation on the relative quantification of PCR products

Labelled PCR products were mixed in known quantities (1 : 1 ratio), digested with various restriction endonucleases, and separated by capillary electrophoresis in order to validate the relative quantification calculated on the basis of peak height and peak area data. In most cases, the original ratio was observed on the T-RFLP chromatogram with good reproducibility and acceptable SD (1±0.1) values (Fig. 1). The ratio of the peak heights, on the other hand, differed significantly from the original concentration ratio when the difference of the resolved T-RF lengths was significant (>50 bp), but even then the ratio of the peak areas approximated the expected value. The deviation from the defined ratio (in these experimental setups representing the empirical margin of error) was highly dependent on the size difference of the fragments separated by capillary electrophoresis but was not influenced by the actual size of the T-RFs (Fig. 1). The increasing deviation from the 1 : 1 ratio was caused by the flattening of peaks of longer terminal fragments as a result of the inherent limits of capillary electrophoresis. This implies that peak area should be preferred to peak height when the semi-quantitative analysis of microbial communities is performed by T-RFLP analysis using capillary electrophoresis. Contradicting this, Lueders & Friedrich (2003) suggested calculating with peak height for T-RFLP analysis using gel electrophoretic separation. Our results draw attention to the importance of experimental conditions (instrument, T-RFLP separation technique and software analysis) in the use of peak height or peak area integration.

1

Relative peak height and peak area ratios of paired T-RFs of different sizes. Purified PCR products from different strains were mixed in replicates in a 1 : 1 concentration ratio and digested with restriction enzymes (Hin6I, TasI, Bsh1236I or Csp6I). The obtained ratios should approximate 1.0; the dotted lines represent the interval of acceptable deviation (1 ± 0.1). When calculating both the peak height and the peak area ratios, the value of the shorter fragment was always correlated with that of the longer T-RF. The horizontal axis shows both the size difference in basepair units and the length of the resolved T-RFs in parentheses. In each figure error bars indicate the SD of the three independent PCRs and the three parallel T-RFLP runs.

1

Relative peak height and peak area ratios of paired T-RFs of different sizes. Purified PCR products from different strains were mixed in replicates in a 1 : 1 concentration ratio and digested with restriction enzymes (Hin6I, TasI, Bsh1236I or Csp6I). The obtained ratios should approximate 1.0; the dotted lines represent the interval of acceptable deviation (1 ± 0.1). When calculating both the peak height and the peak area ratios, the value of the shorter fragment was always correlated with that of the longer T-RF. The horizontal axis shows both the size difference in basepair units and the length of the resolved T-RFs in parentheses. In each figure error bars indicate the SD of the three independent PCRs and the three parallel T-RFLP runs.

Effect of primer mismatch

The effect of primer mismatch was investigated with two widely used forward primers (27F and 63F): 27F showed no mismatch with all of the target sequences, while the 63F primer had three mismatches close to the 5′ end against DNA isolated from the Bacillus strains. The reverse primer (1387R) matched all targets perfectly (Table 1). To demonstrate the effect of the primer mismatch, DNA isolated from different strains was mixed in equal ratios, and 16S rRNA genes were amplified by PCR using different primer sets (TET-27F and 1387R, later referred to as 27F, and TET-63F and 1387R, later referred to as 63F). As expected, A. hydrophila and P. fluorescens were preferentially amplified over both Bacillus strains when the 63F primer was used, while the 27F primer amplified all templates without bias. For further reference, these setups are referred to as biased or mismatch and nonbiased or nonmismatch PCR, respectively. When the mismatch template was represented in a lower amount (1 : 10), the corresponding PCR product could not be detected (data not shown), which indicates that important members of a community may be overlooked as a result of preferential amplification. Further experiments focused on PCR parameters, which may enhance or attenuate the quantitative bias caused by primer mismatches.

Effect of annealing temperature on primer mismatch

The efficiency of primer annealing is a very important factor for the success and stringency of PCR, and can be modified by factors such as the chemical constitution of the buffer (PCR enhancers, cosolvents), primer concentration, Mg2+ concentration, and the annealing temperature (Markoulatos et al., 2002). This latter factor was investigated as the main parameter of PCR optimization, because it is easy both to measure and to modify.

Biased and nonbiased PCRs were performed at various annealing temperatures between 47 and 61°C using mixtures of templates. Figure 2a shows the results of the gradient PCR obtained from the mixture (1 : 1) of the mismatch pair of A. hydrophila (AH) and B. subtilis (BS) DNA template using 27F and 63F primer sets. PCR products were digested with Hin6I restriction enzyme resulting in two T-RFs with a 25-bp size difference, resulting in reproducibility that is within a 5% margin of error (Fig. 1). The deviation from the nearly 1 : 1 template ratio during amplification with the mismatch primer increased almost exponentially with increasing annealing temperature. Owing to this relationship, at annealing temperatures of 59.9 and 61°C the detection of the mismatch template resulted in relatively high SD values of peak area ratios (Fig. 2a). The original ratio was approximated when using lower annealing temperatures. Such a nearly exponential relationship between the annealing temperature and the distortion from the template ratio, greater than the margin of error, was not observed with the perfectly matching 27F primer (Fig. 2a). It is noteworthy that even a very low annealing temperature (47°C) had no negative effect on the PCR product ratios from the nonbiased amplifications.

2

(a) Relative peak area ratios of T-RFs from Aeromonas hydrophila (AH) and Bacillus subtilis (BS) at various annealing temperatures. In each case, the peak area data of AH were correlated with those of BS. The T-RFs were generated with the 27F and the 63F primer sets; the PCR products in both setups were digested with Hin6I. (b) Relative peak area ratios of T-RFs from Pseudomonas fluorescens (PF) and Bacillus cereus (BC) at various annealing temperatures. In each case, the peak area data of PF were correlated with those of BC. The T-RFs were generated with the 27F and the 63F primer sets; the PCR products in both setups were digested with Hin6I and TasI in separate reactions. Expected ratios are indicated with broken lines.

2

(a) Relative peak area ratios of T-RFs from Aeromonas hydrophila (AH) and Bacillus subtilis (BS) at various annealing temperatures. In each case, the peak area data of AH were correlated with those of BS. The T-RFs were generated with the 27F and the 63F primer sets; the PCR products in both setups were digested with Hin6I. (b) Relative peak area ratios of T-RFs from Pseudomonas fluorescens (PF) and Bacillus cereus (BC) at various annealing temperatures. In each case, the peak area data of PF were correlated with those of BC. The T-RFs were generated with the 27F and the 63F primer sets; the PCR products in both setups were digested with Hin6I and TasI in separate reactions. Expected ratios are indicated with broken lines.

A similar result was obtained with another mismatch mixture of DNA template of P. fluorescens (PF) and B. cereus (BC) independently of the restriction enzymes used (Fig. 2b). The bias introduced by primer mismatches was reduced at lower annealing temperatures and reached the ratio obtained with the nonbiased PCR. It should be noted that neither the nonbiased nor the biased PCR at low annealing temperature gave the expected ratio calculated from rRNA operon frequencies (Table 2), possibly as a result of other primer-independent biases (Farrelly et al., 1995; Hansen et al., 1998) (e.g. changes of genome size during cultivation; Vellai et al., 1999). It is obvious from the comparison of the two sets of primers (Fig. 2b) that the effect of the mismatch bias was more significant than that of the difference in genomic properties. When the nonmismatch template pair of AH and PF were subjected to the same range of annealing temperatures, no correlation was found between the temperatures and the peak area ratios. The deviation from the set ratio was constant using both the 63F and the 27F primer sets (data not shown).

2

Genome size and number of rRNA operons of strains used in this study: AH, Aeromonas hydrophila; BC, Bacillus cereus; BS, Bacillus subtilis; PF, Pseudomonas fluorescens

 Operon number Genome size (106 bp) Operon frequency (operon per 106 bp) Operon frequency ratios* 
AH BC BS PF 
AH 10 4.50 2.22 1.00 0.92–1.00 0.94–1.17 2.95 
BC 12–13 5.40–5.41 2.22–2.40 1.00–1.08 1.00 0.93–1.27 2.94–3.19 
BS 8–10 4.21 1.90–2.38 0.86–1.07 0.79–1.07 1.00 2.52–3.15 
PF 6.63 0.75 0.34 0.31–0.34 0.32–0.40 1.00 
 Operon number Genome size (106 bp) Operon frequency (operon per 106 bp) Operon frequency ratios* 
AH BC BS PF 
AH 10 4.50 2.22 1.00 0.92–1.00 0.94–1.17 2.95 
BC 12–13 5.40–5.41 2.22–2.40 1.00–1.08 1.00 0.93–1.27 2.94–3.19 
BS 8–10 4.21 1.90–2.38 0.86–1.07 0.79–1.07 1.00 2.52–3.15 
PF 6.63 0.75 0.34 0.31–0.34 0.32–0.40 1.00 
*

Theoretical calculation of operon frequency ratios according to information on genomic properties of strains listed in the Ribosomal RNA Operon Copy Number Database (http://rrndb.cme.msu.edu/rrndb).

The effect of annealing temperature on the product ratio of multitemplate PCR was previously investigated by DGGE with a different ‘universal’ primer pair, and preferential amplification and reduction of the complexity of community patterns at high annealing temperatures were observed (Ishii & Fukui, 2001). Clone libraries generated from PCR products of an environmental sample obtained at different annealing temperatures with the 63F-1389R primers were compared, and a significant increase in the number of phylotypes was detected by lowering the annealing temperature (Hongoh et al., 2003), which is in good agreement with our results. The results presented here in addition highlight that, at certain low annealing temperatures, the nonmismatch ratios can be approached with mismatch primers and preferential amplification can be minimized.

Effect of PCR cycle number

PCR amplifications were performed in order to assess the effect of cycle number on multitemplate PCR. When genomic DNA was mixed as template, genomic properties affected the extent of deviation from the expected ratio (Fig. 3). However, it is difficult to take into account the effect of different genome sizes and rRNA copy numbers in the case of environmental community analysis, because these parameters are unknown for uncultured microorganisms. Furthermore, only limited information is available for cultivated bacteria in public databases (Klappenbach et al., 2000) to correct for this factor. The effect of the difference in genomic properties was intentionally omitted from the cycle-number experiments using PCR products obtained from amplification with primers 27F and 1492R as templates. Purified 16S PCR products of A. hydrophila and B. subtilis (AH-BS, mismatch) and A. hydrophila and P. fluorescens (AH-PF, nonmismatch) were mixed in equal amounts and amplified with the 63F primer set. Our results showed little effect of the number of cycles on the amplicon ratios, but the impact of the mismatch was apparent (Fig. 3). However, in the case of the AH-BS setup there was a slight decrease of the overrepresentation of AH with increasing cycle number.

3

The effect of cycle number on peak area ratios of PCR products generated from Aeromonas hydrophila (AH) and Bacillus subtilis (BS) and from Aeromonas hydrophila (AH) and Pseudomonas fluorescens (PF); similarly, the effect of cycle number on peak area ratios of genomic DNA generated from Aeromonas hydrophila (AH) and Pseudomonas fluorescens (PF). Purified PCR products obtained with primers 27F-1492R or genomic DNA were mixed in a 1 : 1 concentration ratio, amplified with the primer pair 63F at 54°C annealing temperature, and digested with Hin6I. The starting 1 : 1 template ratios were mixed in three replicates, and three parallel PCRs were run from each mixture: the area ratio value at each cycle number is thus the mean of nine data. The PCR reactions were stopped after 18, 24, 32 and 48 cycles.

3

The effect of cycle number on peak area ratios of PCR products generated from Aeromonas hydrophila (AH) and Bacillus subtilis (BS) and from Aeromonas hydrophila (AH) and Pseudomonas fluorescens (PF); similarly, the effect of cycle number on peak area ratios of genomic DNA generated from Aeromonas hydrophila (AH) and Pseudomonas fluorescens (PF). Purified PCR products obtained with primers 27F-1492R or genomic DNA were mixed in a 1 : 1 concentration ratio, amplified with the primer pair 63F at 54°C annealing temperature, and digested with Hin6I. The starting 1 : 1 template ratios were mixed in three replicates, and three parallel PCRs were run from each mixture: the area ratio value at each cycle number is thus the mean of nine data. The PCR reactions were stopped after 18, 24, 32 and 48 cycles.

Implications for the analysis of environmental samples

Our results obtained from the simple model communities imply that the annealing temperature should be kept as low as possible in order to avoid preferential amplification because the quantitative aspects of product ratios were not affected even at very low annealing temperatures. This hypothesis was tested with environmental DNA isolated from cattail rhizoplane samples. PCR products were generated with two primer sets (27F and 63F) at various annealing temperatures and compared by T-RFLP analysis (Table 3). Most of the T-RFs were in fact detected with both primer pairs. There were, however, also unique peaks in both sets of chromatograms, indicating a primer-induced bias (Fig. 4).

3

T-RFLP analyses of the same rhizoplane sample applying different PCR parameters

Primer set Annealing temperature (°C) Number of cycles Peak numbers Shannon index (H) (A+B)/C* 
27F 47.0 32 71.5 ± 14.8 1.19 ± 0.06 0.57 ± 0.02 
27F 50.9 32 65.0 ± 7.1 1.06 ± 0.04 0.56 ± 0.02 
27F 55.3 32 58.0 ± 1.4 0.99 ± 0.01 0.60 ± 0.03 
27F 59.9 32 49.0 ± 5.7 0.94 ± 0.01 0.67 ± 0.04 
63F 47.0 32 83.0 ± 1.4 1.31 ± 0.01 0.71 ± 0.01 
63F 50.9 32 75.5 ± 13.4 1.21 ± 0.08 0.78 ± 0.04 
63F 55.3 32 72.5 ± 8.7 1.17 ± 0.03 1.38 ± 0.13 
63F 59.9 32 70.7 ± 4.0 1.34 ± 0.05 3.89 ± 0.19 
27F 52.0 18 19.0 ± 2.6 0.89 ± 0.02 0.73 ± 0.02 
27F 52.0 24 22.0 ± 1.4 0.93 ± 0.00 0.82 ± 0.02 
27F 52.0 32 24.7 ± 5.5 0.87 ± 0.05 0.67 ± 0.01 
27F 52.0 48 30.7 ± 7.8 0.91 ± 0.07 0.55 ± 0.04 
63F 52.0 18 30.0 ± 1.7 1.04 ± 0.02 1.26 ± 0.02 
63F 52.0 24 26.8 ± 4.3 1.02 ± 0.02 1.40 ± 0.06 
63F 52.0 32 33.0 ± 2.1 1.07 ± 0.03 1.03 ± 0.05 
63F 52.0 48 26.5 ± 0.7 1.03 ± 0.03 0.83 ± 0.03 
Primer set Annealing temperature (°C) Number of cycles Peak numbers Shannon index (H) (A+B)/C* 
27F 47.0 32 71.5 ± 14.8 1.19 ± 0.06 0.57 ± 0.02 
27F 50.9 32 65.0 ± 7.1 1.06 ± 0.04 0.56 ± 0.02 
27F 55.3 32 58.0 ± 1.4 0.99 ± 0.01 0.60 ± 0.03 
27F 59.9 32 49.0 ± 5.7 0.94 ± 0.01 0.67 ± 0.04 
63F 47.0 32 83.0 ± 1.4 1.31 ± 0.01 0.71 ± 0.01 
63F 50.9 32 75.5 ± 13.4 1.21 ± 0.08 0.78 ± 0.04 
63F 55.3 32 72.5 ± 8.7 1.17 ± 0.03 1.38 ± 0.13 
63F 59.9 32 70.7 ± 4.0 1.34 ± 0.05 3.89 ± 0.19 
27F 52.0 18 19.0 ± 2.6 0.89 ± 0.02 0.73 ± 0.02 
27F 52.0 24 22.0 ± 1.4 0.93 ± 0.00 0.82 ± 0.02 
27F 52.0 32 24.7 ± 5.5 0.87 ± 0.05 0.67 ± 0.01 
27F 52.0 48 30.7 ± 7.8 0.91 ± 0.07 0.55 ± 0.04 
63F 52.0 18 30.0 ± 1.7 1.04 ± 0.02 1.26 ± 0.02 
63F 52.0 24 26.8 ± 4.3 1.02 ± 0.02 1.40 ± 0.06 
63F 52.0 32 33.0 ± 2.1 1.07 ± 0.03 1.03 ± 0.05 
63F 52.0 48 26.5 ± 0.7 1.03 ± 0.03 0.83 ± 0.03 
*

Area ratios of peaks marked in the chromatograms by letters (Fig. 4). Peak area data of peaks A and B were combined for simplicity, because at 48 cycles the two peaks were not resolved perfectly.

4

Effect of various PCR parameters on T-RFLP chromatograms generated from a cattail rhizoplane sample with the primer set 27F and 63F, digested with MspI. The chromatograms have been aligned along the x-axis for clarity. Peaks referred to in the text are marked with letters A, B and C. (a) Effect of annealing temperature on the T-RFLP profile of the rhizoplane sample. (b) Effect of cycle number on the T-RFLP profile of the rhizoplane sample.

4

Effect of various PCR parameters on T-RFLP chromatograms generated from a cattail rhizoplane sample with the primer set 27F and 63F, digested with MspI. The chromatograms have been aligned along the x-axis for clarity. Peaks referred to in the text are marked with letters A, B and C. (a) Effect of annealing temperature on the T-RFLP profile of the rhizoplane sample. (b) Effect of cycle number on the T-RFLP profile of the rhizoplane sample.

Increasing the annealing temperature resulted in fewer peaks, suggesting a less complex community structure, for both primer sets (Table 3). This contradicts the results of Osborn et al. (2000) that the number of T-RFs in samples increases using a higher annealing temperature. The different outcome may be attributed to an uneven sample loading into the wells of the acrylamide gel, which was not relevant for our study where capillary electrophoresis was used and sample loading was automated and optimized for total peak area.

In addition to the appearance of new peaks at lower temperatures that were not present or predominant in the PCR products obtained with high-stringency PCR (Fig. 4a), significant shifts occurred in the relative abundance of peaks, as best demonstrated with the area ratios of peaks A, B and C. In the case of the primer set 27F, the peak area ratio did not change significantly with increasing annealing temperature, in contrast to a major drop in the relative abundance of peak C with the 63F set (Fig. 4a, Table 3). This observation suggests that a mismatch bias contributed to the decrease of the relative area of peak C within the rhizoplane community profile. The Shannon diversity index decreased for both primer pairs and showed a slight increase at 59.9°C with 63F (Table 3). The increase of the Shannon index can be attributed to the significant change in the relative area of peak C. The investigations on environmental samples reinforce the results obtained with the model community experiments and prove that the detected diversity of a microbial community decreases with increasing annealing temperature.

The effect of PCR cycle number was also tested on the rhizoplane sample (Fig. 4b, Table 3). In the case of the 27F primer set there was a slight increase in total peak numbers with increasing cycles, but no significant relationship was observed with the 63F primer set. On the other hand, the Shannon indices did not change with higher cycle numbers for either primer set. Thus, contrary to Osborn et al. (2000), we could not demonstrate a more complex T-RFLP pattern with increasing cycle number. This result supports the opinion that diversity indices are better indicators of diversity than peak numbers alone when analysing complex microbial communities through fingerprinting techniques (Saikaly et al., 2005; Loisel et al., 2006).

Contrary to the overemphasis of PCR cycle-dependent bias owing to reannealing (Suzuki & Giovannoni, 1996), our results support the idea that the effect of different cycle numbers is of minor importance in the case of diverse templates from environmental samples even at a high number of PCR cycles (Lueders & Friedrich, 2003; Acinas et al., 2005). Nevertheless, the use of unnecessarily high cycle numbers is not recommended owing to the possibility of the formation of side products (Qiu et al., 2001), such as chimeras (Wang & Wang, 1997), heteroduplexes, and single-stranded DNA molecules (Jensen & Straus, 1993).

Conclusions

Because the design of perfectly matching universal primers remains an unsolved problem (Forney et al., 2004), the investigation of PCR parameters that could attenuate the quantitative bias associated with preferential amplification caused by primer mismatch is very important. Newly designed domain-specific primers are usually tested with genomic DNA of strains and environmental samples for PCR amplification efficiency (Marchesi et al., 1998; Baker et al., 2003), but preferential amplification should also be considered and tests carried out to determine the importance of this phenomenon. PCR optimization strategies for the assessment of environmental communities by ‘universal’ primer sets are directed to increasing the specificity of amplification using relatively high annealing temperature (Hansen et al., 1998) or ‘touch-down’ PCR protocol (Simpson et al., 2000). The results presented here suggest the opposite strategy for reducing preferential amplification using specific but low-stringency amplification. (1) PCR amplification should be optimized to reach the lowest annealing temperature, where the reaction is still specific and unspecific products (mispriming) are not observed. The cycle number can be high at this step. (2) PCR amplification should be repeated at the optimal temperature using parallel samples at low cycle numbers (around 25). In the case of low yield, parallel runs can be combined to obtain sufficient quantities for subsequent analyses.

Acknowledgements

This work was supported by a grant from the Hungarian National Science Foundation (OTKA T43617). We are indebted to the staff of the Department of Immunology of the Eötvös Loránd University of Science for assistance and access to their gradient thermocycler.

References

Acinas
SG
Sarma-Rupavtarm
R
Klepac-Ceraj
V
Polz
MF
(
2005
)
PCR-induced sequence artifacts and bias: insights from comparison of two 16S rRNA clone libraries constructed from the same sample
.
Appl Environ Microbiol
 
71
:
8966
8969
.
Amann
RI
Binder
BJ
Olson
RJ
Chisholm
SW
Devereux
R
Stahl
DA
(
1990
)
Combination of 16S rRNA-targeted oligonucleotide probes with flow cytometry for analyzing mixed microbial populations
.
Appl Environ Microbiol
 
56
:
1919
1925
.
Baker
GC
Smith
JJ
Cowan
DA
(
2003
)
Review and re-analysis of domain-specific 16S primers
.
J Microbiol Methods
 
55
:
541
555
.
Chandler
DP
Fredrickson
JK
Brockman
FJ
(
1997
)
Effect of PCR template concentration on the composition and distribution of total community 16S rDNA clone libraries
.
Mol Ecol
 
6
:
475
482
.
Crosby
LD
Criddle
CS
(
2003
)
Understanding bias in microbial community analysis techniques due to rrn operon copy number heterogeneity
.
Biotechniques
 
34
:
790
802
.
Dohrmann
AB.
Tebbe
CC.
(
2004
)
Microbial community analysis by PCR-single-strand conformation polymorphism (PCR-SSCP)
.
Molecular Microbial Ecology Manual
 ,
2nd edn
(
Kowalchuk
GA
De Bruijn
FJ
Head
IM
Akkermans
ADL
Van Elsas
JD
, eds), pp.
809
838
.
Kluwer Academic Publisher
,
Dordrecht
.
Farrelly
V
Rainey
FA
Stackebrandt
E
(
1995
)
Effect of genome size and rrn gene copy number on PCR amplification of 16S rRNA genes from a mixture of bacterial species
.
Appl Environ Microbiol
 
61
:
2798
2801
.
Forney
LJ
Zhou
X
Brown
CJ
(
2004
)
Molecular microbial ecology: land of the one-eyed king
.
Curr Opin Microbiol
 
7
:
210
220
.
Frostegard
A
Courtois
S
Ramisse
V
Clerc
S
Bernillon
D
Le Gall
F
Jeannin
P
Nesme
X
Simonet
P
(
1999
)
Quantification of bias related to the extraction of DNA directly from soils
.
Appl Environ Microbiol
 
65
:
5409
5420
.
Hansen
MC
Tolker-Nielsen
T
Givskov
M
Molin
S
(
1998
)
Biased 16S rDNA PCR amplification caused by interference from DNA flanking the template region
.
FEMS Microbiol Ecol
 
26
:
141
149
.
Hongoh
Y
Yuzawa
H
Ohkuma
M
Kudo
T
(
2003
)
Evaluation of primers and PCR conditions for the analysis of 16S rRNA genes from a natural environment
.
FEMS Microbiol Lett
 
221
:
299
304
.
Hugenholtz
P
Goebel
BM
Pace
NR
(
1998
)
Impact of culture-independent studies on the emerging phylogenetic view of bacterial diversity
.
J Bacteriol
 
180
:
4765
4774
.
Ishii
K
Fukui
M
(
2001
)
Optimization of annealing temperature to reduce bias caused by a primer mismatch in multitemplate PCR
.
Appl Environ Microbiol
 
67
:
3753
3755
.
Jensen
MA
Straus
N
(
1993
)
Effect of PCR conditions on the formation of heteroduplex and single-stranded-DNA products in the amplification of bacterial ribosomal DNA spacer regions
.
PCR Meth Appl
 
3
:
186
194
.
Klappenbach
JA
Dunbar
JM
Schmidt
TM
(
2000
)
rRNA operon copy number reflects ecological strategies of bacteria
.
Appl Environ Microbiol
 
66
:
1328
1333
.
Lane
DJ.
(
1991
)
16S/23S rRNA sequencing
.
Nucleic Acid Techniques in Bacterial Systematics
  (
Stackebrandt
E
Goodfellow
M
, eds), pp.
115
147
.
Wiley
,
New York
.
Loisel
P
Harmand
J
Zemb
O
Latrille
E
Lobry
C
Delgenes
JP
Godon
JJ
(
2006
)
Denaturing gradient electrophoresis (DGE) and single-strand conformation polymorphism (SSCP) molecular fingerprintings revisited by simulation and used as a tool to measure microbial diversity
.
Environ Microbiol
 
8
:
720
731
.
Lueders
T
Friedrich
MW
(
2003
)
Evaluation of PCR amplification bias by terminal restriction fragment length polymorphism analysis of small-subunit rRNA and mcrA genes by using defined template mixtures of methanogenic pure cultures and soil DNA extracts
.
Appl Environ Microbiol
 
69
:
320
326
.
Marchesi
JR
Sato
T
Weightman
AJ
Martin
TA
Fry
JC
Hiom
SJ
Dymock
D
Wade
WG
(
1998
)
Design and evaluation of useful bacterium-specific PCR primers that amplify genes coding for bacterial 16S rRNA
.
Appl Environ Microbiol
 
64
:
795
799
.
Markoulatos
P
Siafakas
N
Moncany
M
(
2002
)
Multiplex polymerase chain reaction: a practical approach
.
J Clin Lab Anal
 
16
:
47
51
.
Marsh
TL
(
1999
)
Terminal restriction fragment length polymorphism (T-RFLP): an emerging method for characterizing diversity among homologous populations of amplification products
.
Curr Opin Microbiol
 
2
:
323
327
.
Muyzer
G
Smalla
K
(
1998
)
Application of denaturing gradient gel electrophoresis (DGGE) and temperature gradient gel electrophoresis (TGGE) in microbial ecology
.
Anton Leeuw Int J G
 
73
:
127
141
.
Nasri
M
Thomas
D
(
1987
)
Alteration of the specificity of Pvuii restriction endonuclease
.
Nucleic Acids Res
 
15
:
7677
7687
.
Nikolausz
M
Márialigeti
K
Kovács
G
(
2004
)
Comparison of RNA- and DNA-based species diversity investigations in rhizoplane bacteriology with respect to chloroplast sequence exclusion
.
J Microbiol Meth
 
56
:
365
373
.
Osborn
AM
Moore
ERB
Timmis
KN
(
2000
)
An evaluation of terminal-restriction fragment length polymorphism (T-RFLP) analysis for the study of microbial community structure and dynamics
.
Environ Microbiol
 
2
:
39
50
.
Polz
MF
Cavanaugh
CM
(
1998
)
Bias in template-to-product ratios in multitemplate PCR
.
Appl Environ Microbiol
 
64
:
3724
3730
.
Qiu
XY
Wu
LY
Huang
HS
McDonel
PE
Palumbo
AV
Tiedje
JM
Zhou
JZ
(
2001
)
Evaluation of PCR-generated chimeras: mutations, and heteroduplexes with 16S rRNA gene-based cloning
.
Appl Environ Microbiol
 
67
:
880
887
.
Reysenbach
AL
Giver
LJ
Wickham
GS
Pace
NR
(
1992
)
Differential amplification of ribosomal-RNA genes by polymerase chain-reaction
.
Appl Environ Microbiol
 
58
:
3417
3418
.
Saikaly
PE
Stroot
PG
Oerther
DB
(
2005
)
Use of 16S rRNA gene terminal restriction fragment analysis to assess the impact of solids retention time on the bacterial diversity of activated sludge
.
Appl Environ Microbiol
 
71
:
5814
5822
.
Sambrook
JE
Fritsch
EF
Maniatis
T
(
1989
)
Molecular Cloning: A Laboratory Manual
 .
Cold Spring Harbor Laboratory Press
,
Cold Spring Harbor, NY
.
Schmalenberger
A
Schwieger
F
Tebbe
CC
(
2001
)
Effect of primers hybridizing to different evolutionarily conserved regions of the small-subunit rRNA gene in PCR-based microbial community analyses and genetic profiling
.
Appl Environ Microbiol
 
67
:
3557
3563
.
Schwieger
F
Tebbe
CC
(
1998
)
A new approach to utilize PCR-single-strand-conformation polymorphism for 16S rRNA gene-based microbial community analysis
.
Appl Environ Microbiol
 
64
:
4870
4876
.
Shannon
CE
Weaver
W
(
1949
)
The Mathematical Theory of Communication
 .
University of Illinois Press
,
Urbana, IL
.
Simpson
JM
McCracken
VJ
Gaskins
HR
Mackie
RI
(
2000
)
Denaturing gradient gel electrophoresis analysis of 16S ribosomal DNA amplicons to monitor changes in fecal bacterial populations of weaning pigs after introduction of Lactobacillus reuteri strain MM53
.
Appl Environ Microbiol
 
66
:
4705
4714
.
Suzuki
M
Rappe
MS
Giovannoni
SJ
(
1998
)
Kinetic bias in estimates of coastal picoplankton community structure obtained by measurements of small-subunit rRNA gene PCR amplicon length heterogeneity
.
Appl Environ Microbiol
 
64
:
4522
4529
.
Suzuki
MT
Giovannoni
SJ
(
1996
)
Bias caused by template annealing in the amplification of mixtures of 16S rRNA genes by PCR
.
Appl Environ Microbiol
 
62
:
625
630
.
Vellai
T
Kovacs
AL
Kovacs
G
Ortutay
C
Vida
G
(
1999
)
Genome economization and a new approach to the species concept in bacteria
.
P Roy Soc Lond B Bio
 
266
:
1953
1958
.
Von Wintzingerode
F
Gobel
UB
Stackebrandt
E
(
1997
)
Determination of microbial diversity in environmental samples: pitfalls of PCR-based rRNA analysis
.
FEMS Microbiol Rev
 
21
:
213
229
.
Wang
GCY
Wang
Y
(
1997
)
Frequency of formation of chimeric molecules is a consequence of PCR coamplification of 16S rRNA genes from mixed bacterial genomes
.
Appl Environ Microbiol
 
63
:
4645
4650
.
Weidner
S
Arnold
W
Puhler
A
(
1996
)
Diversity of uncultured microorganisms associated with the seagrass Halophila stipulacea estimated by restriction fragment length polymorphism analysis of PCR-amplified 16S rRNA genes
.
Appl Environ Microbiol
 
62
:
766
771
.

Author notes

Editor: Christoph Tebbe