- Email: [email protected]
Post-amplification Klenow fragment treatment alleviates PCR bias caused by partially single-stranded amplicons
Journal of Microbiological Methods 61 (2005) 69 – 75 www.elsevier.com/locate/jmicmeth
Post-amplification Klenow fragment treatment alleviates PCR bias caused by partially single-stranded amplicons Markus Egert1, Michael W. Friedrich* Max Planck Institute for Terrestrial Microbiology, Karl-von-Frisch-Strabe, 35043 Marburg, Germany Received 12 July 2004; received in revised form 2 November 2004; accepted 2 November 2004 Available online 30 November 2004
Abstract Partially single-stranded amplicons, formed during PCR amplification of single and mixed templates, are a potential source of bias in genetic diversity studies. The analysis of 16S rRNA gene diversity in mixed template samples by the fingerprinting technique terminal restriction fragment length polymorphism (T-RFLP) analysis can be biased by the occurrence of pseudo-TRFs, i.e., restriction fragments occurring in addition to the expected terminal restriction fragments of single amplicons. This bias originates from PCR products, which are single-stranded at their terminal restriction site. Here we show that treatment of PCR amplicons with Klenow fragment prior to restriction digest and T-RFLP analysis minimized effectively the occurrence of pseudo-T-RFs. Klenow fragment activity filled in bases into the partially single-stranded amplicons and thereby restored the affected amplicons to complete double strands. Our method allowed to improve the assessment of genetic diversity and gene ratios from T-RFLP analysis of an original environmental sample. Since partially single-stranded amplicons might influence many PCR-based techniques, post-amplification treatment with Klenow fragment may be useful for a wide range of applications, which assess the composition of amplicon pools, e.g., the analysis of marker gene diversity in mixed template samples by fingerprinting techniques or the analysis of sequence diversity by cloning. D 2004 Elsevier B.V. All rights reserved. Keywords: Genetic diversity analyses; Molecular fingerprinting; Single-stranded amplicons; T-RFLP
1. Introduction
* Corresponding author. Tel.: +49 6421 178830; fax: +49 6421 178809. E-mail address: [email protected] (M.W. Friedrich). 1 Present address: Laboratory of Microbiology, Wageningen University, Hesselink van Suchtelenweg 4, 6703 CT Wageningen, The Netherlands. 0167-7012/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.mimet.2004.11.002
The molecular analysis of genetic diversity in mixed template samples can be biased by several artifacts originating from PCR amplification (for reviews, see Kanagawa, 2003; von Wintzingerode et al., 1997). Sequence errors introduced by the DNA polymerase, the formation of chimeric PCR products, and the conversion of cloned heteroduplex amplicons
70
M. Egert, M.W. Friedrich / Journal of Microbiological Methods 61 (2005) 69–75
into hybrid homoduplex molecules by the mismatch repair system of the host cells can cause an overestimation of the number of sequence types present in the original sample (Qiu et al., 2001; Speksnijder et al., 2001; Thompson et al., 2002). Moreover, differences in primer binding energy (Lueders and Friedrich, 2003; Polz and Cavanaugh, 1998) and template re-annealing (Mathieu-Daude et al., 1996; Suzuki and Giovannoni, 1996) during later amplification cycles
might cause differences of original template ratios compared to their representation in the amplicon pool after amplification. Recently, we discovered that after PCR amplification of 16S rRNA genes from single and mixed template samples, a significant proportion of amplicons can be partially single-stranded (Egert and Friedrich, 2003), which can markedly affect subsequent genetic diversity analysis, e.g., by terminal restriction fragment length polymorphism (T-RFLP)
Fig. 1. Scheme illustrating the effect of pre-treatment of amplicons originating from a single template with mung bean nuclease (B) or Klenow fragment (C) vs. the untreated control (A) on the resulting T-RFLP profiles. Partly single-stranded amplicons, formed during PCR-amplification, lead to the formation of a pseudo-T-RF (red rectangle in schematic profiles), representing a restriction fragment longer than the true, terminal RF (black rectangle in schematic profiles). Pseudo-T-RFs form due to the inability of restriction endonucleases to cut amplicons, which are singlestranded (ss) at their terminal restriction sites (terminal restriction site as black square in schematic amplicons; second restriction site in red; actual cutting site indicated by triangle). Treatment with single-strand degrading mung bean nuclease prior to the restriction digest eliminates all single-stranded amplicons, and the pseudo-T-RF in the profile, but results in underestimation of the relative frequency of amplicons displaying pseudo-T-RF formation. In contrast, by treatment with Klenow fragment prior to the restriction digest, single-stranded amplicons are converted into fully double-stranded ones (newly synthesized bases of the amplicon in green), and thus, all amplicons are rendered accessible for endonucleases at their terminal restriction sites, which is a prerequisite for amplicon quantification in molecular diversity analysis. Note: TRFLP analyses allow to assess the relative (not the absolute) gene frequencies based on relative fluorescence units (RFU), by referring height or area of selected T-RF peaks to the peak height or area of the total profile or by comparing the ratio of peak heights or areas of selected peaks (see Table 1). FAM: 6V-carboxyfluorescein fluorescent label of amplicons.
M. Egert, M.W. Friedrich / Journal of Microbiological Methods 61 (2005) 69–75
profiling (Egert and Friedrich, 2003), denaturing gradient gel electrophoresis (DGGE; Simpson et al., 1999), and possibly also by cloning and sequencing of randomly selected clones. In case of T-RFLP analysis, i.e., a molecular fingerprinting technique widely used to rapidly compare genetic diversity and dynamics among different samples or treatments (for reviews see (Bruce and Hughes, 2000; Kitts, 2001), amplicons which are single-stranded at their terminal restriction site favor the formation of restriction fragments (RFs) longer than the true, expected terminal-RF (representing the second or subsequent restriction sites), because type II restriction enzymes cut double-stranded DNA only (Nishigaki et al., 1985). If not identified, these additional bpseudo-terminal restriction fragmentsQ (pseudo-T-RFs) can lead to an overestimation of genetic diversity in the original sample (Fig. 1A). Treatment of amplicons with single-strand degrading mung bean nuclease prior to T-RFLP analysis effectively eliminates pseudo-T-RFs from T-RFLP profiles (Egert and Friedrich, 2003). However, mung bean digest will eliminate all amplicons affected and thus renders these amplicons unavailable for estimation of relative frequencies by T-RF peak quantification (Lueders and Friedrich, 2003), resulting in a severe underestimation of genes represented by these amplicons (Fig. 1B). Consequently, this PCR bias may be solved much more satisfactorily by preventing the formation of partially single-stranded amplicons. However, the molecular mechanism underlying the formation of partly single-stranded amplicons is not clear. Temporary template secondary structures (Jensen and Straus, 1993) might cause the DNA polymerase to fall off during amplification prior to complete synthesis of the complementary strand (Polz and Cavanaugh, 1998). So far, a lower number of PCR cycles was shown to alleviate the extent of pseudo-TRF formation (Egert and Friedrich, 2003). However, partially single-stranded amplicons clearly occurred even with low cycles numbers (b20), and particularly in case of environmental samples, usually more PCR cycles (~25–30) are required to obtain a PCR signal sufficient for downstream analyses, e.g., fingerprinting. Several other strategies to minimize this bias were not successful (Egert and Friedrich, 2003). Neither varying conditions favoring primer/template annealing (i.e., annealing temperature, ionic strength of the
71
PCR buffer) nor variations of the primer ratios to exclude asymmetric primer utilization (Jensen and Straus, 1993), nor breconditioning PCRQ, i.e., a low cycle number re-amplification of a slightly diluted mixed-template PCR-product (Thompson et al., 2002), were effective in preventing artifactual amplicons (Egert and Friedrich, unpublished results). Here, we show that post-PCR treatment with Klenow fragment effectively reduces the amount of partially single-stranded PCR products in an amplicon pool, irrespective of their mode of formation during PCR. Klenow fragment, the large subunit of the DNA polymerase I of E. coli, is widely used in molecular biology, e.g., to produce DNA molecules with blunt ends from those with sticky ends by filling in complementary bases (Sambrook and Russel, 2001). We used Klenow fragment to convert partially singlestranded stretches of amplicons to the respective double-stranded molecules, rendering their terminal restriction sites accessible for restriction endonucleases (Fig. 1C), which is a prerequisite for a reliable estimation of genetic diversity in the original samples by T-RFLP analysis. Since also other PCR-based techniques in genetic diversity analysis are potentially affected by partially single-stranded PCR products (e.g., cloning and DGGE, (Simpson et al., 1999), post-PCR Klenow fragment treatment may improve original template representation in amplicon pools of PCR-based applications in general.
2. Materials and methods 2.1. Defined template mixture PCR-T-RFLP analysis was performed with a mixture of two clonal M13 products, both of which contained bacterial 16S rRNA gene inserts cloned from a beetle larva gut (Egert et al., 2003). Clone PeH55 (accession number AJ576368) displayed an expected T-RF of 72- and a 565-bp pseudo-T-RF, and was designated bpseudo-T-RF templateQ. Clone PeH59 (AJ538351), showed the expected 81-bp TRF only and no pseudo-T-RF (Fig. 2A), and was designated breference templateQ. PCR conditions and T-RFLP parameters were described in detail previously (Egert and Friedrich, 2003). Briefly, bacterial 16S rRNA gene fragments (~900 bp) were PCR
72
M. Egert, M.W. Friedrich / Journal of Microbiological Methods 61 (2005) 69–75
Fig. 2. Representative 16S rRNA gene T-RFLP profiles of (A) a clonal template mixture, comprising a pseudo-T-RF producing template and a reference template, and (B) an environmental sample, before and after treatment of amplicons with Klenow fragment prior to restriction digest. Treatment with Klenow fragment strongly reduced the respective pseudo-T-RF peaks (565 bp (A) and 165 bp (B), respectively) and simultaneously increased the height of the respective true and terminal restriction fragments (72 bp (A) and 64 and 125 bp (B)), measurable as changes in the peak height ratios of the 81 and 72 bp T-RFs (A) and the 125 and 64 bp T-RFs (B) (see Table 1). C: Crenarchaeota, MB: Methanobacteriaceae, RFU: relative fluorescence units.
amplified (32 cycles; ~25 ng of template DNA in a 50-Al standard PCR reaction mixture) using a primer pair specific for all Bacteria, of which the forward primer was fluorescently labeled. Purified amplicons (~75 ng) were digested with MspI, and restriction fragments were subsequently size-separated on an ABI373A automated sequencer in GeneScan mode. 2.2. Enzymatic treatments Amplicon pools were subjected to different treatments prior to T-RFLP analysis, i.e., before restriction digest. Treatments with mung bean nuclease to eliminate partially single-stranded amplicons were performed as described by Egert and Friedrich
(2003). Amplicons were treated with Klenow fragment to fill in bases into single strands as follows: ~1 Ag of PCR product was incubated with 5 U of Klenow fragment exonuclease minus (Promega, Mannheim, Germany), 10 Al of 10 reaction buffer, and 50 AM of each of the four deoxynucleoside triphosphates (dNTPs) in a total volume of 100 Al for 1 h at 20 8C in the dark. Klenow fragment exonuclease minus, which lacks the 3V-5Vexonuclease (proof-reading) activity (Derbyshire et al., 1988), was used to exclude any exonuclease activity from the incubations. The specificity of the treatment, i.e., DNA polymerase activity of the Klenow fragment, was checked by either incubating amplicons without Klenow fragment or replacing dNTPs by 100 AM of each of the four
M. Egert, M.W. Friedrich / Journal of Microbiological Methods 61 (2005) 69–75
dideoxynucleosid triphosphates (ddNTPS; Amersham Pharmacia Biotech., Freiburg, Germany), which terminates nucleoside incorporation by polymerase activity. All incubations were terminated by purification of amplicons with the MinElute PCR purification kit (Qiagen, Hilden, Germany).
73
were checked pair wise for statistical significance ( Pb0.05) with non-parametric Mann–Whitney tests using SYSTAT (SPSS, Chicago, USA). All treatments were replicated three to six times, based on independent PCR amplifications from a single template mixture or the environmental DNA extract, respectively.
2.3. Environmental sample 3. Results and discussion 16S rRNA gene fragments were amplified (38 cycles, 1 Al of 1:30 diluted DNA extract as template in a 50 Al standard PCR reaction mixture) from an environmental DNA extract (originating from the midgut of a beetle larva, (Egert et al., 2003) using an Archaea-specific primer pair, with the reverse primer fluorescently labeled. Mung bean nuclease and Klenow fragment treatments prior to restriction digest and T-RFLP analysis were performed as described above using AluI as restriction enzyme. The effect of Klenow fragment treatment on the T-RFLP profile of the original sample was estimated in addition using two archaeal clones, which represent the most abundant phylogenetic groups in the original sample. For this, 16S rRNA gene fragments (~800 bp) of clone PeMAr19 (AJ576131), representing Methanobacteriaceae, and clone PeMAr04 (AJ538352), representing Crenarchaeota, were amplified using ~5 ng of clonal M13 products as template in a standard PCR reaction with 38 cycles. Both clones displayed a 165bp pseudo-T-RF in addition to their true terminal RFs sized 64 and 125 bp, respectively (Egert and Friedrich, 2003). 2.4. Evaluation of treatments We used T-RFLP analysis to quantify the formation of partially single-stranded 16S rRNA gene amplicons. Effects of treatments were quantified by comparing the ratio of the heights of selected peaks in profiles before and after the treatments. In case of the defined template mixture, the height of the 81-bp T-RF of the reference template was compared to the height of the true T-RF peak (72 bp) of the pseudo-TRF template. In case of the environmental sample, changes in the peak height ratio of the true T-RFs representing Crenarchaeota (125 bp) and Methanobacteriaceae(64 bp), respectively, were followed. Differences in peak height ratios between treatments
Treatment of the amplicon pool with Klenow fragment prior to restriction digest almost completely removed the respective pseudo-T-RF of 565 bp from the T-RFLP profile of the template mixture (Fig. 2A), which indicates that partially single-stranded amplicons were filled in with bases at their ends, and thus, the double strand of these amplicons was restored. If true, the disappearance of the pseudo-T-RF should be accompanied by a respective increase of the true 72bp terminal restriction fragment of the pseudo-T-RF template. This was tested by quantifying changes in peak heights of the 72-bp T-RF of the pseudo-T-RF template relative to that of the 81-bp T-RF of the reference template by calculating ratios after T-RFLP analysis. Assuming a total conversion of partially single-stranded amplicons into double-strands, the peak height of the true terminal 72-bp RF was expected to equal the sum of heights of the 72-bp TRF and the former 565-bp pseudo-T-RF. Because the 81-bp T-RF of the reference template should not be affected by the Klenow fragment treatment, the T-RF peak height ratio was expected to decrease from 2.6 to 1.9 (Table 1). The latter ratio represents the real 16S rRNA gene ratio of the two clones in the amplicon pool. The effect of the different treatments on the suppression of pseudo-T-RF formation is given in Table 1. In fact, treatment of the amplicons with Klenow fragment prior to T-RFLP analysis not only eliminated the 565-bp T-RF of the pseudo-T-RF template almost completely (Fig. 2A) but also significantly reduced the peak height ratio to 2.1, which was in good agreement with the ratio predicted for a total conversion of the pseudo-T-RFs into true terminal RFs. Neither incubation of amplicons without Klenow fragment nor with ddNTPS instead of dNTPs did significantly change the peak height ratio in comparison to the ratio of the control profiles, indicating that the observed change was due to DNA
74
M. Egert, M.W. Friedrich / Journal of Microbiological Methods 61 (2005) 69–75
Table 1 Influence of different treatments of amplicons prior to restriction digest and T-RFLP analysis on the peak heights ratio of selected TRFs from 16S rRNA gene profiles of a defined clonal mixture, comprising a pseudo-T-RF producing template and a reference template (Fig. 2A), and an environmental DNA extract (Fig. 2B) Treatment
Peak height ratio Template mixture
Environmental sample
81 vs. 72 bp T-RF 125 vs. 64 bp T-RF Untreated Calculated conversion of pseudo-T-RFs into true T-RFs Klenow fragment Klenow fragment +ddNTPs Mung bean nuclease
2.6F0.1 1.9F0.1
3.1F0.3 2.8F0.3
2.1F0.1 2.9F0.1 2.4F0.1 3.4F0.3
2.6F0.2 n.d. n.d. 3.2F0.2
Values are meansFS.E.M. (n=3–6). n.d.=not determined.
polymerase activity of the Klenow fragment. Treatment with mung bean nuclease almost completely eliminated the 565-bp T-RF from profiles, too (not shown), and thus verified its nature as a pseudo-T-RF (Egert and Friedrich, 2003). However, in contrast to the treatment with Klenow fragment, the peak height ratio of the two remaining T-RFs increased. In agreement with our model on pseudo-T-RF formation (Egert and Friedrich, 2003), single-stranded 16S rRNA gene amplicons have the tendency to form extensive secondary structures and exist probably with partly back-folded, double-stranded stretches. In this conformation, they can display their true terminal restriction fragments, when double-stranded restriction sites are present. After mung bean nuclease digest, however, also these single-stranded amplicons were degraded (Egert and Friedrich, 2003), which might explain the increased peak height ratio observed. Clearly, an adequate estimation of gene ratios based on peak heights (or area) from T-RFLP profiles requires restoration of double strands of amplicons and not removal of part of the amplicon pool. As an example of use, Klenow fragment treatment was applied to 16S rRNA gene amplicons obtained from an environmental sample of known genetic diversity (Egert et al., 2003). The untreated sample displayed three major peaks (Fig. 2B) assignable to amplicons affiliated with Methanobacteriaceae (64 bp) and Crenarchaeota (125 bp), and a 165-bp pseudo-T-RF caused by partially single-stranded
amplicons of both of these groups. Under the given PCR-T-RFLP conditions, a representative Methanobacteriaceae-clone displayed a pseudo-T-RF making up 26% of the height of its true T-RF peak, while a selected Crenarchaeota-clone showed a pseudo-T-RF with a height of 14% of the true T-RF (data not shown). Thus, after total conversion of pseudo-T-RFs into true T-RFs by Klenow fragment treatment, the height of the 64-bp peak was expected to increase 1.26-fold , while the height of the 125-bp T-RF was expected to increase just 1.14-fold (i.e., a decrease of the peak height ratio from 3.1 to 2.8 (Table 1)). This prediction was corroborated by treatment with Klenow fragment prior to restriction digest: the 165-bp pseudo-T-RF (Fig. 2B) was not only strongly reduced but the peak height ratio decreased to 2.6 in addition as expected, while this ratio was virtually unchanged after treatment with mung bean nuclease. The higher degree of variability across replicates of PCR products from the environmental sample compared to the profiles of the template mixture (Table 1) was probably due to more PCR cycles (n=38) applied; previously, the variability in the extent of pseudo-TRF formation was shown to increase with the number of PCR cycles applied (Egert and Friedrich, 2003). Therefore, probably far more than four to six replicates per treatment may be required to prove the statistical significance of the observed changes. In conclusion, treatment of amplicons with Klenow fragment is a simple and efficient way to effectively fill in partially single-stranded PCR products in amplicon pools of mixed template reactions. In case of T-RFLP analysis, this treatment not only helps to identify pseudo-T-RFs in T-RFLP profiles (which can also be achieved by a treatment with mung bean nuclease), but also restores the true T-RFs. This allows in turn a more precise assessment of gene ratios from T-RFLP profiles (Fig. 1). Also other PCR-based methods of genetic diversity analysis, known to be affected by partially single-stranded amplicons, might be improved by Klenow fragment treatment. In DGGE analysis, for instance, (partially) singlestranded amplicons form smeary, poorly stainable extra bands due to an altered electrophoretic mobility and stainability compared to the respective full-length amplicons (Simpson et al., 1999). PCR-based cloning analysis of genetic diversity is likely to be affected as well: partially single-stranded amplicons cannot be
M. Egert, M.W. Friedrich / Journal of Microbiological Methods 61 (2005) 69–75
ligated into a cloning vector and thus, these amplicons will either not be represented in the clone library or will occur at lower frequency. Post-PCR treatment of amplicons with Klenow fragment appears to be a promising tool for a range of PCR-based applications in genetic diversity analysis to more adequately estimate genetic diversity in mixed template samples from the composition of the amplicon pool. However, templates affected by the formation of single-stranded amplicons may be under-represented in the final amplicon pool because they will likely be amplified at lower efficiency due to incomplete (i.e., partly single-stranded) template strands. Thus, more research is required to elucidate mechanisms involved and to design strategies that prevent the formation of partially single-stranded amplicons during PCR.
Acknowledgments We thank Gesche Braker (MPI, Marburg) for critically reading the manuscript. This study was supported by grants from the Deutsche Forschungsgemeinschaft (DFG, Bonn), the Max Planck Society (Munich), and the Fonds der Chemischen Industrie (Frankfurt/Main).
References Bruce, K.D., Hughes, M.R., 2000. Terminal restriction fragment length polymorphism monitoring of genes amplified directly from bacterial communities in soils and sediments. Mol. Biotechnol. 16, 261 – 269. Derbyshire, V., Freemont, P.S., Sanderson, M.R., Beese, L., Friedman, J.M., Joyce, C.M., Steitz, T.A., 1988. Genetic and crystallographic studies of the 3V,5V-exonucleolytic site of DNA polymerase I. Science 240, 199 – 201. Egert, M., Friedrich, M.W., 2003. Formation of pseudo-terminal restriction fragments, a PCR-related bias affecting terminal restriction fragment length polymorphism analysis of microbial community structure. Appl. Environ. Microbiol. 69, 2555 – 2562. Egert, M., Wagner, B., Lemke, T., Brune, A., Friedrich, M.W., 2003. Microbial community structure in midgut and hindgut of
75
the humus-feeding larva of Pachnoda ephippiata (Coleoptera: Scarabaeidae). Appl. Environ. Microbiol. 69, 6659 – 6668. Jensen, M.A., 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 Methods Appl. 3, 186 – 194. Kanagawa, T., 2003. Bias and artifacts in multitemplate polymerase chain reactions (PCR). J. Biosci. Bioeng. 96, 317 – 323. Kitts, C.L., 2001. Terminal restriction fragment patterns: a tool for comparing microbial communities and assessing community dynamics. Curr. Issues Intest. Microbiol. 2, 17 – 25. Lueders, T., Friedrich, M.W., 2003. Evaluation of PCR amplification bias by T-RFLP analysis of SSU rRNA and mcrA genes using defined template mixtures of methanogenic pure cultures and soil DNA extracts. Appl. Environ. Microbiol. 69, 320 – 326. Mathieu-Daude, F., Welsh, J., Vogt, T., McClelland, M., 1996. DNA rehybridization during PCR: The dC(0)t effectT and its consequences. Nucleic Acids Res. 24, 2080 – 2086. Nishigaki, K., Kaneko, Y., Wakuda, H., Husimi, Y., Tanaka, T., 1985. Type-II restriction endonucleases cleave single-stranded DNAs in general. Nucleic Acids Res. 13, 5747 – 5760. Polz, M.F., Cavanaugh, C.M., 1998. Bias in template-to-product ratios in multitemplate PCR. Appl. Environ. Microbiol. 64, 3724 – 3730. Qiu, X.Y., Wu, L.Y., Huang, H.S., McDonel, P.E., Palumbo, A.V., Tiedje, J.M., Zhou, J.Z., 2001. Evaluation of PCR-generated chimeras: mutations, and heteroduplexes with 16S rRNA genebased cloning. Appl. Environ. Microbiol. 67, 880 – 887. Sambrook, J., Russel, D., 2001. Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor. Simpson, J.M., McCracken, V.J., White, B.A., Gaskins, H.R., Mackie, R.I., 1999. Application of denaturant gradient gel electrophoresis for the analysis of the porcine gastrointestinal microbiota. J. Microbiol. Methods 36, 167 – 179. Speksnijder, A.G.C.L., Kowalchuk, G.A., De Jong, S., Kline, E., Stephen, J.R., Laanbroek, H.J., 2001. Microvariation artifacts introduced by PCR and cloning of closely related 16S rRNA gene sequences. Appl. Environ. Microbiol. 67, 469 – 472. Suzuki, M.T., Giovannoni, S.J., 1996. Bias caused by template annealing in the amplification of mixtures of 16S rRNA genes by PCR. Appl. Environ. Microbiol. 62, 625 – 630. Thompson, J.R., Marcelino, L.A., Polz, M.F., 2002. Heteroduplexes in mixed-template amplifications: formation, consequence and elimination by dreconditioning PCRT. Nucleic Acids Res. 30, 2083 – 2088. von Wintzingerode, F., Goebel, U.B., Stackebrandt, E., 1997. Determination of microbial diversity in environmental samples: pitfalls of PCR-based rRNA analysis. FEMS Microbiol. Rev. 21, 213 – 229.
Post-amplification Klenow fragment treatment alleviates PCR bias caused by partially single-stranded amplicons Markus Egert1, Michael W. Friedrich* Max Planck Institute for Terrestrial Microbiology, Karl-von-Frisch-Strabe, 35043 Marburg, Germany Received 12 July 2004; received in revised form 2 November 2004; accepted 2 November 2004 Available online 30 November 2004
Abstract Partially single-stranded amplicons, formed during PCR amplification of single and mixed templates, are a potential source of bias in genetic diversity studies. The analysis of 16S rRNA gene diversity in mixed template samples by the fingerprinting technique terminal restriction fragment length polymorphism (T-RFLP) analysis can be biased by the occurrence of pseudo-TRFs, i.e., restriction fragments occurring in addition to the expected terminal restriction fragments of single amplicons. This bias originates from PCR products, which are single-stranded at their terminal restriction site. Here we show that treatment of PCR amplicons with Klenow fragment prior to restriction digest and T-RFLP analysis minimized effectively the occurrence of pseudo-T-RFs. Klenow fragment activity filled in bases into the partially single-stranded amplicons and thereby restored the affected amplicons to complete double strands. Our method allowed to improve the assessment of genetic diversity and gene ratios from T-RFLP analysis of an original environmental sample. Since partially single-stranded amplicons might influence many PCR-based techniques, post-amplification treatment with Klenow fragment may be useful for a wide range of applications, which assess the composition of amplicon pools, e.g., the analysis of marker gene diversity in mixed template samples by fingerprinting techniques or the analysis of sequence diversity by cloning. D 2004 Elsevier B.V. All rights reserved. Keywords: Genetic diversity analyses; Molecular fingerprinting; Single-stranded amplicons; T-RFLP
1. Introduction
* Corresponding author. Tel.: +49 6421 178830; fax: +49 6421 178809. E-mail address: [email protected] (M.W. Friedrich). 1 Present address: Laboratory of Microbiology, Wageningen University, Hesselink van Suchtelenweg 4, 6703 CT Wageningen, The Netherlands. 0167-7012/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.mimet.2004.11.002
The molecular analysis of genetic diversity in mixed template samples can be biased by several artifacts originating from PCR amplification (for reviews, see Kanagawa, 2003; von Wintzingerode et al., 1997). Sequence errors introduced by the DNA polymerase, the formation of chimeric PCR products, and the conversion of cloned heteroduplex amplicons
70
M. Egert, M.W. Friedrich / Journal of Microbiological Methods 61 (2005) 69–75
into hybrid homoduplex molecules by the mismatch repair system of the host cells can cause an overestimation of the number of sequence types present in the original sample (Qiu et al., 2001; Speksnijder et al., 2001; Thompson et al., 2002). Moreover, differences in primer binding energy (Lueders and Friedrich, 2003; Polz and Cavanaugh, 1998) and template re-annealing (Mathieu-Daude et al., 1996; Suzuki and Giovannoni, 1996) during later amplification cycles
might cause differences of original template ratios compared to their representation in the amplicon pool after amplification. Recently, we discovered that after PCR amplification of 16S rRNA genes from single and mixed template samples, a significant proportion of amplicons can be partially single-stranded (Egert and Friedrich, 2003), which can markedly affect subsequent genetic diversity analysis, e.g., by terminal restriction fragment length polymorphism (T-RFLP)
Fig. 1. Scheme illustrating the effect of pre-treatment of amplicons originating from a single template with mung bean nuclease (B) or Klenow fragment (C) vs. the untreated control (A) on the resulting T-RFLP profiles. Partly single-stranded amplicons, formed during PCR-amplification, lead to the formation of a pseudo-T-RF (red rectangle in schematic profiles), representing a restriction fragment longer than the true, terminal RF (black rectangle in schematic profiles). Pseudo-T-RFs form due to the inability of restriction endonucleases to cut amplicons, which are singlestranded (ss) at their terminal restriction sites (terminal restriction site as black square in schematic amplicons; second restriction site in red; actual cutting site indicated by triangle). Treatment with single-strand degrading mung bean nuclease prior to the restriction digest eliminates all single-stranded amplicons, and the pseudo-T-RF in the profile, but results in underestimation of the relative frequency of amplicons displaying pseudo-T-RF formation. In contrast, by treatment with Klenow fragment prior to the restriction digest, single-stranded amplicons are converted into fully double-stranded ones (newly synthesized bases of the amplicon in green), and thus, all amplicons are rendered accessible for endonucleases at their terminal restriction sites, which is a prerequisite for amplicon quantification in molecular diversity analysis. Note: TRFLP analyses allow to assess the relative (not the absolute) gene frequencies based on relative fluorescence units (RFU), by referring height or area of selected T-RF peaks to the peak height or area of the total profile or by comparing the ratio of peak heights or areas of selected peaks (see Table 1). FAM: 6V-carboxyfluorescein fluorescent label of amplicons.
M. Egert, M.W. Friedrich / Journal of Microbiological Methods 61 (2005) 69–75
profiling (Egert and Friedrich, 2003), denaturing gradient gel electrophoresis (DGGE; Simpson et al., 1999), and possibly also by cloning and sequencing of randomly selected clones. In case of T-RFLP analysis, i.e., a molecular fingerprinting technique widely used to rapidly compare genetic diversity and dynamics among different samples or treatments (for reviews see (Bruce and Hughes, 2000; Kitts, 2001), amplicons which are single-stranded at their terminal restriction site favor the formation of restriction fragments (RFs) longer than the true, expected terminal-RF (representing the second or subsequent restriction sites), because type II restriction enzymes cut double-stranded DNA only (Nishigaki et al., 1985). If not identified, these additional bpseudo-terminal restriction fragmentsQ (pseudo-T-RFs) can lead to an overestimation of genetic diversity in the original sample (Fig. 1A). Treatment of amplicons with single-strand degrading mung bean nuclease prior to T-RFLP analysis effectively eliminates pseudo-T-RFs from T-RFLP profiles (Egert and Friedrich, 2003). However, mung bean digest will eliminate all amplicons affected and thus renders these amplicons unavailable for estimation of relative frequencies by T-RF peak quantification (Lueders and Friedrich, 2003), resulting in a severe underestimation of genes represented by these amplicons (Fig. 1B). Consequently, this PCR bias may be solved much more satisfactorily by preventing the formation of partially single-stranded amplicons. However, the molecular mechanism underlying the formation of partly single-stranded amplicons is not clear. Temporary template secondary structures (Jensen and Straus, 1993) might cause the DNA polymerase to fall off during amplification prior to complete synthesis of the complementary strand (Polz and Cavanaugh, 1998). So far, a lower number of PCR cycles was shown to alleviate the extent of pseudo-TRF formation (Egert and Friedrich, 2003). However, partially single-stranded amplicons clearly occurred even with low cycles numbers (b20), and particularly in case of environmental samples, usually more PCR cycles (~25–30) are required to obtain a PCR signal sufficient for downstream analyses, e.g., fingerprinting. Several other strategies to minimize this bias were not successful (Egert and Friedrich, 2003). Neither varying conditions favoring primer/template annealing (i.e., annealing temperature, ionic strength of the
71
PCR buffer) nor variations of the primer ratios to exclude asymmetric primer utilization (Jensen and Straus, 1993), nor breconditioning PCRQ, i.e., a low cycle number re-amplification of a slightly diluted mixed-template PCR-product (Thompson et al., 2002), were effective in preventing artifactual amplicons (Egert and Friedrich, unpublished results). Here, we show that post-PCR treatment with Klenow fragment effectively reduces the amount of partially single-stranded PCR products in an amplicon pool, irrespective of their mode of formation during PCR. Klenow fragment, the large subunit of the DNA polymerase I of E. coli, is widely used in molecular biology, e.g., to produce DNA molecules with blunt ends from those with sticky ends by filling in complementary bases (Sambrook and Russel, 2001). We used Klenow fragment to convert partially singlestranded stretches of amplicons to the respective double-stranded molecules, rendering their terminal restriction sites accessible for restriction endonucleases (Fig. 1C), which is a prerequisite for a reliable estimation of genetic diversity in the original samples by T-RFLP analysis. Since also other PCR-based techniques in genetic diversity analysis are potentially affected by partially single-stranded PCR products (e.g., cloning and DGGE, (Simpson et al., 1999), post-PCR Klenow fragment treatment may improve original template representation in amplicon pools of PCR-based applications in general.
2. Materials and methods 2.1. Defined template mixture PCR-T-RFLP analysis was performed with a mixture of two clonal M13 products, both of which contained bacterial 16S rRNA gene inserts cloned from a beetle larva gut (Egert et al., 2003). Clone PeH55 (accession number AJ576368) displayed an expected T-RF of 72- and a 565-bp pseudo-T-RF, and was designated bpseudo-T-RF templateQ. Clone PeH59 (AJ538351), showed the expected 81-bp TRF only and no pseudo-T-RF (Fig. 2A), and was designated breference templateQ. PCR conditions and T-RFLP parameters were described in detail previously (Egert and Friedrich, 2003). Briefly, bacterial 16S rRNA gene fragments (~900 bp) were PCR
72
M. Egert, M.W. Friedrich / Journal of Microbiological Methods 61 (2005) 69–75
Fig. 2. Representative 16S rRNA gene T-RFLP profiles of (A) a clonal template mixture, comprising a pseudo-T-RF producing template and a reference template, and (B) an environmental sample, before and after treatment of amplicons with Klenow fragment prior to restriction digest. Treatment with Klenow fragment strongly reduced the respective pseudo-T-RF peaks (565 bp (A) and 165 bp (B), respectively) and simultaneously increased the height of the respective true and terminal restriction fragments (72 bp (A) and 64 and 125 bp (B)), measurable as changes in the peak height ratios of the 81 and 72 bp T-RFs (A) and the 125 and 64 bp T-RFs (B) (see Table 1). C: Crenarchaeota, MB: Methanobacteriaceae, RFU: relative fluorescence units.
amplified (32 cycles; ~25 ng of template DNA in a 50-Al standard PCR reaction mixture) using a primer pair specific for all Bacteria, of which the forward primer was fluorescently labeled. Purified amplicons (~75 ng) were digested with MspI, and restriction fragments were subsequently size-separated on an ABI373A automated sequencer in GeneScan mode. 2.2. Enzymatic treatments Amplicon pools were subjected to different treatments prior to T-RFLP analysis, i.e., before restriction digest. Treatments with mung bean nuclease to eliminate partially single-stranded amplicons were performed as described by Egert and Friedrich
(2003). Amplicons were treated with Klenow fragment to fill in bases into single strands as follows: ~1 Ag of PCR product was incubated with 5 U of Klenow fragment exonuclease minus (Promega, Mannheim, Germany), 10 Al of 10 reaction buffer, and 50 AM of each of the four deoxynucleoside triphosphates (dNTPs) in a total volume of 100 Al for 1 h at 20 8C in the dark. Klenow fragment exonuclease minus, which lacks the 3V-5Vexonuclease (proof-reading) activity (Derbyshire et al., 1988), was used to exclude any exonuclease activity from the incubations. The specificity of the treatment, i.e., DNA polymerase activity of the Klenow fragment, was checked by either incubating amplicons without Klenow fragment or replacing dNTPs by 100 AM of each of the four
M. Egert, M.W. Friedrich / Journal of Microbiological Methods 61 (2005) 69–75
dideoxynucleosid triphosphates (ddNTPS; Amersham Pharmacia Biotech., Freiburg, Germany), which terminates nucleoside incorporation by polymerase activity. All incubations were terminated by purification of amplicons with the MinElute PCR purification kit (Qiagen, Hilden, Germany).
73
were checked pair wise for statistical significance ( Pb0.05) with non-parametric Mann–Whitney tests using SYSTAT (SPSS, Chicago, USA). All treatments were replicated three to six times, based on independent PCR amplifications from a single template mixture or the environmental DNA extract, respectively.
2.3. Environmental sample 3. Results and discussion 16S rRNA gene fragments were amplified (38 cycles, 1 Al of 1:30 diluted DNA extract as template in a 50 Al standard PCR reaction mixture) from an environmental DNA extract (originating from the midgut of a beetle larva, (Egert et al., 2003) using an Archaea-specific primer pair, with the reverse primer fluorescently labeled. Mung bean nuclease and Klenow fragment treatments prior to restriction digest and T-RFLP analysis were performed as described above using AluI as restriction enzyme. The effect of Klenow fragment treatment on the T-RFLP profile of the original sample was estimated in addition using two archaeal clones, which represent the most abundant phylogenetic groups in the original sample. For this, 16S rRNA gene fragments (~800 bp) of clone PeMAr19 (AJ576131), representing Methanobacteriaceae, and clone PeMAr04 (AJ538352), representing Crenarchaeota, were amplified using ~5 ng of clonal M13 products as template in a standard PCR reaction with 38 cycles. Both clones displayed a 165bp pseudo-T-RF in addition to their true terminal RFs sized 64 and 125 bp, respectively (Egert and Friedrich, 2003). 2.4. Evaluation of treatments We used T-RFLP analysis to quantify the formation of partially single-stranded 16S rRNA gene amplicons. Effects of treatments were quantified by comparing the ratio of the heights of selected peaks in profiles before and after the treatments. In case of the defined template mixture, the height of the 81-bp T-RF of the reference template was compared to the height of the true T-RF peak (72 bp) of the pseudo-TRF template. In case of the environmental sample, changes in the peak height ratio of the true T-RFs representing Crenarchaeota (125 bp) and Methanobacteriaceae(64 bp), respectively, were followed. Differences in peak height ratios between treatments
Treatment of the amplicon pool with Klenow fragment prior to restriction digest almost completely removed the respective pseudo-T-RF of 565 bp from the T-RFLP profile of the template mixture (Fig. 2A), which indicates that partially single-stranded amplicons were filled in with bases at their ends, and thus, the double strand of these amplicons was restored. If true, the disappearance of the pseudo-T-RF should be accompanied by a respective increase of the true 72bp terminal restriction fragment of the pseudo-T-RF template. This was tested by quantifying changes in peak heights of the 72-bp T-RF of the pseudo-T-RF template relative to that of the 81-bp T-RF of the reference template by calculating ratios after T-RFLP analysis. Assuming a total conversion of partially single-stranded amplicons into double-strands, the peak height of the true terminal 72-bp RF was expected to equal the sum of heights of the 72-bp TRF and the former 565-bp pseudo-T-RF. Because the 81-bp T-RF of the reference template should not be affected by the Klenow fragment treatment, the T-RF peak height ratio was expected to decrease from 2.6 to 1.9 (Table 1). The latter ratio represents the real 16S rRNA gene ratio of the two clones in the amplicon pool. The effect of the different treatments on the suppression of pseudo-T-RF formation is given in Table 1. In fact, treatment of the amplicons with Klenow fragment prior to T-RFLP analysis not only eliminated the 565-bp T-RF of the pseudo-T-RF template almost completely (Fig. 2A) but also significantly reduced the peak height ratio to 2.1, which was in good agreement with the ratio predicted for a total conversion of the pseudo-T-RFs into true terminal RFs. Neither incubation of amplicons without Klenow fragment nor with ddNTPS instead of dNTPs did significantly change the peak height ratio in comparison to the ratio of the control profiles, indicating that the observed change was due to DNA
74
M. Egert, M.W. Friedrich / Journal of Microbiological Methods 61 (2005) 69–75
Table 1 Influence of different treatments of amplicons prior to restriction digest and T-RFLP analysis on the peak heights ratio of selected TRFs from 16S rRNA gene profiles of a defined clonal mixture, comprising a pseudo-T-RF producing template and a reference template (Fig. 2A), and an environmental DNA extract (Fig. 2B) Treatment
Peak height ratio Template mixture
Environmental sample
81 vs. 72 bp T-RF 125 vs. 64 bp T-RF Untreated Calculated conversion of pseudo-T-RFs into true T-RFs Klenow fragment Klenow fragment +ddNTPs Mung bean nuclease
2.6F0.1 1.9F0.1
3.1F0.3 2.8F0.3
2.1F0.1 2.9F0.1 2.4F0.1 3.4F0.3
2.6F0.2 n.d. n.d. 3.2F0.2
Values are meansFS.E.M. (n=3–6). n.d.=not determined.
polymerase activity of the Klenow fragment. Treatment with mung bean nuclease almost completely eliminated the 565-bp T-RF from profiles, too (not shown), and thus verified its nature as a pseudo-T-RF (Egert and Friedrich, 2003). However, in contrast to the treatment with Klenow fragment, the peak height ratio of the two remaining T-RFs increased. In agreement with our model on pseudo-T-RF formation (Egert and Friedrich, 2003), single-stranded 16S rRNA gene amplicons have the tendency to form extensive secondary structures and exist probably with partly back-folded, double-stranded stretches. In this conformation, they can display their true terminal restriction fragments, when double-stranded restriction sites are present. After mung bean nuclease digest, however, also these single-stranded amplicons were degraded (Egert and Friedrich, 2003), which might explain the increased peak height ratio observed. Clearly, an adequate estimation of gene ratios based on peak heights (or area) from T-RFLP profiles requires restoration of double strands of amplicons and not removal of part of the amplicon pool. As an example of use, Klenow fragment treatment was applied to 16S rRNA gene amplicons obtained from an environmental sample of known genetic diversity (Egert et al., 2003). The untreated sample displayed three major peaks (Fig. 2B) assignable to amplicons affiliated with Methanobacteriaceae (64 bp) and Crenarchaeota (125 bp), and a 165-bp pseudo-T-RF caused by partially single-stranded
amplicons of both of these groups. Under the given PCR-T-RFLP conditions, a representative Methanobacteriaceae-clone displayed a pseudo-T-RF making up 26% of the height of its true T-RF peak, while a selected Crenarchaeota-clone showed a pseudo-T-RF with a height of 14% of the true T-RF (data not shown). Thus, after total conversion of pseudo-T-RFs into true T-RFs by Klenow fragment treatment, the height of the 64-bp peak was expected to increase 1.26-fold , while the height of the 125-bp T-RF was expected to increase just 1.14-fold (i.e., a decrease of the peak height ratio from 3.1 to 2.8 (Table 1)). This prediction was corroborated by treatment with Klenow fragment prior to restriction digest: the 165-bp pseudo-T-RF (Fig. 2B) was not only strongly reduced but the peak height ratio decreased to 2.6 in addition as expected, while this ratio was virtually unchanged after treatment with mung bean nuclease. The higher degree of variability across replicates of PCR products from the environmental sample compared to the profiles of the template mixture (Table 1) was probably due to more PCR cycles (n=38) applied; previously, the variability in the extent of pseudo-TRF formation was shown to increase with the number of PCR cycles applied (Egert and Friedrich, 2003). Therefore, probably far more than four to six replicates per treatment may be required to prove the statistical significance of the observed changes. In conclusion, treatment of amplicons with Klenow fragment is a simple and efficient way to effectively fill in partially single-stranded PCR products in amplicon pools of mixed template reactions. In case of T-RFLP analysis, this treatment not only helps to identify pseudo-T-RFs in T-RFLP profiles (which can also be achieved by a treatment with mung bean nuclease), but also restores the true T-RFs. This allows in turn a more precise assessment of gene ratios from T-RFLP profiles (Fig. 1). Also other PCR-based methods of genetic diversity analysis, known to be affected by partially single-stranded amplicons, might be improved by Klenow fragment treatment. In DGGE analysis, for instance, (partially) singlestranded amplicons form smeary, poorly stainable extra bands due to an altered electrophoretic mobility and stainability compared to the respective full-length amplicons (Simpson et al., 1999). PCR-based cloning analysis of genetic diversity is likely to be affected as well: partially single-stranded amplicons cannot be
M. Egert, M.W. Friedrich / Journal of Microbiological Methods 61 (2005) 69–75
ligated into a cloning vector and thus, these amplicons will either not be represented in the clone library or will occur at lower frequency. Post-PCR treatment of amplicons with Klenow fragment appears to be a promising tool for a range of PCR-based applications in genetic diversity analysis to more adequately estimate genetic diversity in mixed template samples from the composition of the amplicon pool. However, templates affected by the formation of single-stranded amplicons may be under-represented in the final amplicon pool because they will likely be amplified at lower efficiency due to incomplete (i.e., partly single-stranded) template strands. Thus, more research is required to elucidate mechanisms involved and to design strategies that prevent the formation of partially single-stranded amplicons during PCR.
Acknowledgments We thank Gesche Braker (MPI, Marburg) for critically reading the manuscript. This study was supported by grants from the Deutsche Forschungsgemeinschaft (DFG, Bonn), the Max Planck Society (Munich), and the Fonds der Chemischen Industrie (Frankfurt/Main).
References Bruce, K.D., Hughes, M.R., 2000. Terminal restriction fragment length polymorphism monitoring of genes amplified directly from bacterial communities in soils and sediments. Mol. Biotechnol. 16, 261 – 269. Derbyshire, V., Freemont, P.S., Sanderson, M.R., Beese, L., Friedman, J.M., Joyce, C.M., Steitz, T.A., 1988. Genetic and crystallographic studies of the 3V,5V-exonucleolytic site of DNA polymerase I. Science 240, 199 – 201. Egert, M., Friedrich, M.W., 2003. Formation of pseudo-terminal restriction fragments, a PCR-related bias affecting terminal restriction fragment length polymorphism analysis of microbial community structure. Appl. Environ. Microbiol. 69, 2555 – 2562. Egert, M., Wagner, B., Lemke, T., Brune, A., Friedrich, M.W., 2003. Microbial community structure in midgut and hindgut of
75
the humus-feeding larva of Pachnoda ephippiata (Coleoptera: Scarabaeidae). Appl. Environ. Microbiol. 69, 6659 – 6668. Jensen, M.A., 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 Methods Appl. 3, 186 – 194. Kanagawa, T., 2003. Bias and artifacts in multitemplate polymerase chain reactions (PCR). J. Biosci. Bioeng. 96, 317 – 323. Kitts, C.L., 2001. Terminal restriction fragment patterns: a tool for comparing microbial communities and assessing community dynamics. Curr. Issues Intest. Microbiol. 2, 17 – 25. Lueders, T., Friedrich, M.W., 2003. Evaluation of PCR amplification bias by T-RFLP analysis of SSU rRNA and mcrA genes using defined template mixtures of methanogenic pure cultures and soil DNA extracts. Appl. Environ. Microbiol. 69, 320 – 326. Mathieu-Daude, F., Welsh, J., Vogt, T., McClelland, M., 1996. DNA rehybridization during PCR: The dC(0)t effectT and its consequences. Nucleic Acids Res. 24, 2080 – 2086. Nishigaki, K., Kaneko, Y., Wakuda, H., Husimi, Y., Tanaka, T., 1985. Type-II restriction endonucleases cleave single-stranded DNAs in general. Nucleic Acids Res. 13, 5747 – 5760. Polz, M.F., Cavanaugh, C.M., 1998. Bias in template-to-product ratios in multitemplate PCR. Appl. Environ. Microbiol. 64, 3724 – 3730. Qiu, X.Y., Wu, L.Y., Huang, H.S., McDonel, P.E., Palumbo, A.V., Tiedje, J.M., Zhou, J.Z., 2001. Evaluation of PCR-generated chimeras: mutations, and heteroduplexes with 16S rRNA genebased cloning. Appl. Environ. Microbiol. 67, 880 – 887. Sambrook, J., Russel, D., 2001. Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor. Simpson, J.M., McCracken, V.J., White, B.A., Gaskins, H.R., Mackie, R.I., 1999. Application of denaturant gradient gel electrophoresis for the analysis of the porcine gastrointestinal microbiota. J. Microbiol. Methods 36, 167 – 179. Speksnijder, A.G.C.L., Kowalchuk, G.A., De Jong, S., Kline, E., Stephen, J.R., Laanbroek, H.J., 2001. Microvariation artifacts introduced by PCR and cloning of closely related 16S rRNA gene sequences. Appl. Environ. Microbiol. 67, 469 – 472. Suzuki, M.T., Giovannoni, S.J., 1996. Bias caused by template annealing in the amplification of mixtures of 16S rRNA genes by PCR. Appl. Environ. Microbiol. 62, 625 – 630. Thompson, J.R., Marcelino, L.A., Polz, M.F., 2002. Heteroduplexes in mixed-template amplifications: formation, consequence and elimination by dreconditioning PCRT. Nucleic Acids Res. 30, 2083 – 2088. von Wintzingerode, F., Goebel, U.B., Stackebrandt, E., 1997. Determination of microbial diversity in environmental samples: pitfalls of PCR-based rRNA analysis. FEMS Microbiol. Rev. 21, 213 – 229.