OUP user menu

Identification of mixed bacterial DNA contamination in broad-range PCR amplification of 16S rDNA V1 and V3 variable regions by pyrosequencing of cloned amplicons

Niclas Grahn, Margaretha Olofsson, Katarina Ellnebo-Svedlund, Hans-Jürg Monstein, Jon Jonasson
DOI: http://dx.doi.org/10.1016/S0378-1097(02)01190-4 87-91 First published online: 1 February 2003


Using a sensitive and rapid method combining broad-range PCR amplification of bacterial 16S rDNA fragments and pyrosequencing for detection, identification and typing, we have found contaminating bacterial DNA in our reagents used for PCR. Identified bacteria are the water-borne bacterial genera Pseudomonas, Stenotrophomonas, Xanthomonas, Ralstonia and Bacillus. Our results are in concordance with recent reports of contaminated industrial water systems. In light of this conclusion, we believe that there is a need for increased awareness of possible contamination in uncertified widely used molecular biology reagents, including ultra-pure water. Since sequence-based 16S rDNA techniques are used in a variety of settings for bacterial typing and the characterization of microbial communities, we feel that future certification of molecular biology reagents, as free of nucleic acids, would be advantageous.

  • 16S rRNA
  • Pyrosequencing
  • Bacterial DNA contamination

1 Introduction

Characterization of bacteria using 16S rDNA sequence-based molecular techniques has become a common tool for phylogenetic analyses [1], for characterization of bacteria present in different environmental samples [2], and for the detection of contaminating bacteria in groundwater [3] and industrial water systems [45]. These techniques can also be used advantageously for the identification of pathogens in clinical samples [68] and food [9]. For many bacterial species, the discriminatory power of sequence analysis is high, even for very short fragments of the 16S rRNA gene as obtained with the pyrosequencing™ technique [8], which is a fast and reliable sequencing method based on pyrophosphate detection upon nucleotide incorporation in a real-time primer extension assay [10]. Sometimes, bacteria can be both unambiguously characterized to the species level and sub-typed using very short (<30 nucleotides) sequences [11].

The recent reports of bacterial contamination in industrial ultra-pure water systems [4,5] led us to reinvestigate some inconclusive results from our own experiments involving high-throughput broad-range PCR amplification and pyrosequencing of 16S rDNA fragments. We observed that when working with low-copy number starting material (few bacteria) or when increased sensitivity of detection was needed, i.e. increased number of cycles (∼35), contaminating amplicons rendered the results inconclusive (data not shown). The pyrograms could invariably be interpreted as representing mixed sequences and no sequence match could be identified using the BLAST tool [12] in public databases. In search of possible causes of this trouble and in light of the reports by Kulakow et al. [4] and Kawai et al. [5], we have focused in this work on the identification of contaminating bacterial DNA yielding specific amplicons in non-templated (‘negative’) controls for our broad-range 16S rDNA-PCR.

2 Materials and methods

2.1 Broad-range PCR of 16S rDNA fragments

Broad-range PCR was performed using puReTaq™ Ready-To-Go™ PCR beads (Amersham Biosciences, Uppsala, Sweden) in a reaction volume of 25 µl and 5 pmol of each primer. The Ready-To-Go PCR beads are lyophilized beads containing all necessary reagents for PCR except primers and template. Primers used were purchased from Scandinavian Gene Synthesis (Köping, Sweden) and were as given in Table 1 together with specification of region of the 16S rRNA gene amplified. PCR conditions were as follows: denaturing step at 95°C for 10 min, 35 cycles of 40 s denaturing at 95°C, 40 s annealing at 55°C and 1 min extension at 72°C, followed by a final extension step of 10 min at 72°C. No template controls were used where no DNA was added; positive controls using genomic Helicobacter pylori 26695 DNA were used for reference. PCR amplicons were analyzed by agarose gel electrophoresis. The gel was stained with ethidium bromide and amplicons were visualized using UV trans-illumination.

View this table:
Table 1

Oligonucleotide sequences, 16S rDNA amplified region, and sizes of the 16S rDNA PCR amplicons described in this study

Primer setPrimer nameSequence 5′ to 3′ orientationAmplified regionAmplicon length (bp)
1pBR5′.segaa gag ttt gat cat ggc tca gNear full length
p13B.asgtg tac tag gcc cgg gaa cgt att c16S rDNA1391
2pBR5′.seas above
pBR-V1.astta ctc acc cgt ccg cca ctV1 region115
3pJBS-V3.segca acg cga aga acc tta cc
B-V3.asacg aca gcc atg cag cac ctV3 region105
4Bio-pBR5′.seBiotin-gaa gag ttt gat cat ggc tca g
pBR-V1.asas aboveV1 → PSQ115
5pJBS.V3.seas above
Bio-B-V3.asBiotin-acg aca gcc atg cag cac ctV3→PSQ105
6M13(–20)Fgta aaa cga cgg cca gInserted fragment
M13Rcag gaa aca gct atg acin pCRII®-TOPO®241+insert
  • Primers are from [8].

  • Primer is from [15].

  • PCR amplicons were used for pyrosequencing.

2.2 Cloning of 16S rDNA amplicons and retrieval of cloned fragments

Cloning of PCR amplicons was performed using a TOPO T/A cloning® kit (Invitrogen, Groningen, the Netherlands) according to the manufacturer's protocol using 4 µl PCR products as starting material, pCRII®-TOPO® as cloning vector and chemically competent TOP 10 Escherichia coli cells for the transformation. From each transformation 15 colonies were picked and then subjected to retrieval of the inserted fragments according to the manufacturer's protocol using vector-specific primers. Insert-specific primers could not be used because they would also target the 16S rRNA genes of the E. coli used for transformation, which might spoil the cloning experiment. PCR amplicons were purified using a GFX™ Gel band and PCR purification kit (Amersham Biosciences) before PCR amplification for pyrosequencing and sequence analysis.

2.3 PCR using biotinylated primers and pyrosequencing analysis

For analysis by pyrosequencing the amplicons need to be biotinylated at one end. GFX purified amplicons (1 µl) were used as template in PCR using one biotinylated primer and a regular primer according to Table 1. PCR conditions were as stated earlier but with the number of cycles lowered to 25 cycles. Pyrosequencing analysis was performed using a PSQ96 MA and the new and improved SQA reagents kit (Pyrosequencing, Uppsala, Sweden) and a method of cyclic dispensation of nucleotides [8]. Sample preparation was according to the manufacturer and as described [11]. The results from the pyrosequencing analysis were analyzed using a signature-matching algorithm [8], and bacteria were identified using either catalogued sequences in GenBank [12], with the BLAST tool [13] at the National Center for Biotechnology Information (NCBI), or compared to a local database of 16S rDNA sequences.

3 Results and discussion

The steps outlined in the previous sections for the identification of contaminating bacterial DNA started with an initial 35 cycles broad-range PCR amplification of 16S rDNA variable V1 and V3 regions, which yielded specific amplicons of the expected sizes for bacterial DNA in all of the non-templated ‘negative’ controls for all primer sets used (see Table 1 for reference). The results after agarose gel electrophoresis are shown in Fig. 1. Direct pyrosequencing of these amplicons yielded mixed sequences with the result that no bacteria could be identified (Fig. 2, top panel). To obtain more detail, cloning of the PCR amplicons was performed using a TOPO T/A vector. Pyrosequencing of the 16S rDNA inserts after this cloning led to the identification of sequences unambiguously representing the bacterial genera Pseudomonas, Stenotrophomonas, Xanthomonas, Ralstonia and Bacillus. Pyrosequencing of the variable V3 region was the most informative (Table 2). Outputs from pyrosequencing analysis can be viewed in Fig. 2. The identified bacteria are all usually water-borne and the results are in agreement with those of Kulakow et al. [4] and Kawai et al. [5]. One may therefore safely conclude that the contaminating bacterial DNA emanated neither from random carry-over of PCR amplicons in the laboratory nor did they represent human detritus. In such case one would have expected an entirely different bacterial flora.

Figure 1

Ethidium bromide stained agarose gel showing fragments of the expected sizes using the three primer combinations covering A: the near full length, B: the V1 region and C: the V3 region of 16S rDNA. M: 100 base pair size marker ladder. Arrow indicates the position of an 800 base pair band. Non-templated negative controls are in lanes 1 and 2 followed by the positive control (H. pylori DNA template) in lane 3.

Figure 2

Identification of bacterial 16S rDNA gene fragments. A: Representative pyrogram of direct pyrosequence analysis of variable V3 region PCR-amplicons from a negative control. B, C: Pyrograms displaying sequences identified as Bacillus sp. and Xanthomonas sp., respectively. D: DNA sequence alignment derived from B and C.

View this table:
Table 2

Bacterial genera characterized by 16S rDNA V1 and V3 region pyrosequencing

16S rDNA regionBacterial species identified
V1 variable regionStenotrophomonas (7/15)
Pseudomonas (2/15)
n.d. (6/15)
V3 variable regionStenotrophomonas (4/15)
Ralstonia (4/15)
Bacillus (4/15)
Xanthomonas (3/15)
  • n.d. denotes not determinable sequences.

  • Number of clones identified in 15 clones examined.

Therefore, the results presented here provide proof that low levels of contaminating bacterial 16S rDNA fragments were present in either our ultra-pure water system or had been distributed with the purchased molecular biology reagents or disposable plastic articles. Attempts to culture bacteria from our ultra-pure water system have been unsuccessful, which unfortunately will not exclude the possibility of contaminating DNA fragments. Obviously, the problems concerning bacterial DNA contamination of water as suggested here are not pyrosequencing-related. All methods based on broad-range PCR for detection and identification of low levels of bacteria will be similarly affected. Clearly, bacterial DNA contamination of water may arise accidentally; the water supplies of some PCR reagent manufacturer could well be the source, which would cause problems in many situations, especially when using samples and matrices with a low occurrence of bacteria, e.g. in the detection and characterization of bacteria in human biopsy specimens. However, these low background levels are usually not detected and seem not to interfere with bacterial typing when working with high amounts of template, e.g. DNA isolated from bacterial cultures, and a relatively low number of amplification cycles.

In the present case, the source of the contamination can be argued. We point in no specific direction except that the bacterial DNA contamination seems most likely to have been water-borne and maybe represents a rather underestimated cause of falsely primed PCR reactions [14]. Therefore, we want to raise our voices for an increased awareness of this phenomenon. We feel that the need for reagents certified as free of nucleic acids, not only as free of nucleases, will increase with new higher demands for sensitivity and specificity of analyses performed in molecular biology laboratories around the world. This is important not only in a clinical routine diagnostic setting but also for those laboratories performing microbiological analyses of environmental samples, foods or drinking water. A parallel can be drawn to forensic science, where contaminating eukaryotic nucleic acids during DNA isolation procedures have to be carefully avoided. Consequently, there are companies that market reagents as certified free of eukaryotic nucleic acids for the intended use in a forensic setting. Proper identification of low levels of bacteria and other microorganisms in tissue fragments by broad-range PCR might seem equally important. However, consensus on the need for quality certification of reagents used in molecular biology seems necessary before one would be able to fully take advantage of the inherent power of these broad-range PCR techniques.


The LMÖ Molecular Biology program supported this work.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
View Abstract