42
A. Meyerdierks and F.O. Glöckner
2.3.4 Amplification of Genomic DNA
Comprehensive metagenome analyses generally require microgram amounts of
genomic DNA in order to be representative (see below). Theoretically, less DNA
would be sufficient for a representative metagenome analysis, if the microbial diversity in the sample is low or strategies have been applied to reduce the diversity and
to enrich for genomic DNA of target cell populations. However, standard metagenomic library construction and sequencing techniques are not compatible with
DNA amounts much below the microgram level. In some studies, large amounts
of genomic DNA cannot be obtained, e.g., from the deep biosphere (Webster et al.
2003, Biddle et al. 2008). These small quantities of starting DNA need to be amplified before metagenome analysis. To overcome this problem, different genomic
DNA amplification methods have been developed (Telenius et al. 1992, Zhang
et al. 1992, Breitbart et al. 2002, Breitbart and Rohwer 2005, Pinard et al. 2006)
of which one, the multiple displacement amplification (MDA) with bacteriophage
φ29 polymerase, is now widely used in metagenome analyses.
The DNA polymerase of bacteriophage φ29 from Bacillus subtilis is required
for the replication of the 19,285 bp phage genome (Blanco and Salas 1985b). In
addition to its DNA polymerase activity, this single polypeptide of 66,520 dalton
(Blanco and Salas 1984, Watabe et al. 1984) has a 3 → 5 exonuclease activity
(Blanco and Salas 1985a). The error rate of the φ29 polymerase (∼10 –5 to 10 –7 )
(Esteban et al. 1993, Nelson et al. 2002, Zhang et al. 2006) is about 10–100-fold
lower than the mutant frequency determined for Taq DNA polymerase (Eckert and
Kunkel 1990) and the processivity is high. The replication of the bacteriophage φ29
genome in a protein primed reaction was determined to last 8 min (Blanco et al.
1989). The φ29 polymerase can also elongate DNA-primed reactions. The circular
single stranded M13 genome (∼7,250 bases) was isothermally replicated in 5 min,
producing a strand of more than 70 kb after 40 min, due to the strand displacement
capability of the φ29 polymerase (Blanco et al. 1989). This strand displacement
capability leads to a successive detachment of the 5 -end of the synthesised DNA
molecule from the template strand by the φ29 polymerase during DNA synthesis, resulting in a rolling circle amplification (RCA) (Fire and Xu 1995, Liu et al.
1996).
In (meta)genome analyses, the φ29 DNA polymerase is used in a multiplyprimed amplification with 3 -terminally modified, exonuclease resistant, random
hexamers as primers. This results in an exponential, hyperbranched amplification
of DNA of 10,000-fold (Dean et al. 2001), or even more (Dean et al. 2002). The
hexanucleotides bind at several sites to the template DNA, and are elongated by the
φ29 polymerase. If the polymerase reaction reaches the 5 -end of the next hexamer
binding to the DNA, the preceding DNA strand is displaced and the displaced single strand is again subject to the binding of hexamers and elongation. The result is
a network of amplified DNA (Fig. 2.2). Disadvantages of this method include the
formation of chimeras, amplification biases, and non-specific DNA amplification.
Chimera formation during MDA occurs with a frequency of one rearrangement per
10–22 kb of MDA product (Zhang et al. 2006, Lasken and Stockwell 2007). In 85%
A. Meyerdierks and F.O. Glöckner
2.3.4 Amplification of Genomic DNA
Comprehensive metagenome analyses generally require microgram amounts of
genomic DNA in order to be representative (see below). Theoretically, less DNA
would be sufficient for a representative metagenome analysis, if the microbial diversity in the sample is low or strategies have been applied to reduce the diversity and
to enrich for genomic DNA of target cell populations. However, standard metagenomic library construction and sequencing techniques are not compatible with
DNA amounts much below the microgram level. In some studies, large amounts
of genomic DNA cannot be obtained, e.g., from the deep biosphere (Webster et al.
2003, Biddle et al. 2008). These small quantities of starting DNA need to be amplified before metagenome analysis. To overcome this problem, different genomic
DNA amplification methods have been developed (Telenius et al. 1992, Zhang
et al. 1992, Breitbart et al. 2002, Breitbart and Rohwer 2005, Pinard et al. 2006)
of which one, the multiple displacement amplification (MDA) with bacteriophage
φ29 polymerase, is now widely used in metagenome analyses.
The DNA polymerase of bacteriophage φ29 from Bacillus subtilis is required
for the replication of the 19,285 bp phage genome (Blanco and Salas 1985b). In
addition to its DNA polymerase activity, this single polypeptide of 66,520 dalton
(Blanco and Salas 1984, Watabe et al. 1984) has a 3 → 5 exonuclease activity
(Blanco and Salas 1985a). The error rate of the φ29 polymerase (∼10 –5 to 10 –7 )
(Esteban et al. 1993, Nelson et al. 2002, Zhang et al. 2006) is about 10–100-fold
lower than the mutant frequency determined for Taq DNA polymerase (Eckert and
Kunkel 1990) and the processivity is high. The replication of the bacteriophage φ29
genome in a protein primed reaction was determined to last 8 min (Blanco et al.
1989). The φ29 polymerase can also elongate DNA-primed reactions. The circular
single stranded M13 genome (∼7,250 bases) was isothermally replicated in 5 min,
producing a strand of more than 70 kb after 40 min, due to the strand displacement
capability of the φ29 polymerase (Blanco et al. 1989). This strand displacement
capability leads to a successive detachment of the 5 -end of the synthesised DNA
molecule from the template strand by the φ29 polymerase during DNA synthesis, resulting in a rolling circle amplification (RCA) (Fire and Xu 1995, Liu et al.
1996).
In (meta)genome analyses, the φ29 DNA polymerase is used in a multiplyprimed amplification with 3 -terminally modified, exonuclease resistant, random
hexamers as primers. This results in an exponential, hyperbranched amplification
of DNA of 10,000-fold (Dean et al. 2001), or even more (Dean et al. 2002). The
hexanucleotides bind at several sites to the template DNA, and are elongated by the
φ29 polymerase. If the polymerase reaction reaches the 5 -end of the next hexamer
binding to the DNA, the preceding DNA strand is displaced and the displaced single strand is again subject to the binding of hexamers and elongation. The result is
a network of amplified DNA (Fig. 2.2). Disadvantages of this method include the
formation of chimeras, amplification biases, and non-specific DNA amplification.
Chimera formation during MDA occurs with a frequency of one rearrangement per
10–22 kb of MDA product (Zhang et al. 2006, Lasken and Stockwell 2007). In 85%