All general molecular biology techniques were carried out as described in the Molecular Cloning Laboratory Manuals by Sambrook and Russell (Cold Spring Harbor Laboratory Press). DNA- and RNA-concentrations were measured in one-microliter samples using a NanoDrop spectrophotometer (NanoDrop Technologies, Wilmington, Delaware, USA).
Template libraries and RNAs
The MegaMan library (Stratagene, La Jolla, CA, USA) is a transcriptome library which contains pooled plasmids harboring cDNAs from many different human tissues and cell lines. Chicken total RNA was prepared from frozen E5–E9 stage chicken embryos (supplied by Esther Stöckli, University of Zürich). The RNA was purified with Ambion's MELT total RNA preparation system (Ambion, Austin, TX, USA). The other RNAs were purchased from commercial sources; adult Drosophila melanogaster poly(A)+ mRNA was from Clontech, BD Biosciences (Palo Alto, CA, USA) and Arabidopsis thaliana total mRNA from BioChain Institute (Hayward, CA, USA).
Construction of template for PCR amplification from plasmid cDNA library
The MegaMan plasmid cDNA library was amplified by RCA with the bacteriophage Φ29 DNA polymerase . The random octameric oligonucleotide primers were modified with two Nitroindole bases and two phosphorothioate links: 5' Nitroindole-Nitroindole-NNNN-s-N-s-N 3' (N = any base, s = phosphorothioate link; obtained from Microsynth, Balgach, Switzerland) in order to increase template-specific synthesis while keeping the background at a low level . A 180 microliter reaction contained 0.1 mM random primers, 75 ng cDNA library, 1 × Tango buffer (33 mM Tris-acetate pH 7.9 at 37°C, 10 mM Mg-acetate, 66 mM K-acetate, 0.1 mM BSA, Fermentas, Vilnius, Lithuania), 0.4 mM desoxynucleoside triphosphates, 1 × Φ29 polymerase buffer, 2.5 mU ~ 1.3 μg yeast pyrophosphatase (Roche Diagnostics, Basel, Switzerland) and 35 U Φ29 DNA polymerase (New England Biolabs, Ipswich, MA, USA or Fermentas). The mixture was incubated for 24–48 h at 34°C, thereafter EDTA was added to 20 mM. The stopped reaction was subjected to 3 freeze-thaw cycles to shear the large DNA complexes and precipitated with 2.5 volumes of ethanol in the presence of 2 M ammonium acetate pH 5.0 for 15 minutes on ice. The precipitate was spun down for 10 minutes at 4°C at 13'000 × g. The pellet was washed 2 × with 70% ethanol, dried for 5 min at room temperature and dissolved with 500 μl 10 mM Tris-Cl pH 8.0, 1 mM EDTA during 30 min at 60°C. Thereafter, the solution was kept overnight on a turning wheel at room temperature. This procedure yields at least 150 μg DNA, i.e. sufficient template DNA for 1000 standard 50 μl PCR reactions (see below). Other ways of purifying the DNA, e.g. spin column adsorption and elution, resulted in loss of most of the large branched DNA complexes; e.g. the purification of the DNA using QIAquick PCR purification columns (Qiagen) generated tenfold less DNA (cf. ).
Construction of template for PCR amplification from poly(A)+- or total RNA via ds cDNA
The first strand cDNA was synthesized using a cDNA synthesis kit (Roche). The cDNA was cleaned by phenol-chloroform extraction and ethanol precipitation in presence of 2 M ammonium acetate pH 5.0. The products obtained from 1 μg of poly(A)+ RNA or from 10 μg of total RNA were dissolved in 25 μl of 10 mM Tris-Cl pH 8.0, 0.1 mM EDTA. The double-stranded cDNA was ligated to concatemers and circles with T4 DNA ligase in presence of 5% polyethylene glycol 4000 according to the specifications of the supplier (Fermentas). The overall yield was 1–2 μg ligated cDNA. Approximately 100 ng of this cDNA were amplified with Φ29 DNA polymerase as described above and yielded template for about 1000 PCR reactions.
Construction of template for PCR amplification from poly(A)+RNA or total RNA via direct circularization of the ss cDNA
The oligodT(15) primer was ordered as a 5' phosphorylated primer (Microsynth) or was treated with T4 polynucleotide kinase and rATP (Fermentas). The phoshorylated primer was annealed to the chicken embryonal total RNA (4–6 μg) and the first strand cDNA synthesized with AccuScript reverse transcriptase (Stratagene) according to the specifications of the supplier. EDTA was added to final 20 mM to the ss cDNA product (about 3–5 μg cDNA in 40 μl), heated to 98°C for 2 min to melt the cDNA-RNA hybrids. The solution was chilled and 15 U RNAse I (from the Roche cDNA synthesis kit) was added. After 30 min at 37°C, 3 U Proteinase K were added and the incubation was continued at 37°C for 30 more minutes. Thereafter, Proteinase K was inactivated by heating at 70°C for 15 minutes. The products were precipitated with ethanol and 2 M ammonium acetate pH 5.0. After the wash with 70% ethanol, the accrued 2.5–3.0 μg DNA was dissolved in 15 μl H2O. To this sample, 2 μl of Circle Ligase buffer, 50 μM rATP and 2 μl of CircLigase (Epicentre, 100 U/μl) were added. The ss cDNA was circularized by incubation for 1 h at 60°C and 10 min at 80°C. This circular DNA could be used for ramification amplification with Φ29 DNA polymerase as described above. A sample containing 400 ng of circular ssDNA yielded about 80 μg ds DNA to feed about 1000 standard PCR reactions with template.
When the novel Phusion DNA polymerase PCR mixture (Finnzymes) became available, we decided to compare its performance with that of the most successful PCR set up known to us. Its application according to the supplier's specifications (both HF and GC buffer) improved our results from about 50 to 70% positive reactions at first go. Both addition of methyl sulfoxides  and changing the primer concentration, are known to largely influence the PCR outcome. When tetramethylene sulfoxide (TMSO, Acros Chemicals) was introduced as an additive with high and low primer concentration (1 μM or 100 nM final concentration, respectively) we achieved the highest level of positive reactions. The most successful combinations required four separate PCR reactions per target: a) high primer, HF buffer, b) low primer, HF buffer c) high primer, GC buffer, 2% TMSO, d) low primer, GC buffer, 2% TMSO. Tests with commercial PCR optimization kits like the FailSafe PCR system (Epicentre, 12 different conditions) did not outperform the a/b/c/d conditions with Phusion polymerase. The cycling parameters are also crucial. The most reliable result was obtained using a step-down protocol: 1 min 98°C; 20 cycles 30 s 98°C, 1 min 60°C to 50°C (decrease of 0.5°C per step), X min 72°C (depends on target length, about 20 s per kb) followed by 20 cycles 30 s 98°C, 30 s 52°C, 1 min 72°C. Finally cool to 10°C. The PCR products are stable at this temperature for a few days. For the classical DNA polymerases (Pfu plus Taq) the extension time was 1 min per kb.
We set up a systematic survey comparing the Phusion polymerase conditions with the well-known Pfu-Taq mixture and one of the most recently available other fusion polymerases, namely Stratagene's Herculase II Fusion polymerase. The three polymerase formulations were used at high or low primer concentrations (1 μM or 100 nM, respectively) as well as in absence or presence of 2% TMSO (the maximally tolerated TMSO concentration with Phusion polymerase is about 6%, data not shown). To ensure highly reproducible pipetting, we ran the experiments on a Tecan Freedom Evo II liquid handling workstation. The master mixes including the standard buffer, enzyme(s), nucleotides and water to 50% of the final volume were mixed manually. The remainder was added by the liquid handling workstation. The PCR was carried out by an MJ Research (BioRad) thermocycler installed on the deck of the machine. The agarose gels were prepared and run manually or were E-gels (InVitrogen) and run by the workstation.