Long-PCR Reagents and Guidelines

from George Church as Modified from Cheng et al. (1)

General Guidelines for Long-PCR Conditions and Enzyme Mixtures
==========
Following the results of Cheng et al. (1) we have had success using Tth (ABI/Perkin-Elmer) as the main-component polymerase and Vent (New England Biolabs) as the fractional-component polymerase.
For PCR with low-complexity templates (e.g., phage l, plasmid and cosmid inserts)

50ml reaction
1-2.5 U Tth

0.02 U Vent

0.2 mM each dNTP

20-40 pmoles each primer (400-800 nM)

1.1-1.2 mM Mg(OAc)2

105 -107 template molecules
For PCR with moderate-complexity templates (e.g., bacterial genomic DNA)

50ml reaction same as above except for the following change

0.1 U Vent

This Vent concentration has not been completely optimized and more Vent might be better, but this concentration works well. Please experiment.
For PCR with high-complexity templates (e.g., human genomic DNA)
No data available yet; please inform us of your data.

Cycle times and temperatures
==========
Generally, we have been using two temperature cycles with one annealing/extension step @ 68? and a short melting step @ 94?.
Presently, a rough formula for calculating annealing/extension times is
1 min + (2.5 sec/100 bases)
The constant one minute is probably necessary for primer extension to occur; at 68?; the kinetics of primer-template annealing and melting may become the limiting factor in the rate of primer extension.

Generic Long-PCR Program
==========

Initial melting 94?;, 10-15 sec

Cycles 1-15 94?;, 10 sec @; 68?;, x min (15 times)

Cycles 16-30 94?;, 10 sec @; 68?;, x min +15 sec/cycle (15 times)

The 15 sec cycle extension for cycles 16-30 may be necessary for only the longest PCR (>15-20 kb), please experiment.

Picking Primers
==========
We have had success using the following guidelines for primers (these are not inviolable rules, they are simply guidelines)
Primers are 20 to 23 bases in length

G+C = 12 bases

A+T = 8 to 11 bases

Ideal Tm = 60?; to 68? in 85 mM salt.
These values were calculated using the PrimerSelect program of DNAStar. A working primer might have a Tm significantly lower than 60?;, and we have used successfully primers with Tm values below 50?;; but avoid this if possible‹especially if you are doing genomic DNA templated reactions. The largest PCR we have done is about 20 kb using the positive control primers from the PCR-XL kit available from Perkin-Elmer; these primers both have Tm values of about 60?; using PrimerSelect from DNAStar. Other programs probably give similar results.
Avoid primer hairpins.

Avoid primers with 3' complementarity (results in primer-dimers).
These last two problems can be avoided with the aid of a primer-picking program like PrimerSelect.

One additional piece of advice: Pick primers with A/T-rich 3' ends if possible (2). G/C-rich 3' ends may be too sticky and give non-specific products. Using this additional rule primer length might increase to 25 or 26 bases. We have not used this rule in our primer-picking but are planning to in the near future.

Materials
========

Long-PCR buffer

85 mM KOAc

25 mM Tricine pH 8.7 (adjust pH of stock solution with KOH)

8% glycerol

1% DMSO (1 to 4% works)

A 5X buffer stock containing 5% DMSO (1% final conc.) can easily be made using 1 M Tricine and 1 M KOAc stock solutions as follows.

10 ml 5X buffer

4.25 ml 1M KOAc

1.25 ml 1M Tricine, pH8.7 @ 25? (with KOH)

4.00 ml glycerol

0.50 ml DMSO