In multi-template PCR, chimeric sequences come out from template switching during PCR reaction and mismatch repair of heteroduplexes by the host cell upon cloning. How to reduce such artifacts to avoid overestimation of haplotypes? I present here a trick to reduce such artifacts with economy use of TA cloning kit.

Incomplete extension products work as primers at the next PCR cycle to make chimeric sequences.
Inactivating enzyme in later cycles is prone to make incomplete extension.

In later PCR cycles when the reaction reaches a plateau, rennealing between PCR products making
heteroduplexes becomes predominant over annealing between PCR product/template and primer.
Cloning heteroduplex sequence into a host cell, the mismatch repair system works randomly
using either strand as template to make chimeric sequence.
Optimization of number of cycles with a real time PCR, use of heteroduplex-specific decomposition enzyme, and addition of three cycles in refreshed PCR mixture have been reported to reduce chimeric sequences upon PCR and cloning. The latter method is cheap and easy without additional machine or enzyme. Recently I tested it, and it turned out that only a single re-extension is enough for cloning reducing chimeras.

Heteroduplexes are heat denatured in a refreshed PCR mix containing sufficient primers, dNTPs and
enzyme. Primers competitively and preferentially anneal to single-strand DNAs suppressing
heteroduplex formation. Homoduplexes are subsequently recovered by a single re-extension of nascent
strand.
Outline
1st round PCR
not more than 30 cycles (IMPORTANT)1)
Electrophoresis
to quantify PCR product
Dilute
to c.a. 10ng/uL with MQ2)
Single re-extension
1uL diluted PCR product
+ 9uL PCR mix
{dNTPs (Takara) 0.8uL, 10x ExTaq buffer 1uL, 1/10x ExTaq 0.5uL,
F-primer (5pM/uL) 1uL, R-primer (5pM/uL) 1uL, MQ 5.2uL}
1/10x ExTaq =
{ExTaq (5U/uL) diluted 10 times with 10mg/mL BSE (Boehringer #238040)}
{94°C (180sec) -> X°C (30sec) -> 72°C (30min) -> 4°C} x 1
Ligation
1uL re-extension product
+ 1uL 1/10x pGEM-t {pGEM-t (Promega) diluted 10 times with MQ}3)
+ 3uL ligation mix {2x buffer 2.5uL, ligase 0.5uL (pGEM-t kit)}
IMPORTANT; add ligation mix finally4)
15°C 4hrs -> 4°C several hrs (optional)
+ 5uL TE
TF
Put competent DH5alpha on ice to thaw5)
4uL of the diluted ligation product in 2ml Eppendorf tube on ice
Pour 50uL of DH5alpha onto the ligation product
On ice 15min
44°C 45sec
On ice 1min (immediately prepare the next step)
+500uL SOC (37°C)
Rotate 10min (if SOC is chilled, 15min)
Spread 100uL on an LB amp+ X-gal+ plate
37°C o/n
Efficiency
100-500 colonies, more than 1/2 white
An example from a tetraploid loach (RAG1, 1.5kb, 4 alleles
with 0.9 - 3.8% nt differences)6)
----------------------------------------------------------
reaction condition cycles chimera/total clones
----------------------------------------------------------
no re-extension 35 22/25
no re-extension 30 5/16*
single re-extension 35 17/28
single re-extension 30 3/19
----------------------------------------------------------
* An allele out of four was not recovered.
<Note>
1) Reaction with ExTaq (Takara). Many chimeras appeared even upon re-extension from PCR product of 35 cycles.
2) The value is in case of 1.5kb insert. Adjust DNA amount according to length of the insert.
3) This combination makes a good insert-vector ratio of about 2ng of insert (1.5kb) and 5ng vector (3kb).
4) When there are a number of inserts to ligate, mix insert and vector first, and then add ligation enzyme mix later. Frequent self-ligation occurs if mixture of enzyme and vector is added to insert (mixing enzyme and vector first), though it is easier.
5) Home-made competent cell.
6) Chimeras cannot be completely eliminated even upon single re-extension. Use the results cautionarily.