PCR produced probes

Tue Oct 11 11:26:46 EST 1994

I've been having some difficulties producing DNA probes greater than 6 kb by 
PCR and was hoping someone could point me in the right direction. Here's what 
we've done in a nutshell. We knocked out pathogenicity in a plant pathogenic 
fluorescent pseudomonad by transposon mutagenesis (Tn5). I think there were 34
mutants originally. Screening ClaI restricted mutants with DIG labelled Tn5
revealed 9 unique sites were present. We prepared a cosmid library of the 
wild-type pseudomonad in E. coli HB101 and would like to find the clones with 
with the genes of interest. We've been able to prepare probes from 2 of the 9
mutants by following the protocol of Rich and Willis (Nucleic Acid Research,
vol 18, No. 22, pages 6673-6676) with some minor revisions. In this method the
fragment containing the Tn5 is circularized by ligating the sticky ends with 
T4 ligase (we cut with ClaI to get compatible ends). Since the sequence of the
terminal repeats of the Tn5 are known we made a 20-mer primer from this region
for PCR.  The 2 mutants with the smallest ClaI fragments (7.3 kb and 9.1 kb)
worked perfectly giving a single band of the predicted size (1.4 kb and 3.2 kb
respectively) since amplification of the circularized fragments results in the
exclusion of the transposon from the final product (Tn5 is around 5.9 kb). The
other 7 mutants have fragment sizes ranging from 12.6 kb to 15.4 kb. The lanes
were clean for these guys, not even a slight smear. The first thing that came
to mind was that we were outside the range of Taq polymerase's amplification 
capabilities (desired products are about 6.7 kb to 9.5 kb) so we tried the Taq
extender by Stratagene. This time there was a weak smear in all the lanes 
ranging from 3500 bp to 4100 bp but no definate bands. 

Does the fact that we got some smearing mean that the PCR conditions simply
need to be optimized to get larger products. Maybe the configuration of the 
template is wrong. Is it possible to circularize 12-15 kb fragments or is the 
probability that the sticky ends of adjacent fragments will find each other 
before the ends on the same fragment, too great. If anyone has any suggestions
your input would be greatly appreciated. The reason we adopted this method for
constructing our probes was it seemed quick and easy and for the 2 that worked 
this was certainly the case, however I'm definately open to some alternative 
approaches as well.   

Thanks in advance,

Tim Ells
Agriculture Canada 
Research Station
Kentville N.S.
B4N 1J5

More information about the Methods mailing list