It is possible that this template is GC rich and the secondary structure interferes with the pcr. You might try a dmso gradient in 1% steps up to 8% ( use a lower annealing temp for your primers for this test.

It might be interesting to know which primer is the biggest problem so try each of your primers in a single primer "pcr". Some sequences are found near each other and in opposite orientation so if any primer gives lotds of amplimers on its own then it needs redesigning. Try nested/hemi nested pcr. If the outer primers give a mess then re amplify 1/10th ul with nested internal primers 20 cycles.If you are lucky most of the wrong bands do not also contain the internal primer sequences.

On the assumption that the template is high Gc try a pcr with 1M betaine in the pcr mix.

I think that long pcr works best with high annealing primers because the template is long and reanneals quickly so the answer may be to design longer primers and move towards a 2 step pcr or at least an annealing temperature close to the extension temperature of the enzymes

Also you are getting primer dimer at the higher annealing temperatures. This may make the amplification less efficient.Try using less primer

This template looks like the FMR gene which is high GC with long runs of GC triplets. If this is the case than recently published work concerning the amplification of this gene should give you good primer pairs and pcr reaction conditions . Many people have worked on this problem amplification and there will be some good stuff published

Hi both Paul Rutland Nahed Hussien

Would you say an approach using both Betaine and DMSO could work or are they mutually exclusive optimizations (ie can I use both reagents in the same PCR reaction or should I use them separately?)

Thanks!

Bea