Problems with AT-rich sequence

[Bengt Oxelman]

We have recently started a project that involves sequencing of a 
AT-rich (60-70 %) sequence in the chloroplast genome. After amplifying
a 900-bp fragment, we have tried cycle sequencing, using a protocol 
that has worked well on amplified fragments of the nuclear rDNA 
(c. 40% AT). Now, we get sequences that are almost unreadable, because
of lots of 'false stops'. We would appreciate very much if anyone could
give us a hint on how to solve this problem. Lower extension temp? 
Different d/ddNTP concentrations? Tetramethyl ammonium chloride? Could 
the sequence close to the 3' of the primer cause problems (i.e. like
AGATGAGCAACACCCCCCCTAGAAA)?

Any thoughts greatly appreciated!

Bengt