We have further analyzed parameters affecting stable transformation of
Trypanosoma brucei. Linear DNA was much more efficient than circular
DNA and in the vast majority of transformants analyzed the plasmid DNA
had inserted into the chromosomes by homologous recombination. The pr
esence of non-homologous (vector) DNA at one or both ends of linear co
nstructs inhibited transformation efficiency. Less than 1 kb of homolo
gous flanking sequence was sufficient for efficient targeting of a mar
ker gene into the tubulin gene array. When transformants with a single
neomycin phosphotransferase (neo(r)) gene replacing a beta-tubulin ge
ne were selected for higher levels of G418 resistance, the neo(r) gene
was amplified and spread through the tubulin gene cluster. The additi
onal neo(r) gene copies were adjacent in the tubulin gene array and we
re added to the array rather than replacing beta-tubulin genes. These
results are compatible with asymmetric post-replication recombination
(unequal sister chromatid exchange) as the mechanism for neo(r) gene a
mplification. Starting with a circular construct containing the neo(r)
gene between tubulin intergenic regions, we obtained a single transfo
rmant that maintained the neo(r) genes as an extrachromosomal plasmid.
We show this plasmid to consist of a circular pentamer of the input c
onstruct. All other attempts to derive a shuttle vector that replicate
s extrachromosomally in T. brucei were unsuccessful. Our experiments e
xtend previous observations suggesting that T. brucei has a strong pre
ference for chromosomal insertion of exogenous DNA by homologous recom
bination.