Cryptic 3′ mRNA processing signals hinder the expression of Schistosoma mansoni integrins in yeast

- Codon usage was hypothesised as main reason for failing to express parasite genes.
- Schistosome genomes contain many AT-rich genes.
- We show that AT-rich stretches are interpreted as mRNA 3′-end processing signals.
- Optimising these sites prevents truncation and allows full-length protein expression.

The expression of parasite genes has often proven difficult in heterologous systems such as yeast or E. coli. Most often, promoter choice and codon usage were hypothesised to be the main reason for expression failures. The trematode parasite Schistosoma mansoni has five integrin genes named Smα-Int1-4 and Smβ-Int1, which we aimed to express in the yeast Saccharomyces cerevisiae. This has not been achieved, however, as only Smβ-Int1 integrin could be expressed. When the four α integrins were driven by a stronger promoter, this enabled Smα-Int1 to be expressed as well, but the remaining integrins, Smα-Int2-4, still could not be expressed. Evidence from RT-PCR experiments suggested that this was due to premature transcription termination. Using detailed in silico sequence analyses we identified AT-rich stretches in these integrin genes, which have high similarity to yeast mRNA 3′-end processing signals. We hypothesised that these signals were causing the premature truncation. To test this, we designed an optimised version of Smα-Int3, in which the sequence was modified to replace the yeast 3′ processing signals. This strategy allowed us to express Smα-Int3 integrin successfully in S. cerevisiae. These findings show that the misinterpretation of AT-rich sequences by yeast 3′-mRNA processing machinery can cause problems when attempting to express genes containing such sequences in this host.

245