mrna - Is exon order always preserved in splicing?

After performing a quick literature search, I am, as with GWW, unable to provide any literature against this occurring, although this paper by Black (2005) states that exons in multi-exon pre-mRNAs are always maintained in order.

Black DL. 2005. A simple answer for a splicing conundrum. Proceedings of the National Academy of Sciences of the United States of America 102: 4927–8..

Separate components of the spliceosome recognise sequences at the the start and end of each intron, along with a branch point adenine and some other conserved (highly in the case of yeast) regions upstream of the 3' end. The components seem to assemble in a particular order, which is 5' to 3' directional.

The general consensus seems to be that the spliceosome moves along the pre-mRNA from 5' to 3' removing the introns and splicing the exons together to form the mature transcript, ready for translation. The introns are subsequently degraded. In the case of eukaryotes some processing is required to transport the mRNA out of the nucleus to the ribosomes in the cytoplasm.

The mRNA would have a start codon near the 5' terminus and stop codon near the 3' end, at the boundaries of the CDS, and it isn't clear to me how this structure could be maintained if exon order didn't follow the 5' to 3' assemblage. I should imagine that the protein could be affected in much a similar way to exon skipping, possibly resulting in a truncated, inactive or ineffective protein. As protein folding is determined by the interactions between the amino acids, any change in exon order (secondary structure) is likely to result in a different tertiary structure.