The evolution of complex genomes requires that new combinations of pre-existing protein domains successfully fold into modular polypeptides. During eukaryotic translation model two-domain polypeptides fold efficiently by sequential and co-translational folding of their domains. In contrast, folding of the same proteins in Escherichia coli is post-translational, and leads to intramolecular misfolding of concurrently folding domains. Sequential domain folding in eukaryotes may have been critical in the evolution of modular polypeptides, by increasing the probability that random gene-fusion events resulted in immediately foldable protein structures.