Cellular functions are regulated by complex webs of interactions that might be schematically represented as networks. Two major examples are transcriptional regulatory networks, describing the interactions among transcription factors and their targets, and protein-protein interaction networks. Some patterns, dubbed motifs, have been found to be statistically over-represented when biological networks are compared to randomized versions thereof. Their function in vitro has been analyzed both experimentally and theoretically, but their functional role in vivo, that is, within the full network, and the resulting evolutionary pressures remain largely to be examined. We investigated an integrated network of the yeast Saccharomyces cerevisiae comprising transcriptional and protein-protein interaction data. A comparative analysis was performed with respect to Candida glabrata, Kluyveromyces lactis, Debaryomyces hansenii and Yarrowia lipolytica, which belong to the same class of hemiascomycetes as S. cerevisiae but span a broad evolutionary range. Phylogenetic profiles of genes within different forms of the motifs show that they are not subject to any particular evolutionary pressure to preserve the corresponding interaction patterns. The functional role in vivo of the motifs was examined for those instances where enough biological information is available. In each case, the regulatory processes for the biological function under consideration were found to hinge on post-transcriptional regulatory mechanisms, rather than on the transcriptional regulation by network motifs. The overabundance of the network motifs does not have any immediate functional or evolutionary counterpart. A likely reason is that motifs within the networks are not isolated, that is, they strongly aggregate and have important edge and/or node sharing with the rest of the network.