Domain-domain interactions are a critical type of the mechanisms mediating protein-protein interactions (PPIs). For a given protein domain, its ability to combine with distinct domains is usually referred to as promiscuity or versatility. Interestingly, a previous study has reported that a domain's promiscuity may reflect its ability to interact with other domains in human proteins. In this work, promiscuous domains were first identified from the yeast genome. Then, we sought to determine what roles promiscuous domains might play in the PPI network. Mapping the promiscuous domains onto the proteins in this network revealed that, consistent with the previous knowledge, the hub proteins were significantly enriched with promiscuous domains. We also found that the set of hub proteins were not the same set as those proteins with promiscuous domains, although there was some overlap. Analysis of the topological properties of this yeast PPI network showed that the characteristic path length of the network increased significantly after deleting proteins with promiscuous domains. This indicated that communication between two proteins was longer and the network stability decreased. These observations suggested that, as the hub proteins, proteins with promiscuous domains might play a role in maintaining network stability. In addition, functional analysis revealed that proteins with promiscuous domains mainly participated in the "Folding, Sorting, and Degradation" and "Replication and Repair" biological pathways, and that they significantly execute key molecular functions, such as "nucleoside-triphosphatase activity (GO:0017111)."