The nucleotides cyclic adenosine 3′,5′-monophosphate (cAMP) and cyclic guanosine 3′,5′-monophosphate (cGMP) are intracellular second messengers that play a key role in mediating cellular responses to various hormones and neurotransmitters. Levels of cyclic nucleotides are tightly regulated at the level of synthesis (by adenylate and guanylate cyclase) as well as degradation (by cyclic nucleotide phosphodiesterases, PDEs). The regulation of cyclic nucleotide activity by PDEs is rendered extraordinarily complex because of differential expression, activation, and cross-talk between these isoforms. Historically, it is nearly 50 years since Rall and Sutherland described pharmacological inhibition of PDE activity by theophylline. Together with the application of biochemical and molecular cloning techniques, PDE inhibitors have been pivotal tools in the characterization of the diversity and complexity of cyclic nucleotide signaling and regulation. In 1985, Weishaar et al. published a Perspective in this journal that described the discovery of multiple phosphodiesterase (PDE) isoforms and the therapeutic potential of selective PDE inhibitors. That article stands as an excellent point at which to measure the changed landscape of medicinal chemistry in this area. What was understood at that time to be a family of three isozymes has become a total of 21 human PDE genes falling into 11 families with over 60 isoforms. Although the challenges have become more complex, the demand for isoform-selective inhibitors by which to understand PDE physiology has not changed, and the therapeutic potential of phosphodiesterase inhibitors is not dissimilar to that described 2 decades ago. The publication in 2000 of the first crystal structure of a PDE isoform, PDE4B2B, shed new light onto the mechanism of PDE catalysis and provided a refined template from which to design a new generation of PDE inhibitors. Recently, new crystal structures of PDE enzymes have been published at an extraordinary rate such that there is now almost a surfeit of information regarding PDE structure that is difficult to digest. However, within the increasingly large data set resides information that may provide a crucial edge in the design of novel PDE inhibitors, particularly for those isoforms where selective inhibitors are not yet available. To date, it has been necessary to rely on massive medicinal chemistry efforts to generate the potent and isoform-selective ligands that have been developed for PDE3, PDE4, and PDE5. Even then, for sildenafil, a lack of selectivity for PDE5 over PDE6 has emerged as a clinical issue. The release of the crystal structures mentioned above should allow structure-based elements to be used more often in the discovery of new inhibitors. The potential also exists to develop pathways to perform virtual screening for both potency and selectivity for the PDE isoforms of interest, thereby increasing the efficiency of the drug design process. At the time of writing, crystal structure coordinates relating to five PDE isozymes have been released covering PDEs 1, 3, 4, 5, and 9. In this Perspective, we look at the information that these structures provide to us about the PDE superfamily as it relates to homology and structural alignment, cyclic nucleotide substrate specificity, inhibitor selectivity, and the prospects for the design of isoform-selective PDE inhibitors.