PDZ domains (also called Discs-large homologous regions (DHR) or GLGF) are conserved structural elements of 80 to 100 amino acids that were originally identified in the post-synaptic density protein PSD-95, the Drosophila tumor suppressor discs-large, and the tight-junction protein ZO-1. PDZ domains occur as single or, more frequently, as multiple tandemly repeated copies in a large and diverse set of proteins from all eukaryotes [1,2]. The presence of one or two copies of PDZ domains in some bacteria may be explained by horizontal gene transfer [3].

Eukaryotic PDZ domains are multifunctional protein-protein interaction modules that are involved in the clustering of signaling molecules and play important role in organizing protein networks on membranes. In many cases, PDZ domains bind to a signature motif ([FYST]-X-[FVA]) occurring at the very C-terminus of target proteins. PDZ domains are also able to form heterodimers, and to interact with internal peptide fragments of target proteins [1,2,4]. It has been proposed that bacterial PDZ domains also possess C-terminal polypeptide binding functions [3].

PDZ domains are compact globular domains that were originally called GLGF as Gly-Leu-Gly-Phe is a relatively conserved element of their sequences [1]. Resolution of crystal and solution structures of PDZ domains with and without ligands has revealed that the PDZ domain contains six β-strands and two α-helices [5,6,7,8]. The C-terminal motif of target proteins has been shown to bind in an anti-parallel fashion to a groove formed by the principal α-helix and the second β-strand of the PDZ domains [5,8]. The binding site involved in the heterodimerization of some PDZ domains is situated opposite this canonical peptide binding groove. It is an about 30 residue C-terminal extension of the PDZ domain, which forms a β-hairpin finger. The first strand of the β-hairpin finger mimics a canonical C-terminal peptide ligand, inserting into the peptide binding groove of the PDZ domain from the partner protein [7,8]. It has been proposed that PDZ domains may bind in a general fashion to non-terminal sequences, possibly those with β-finger-type structures, regardless of whether or not they are found downstream of other PDZ domains [9].

Some proteins known to contain a PDZ domain are listed below:

• Eukaryotic membrane associated guanylate kinases (MAGUKs).
• Mammalian amyloid β A4 precursor protein-binding family A proteins, putative function in synaptic vesicle exocytosis by binding to Munc18-1, an essential component of the synaptic vesicle exocytotic machinery.
• Mammalian neuronal nitric oxide synthase (nNOS).
• Interleukin-16, the only known secreted member of the PDZ protein family.
• Vertebrate LIM-kinases (LIMK), which display serine/threonine-specific phosphorylation of myelin basic protein and histone in vitro.
• Drosophila and vertebrate dishevelled (Dsh and Dvl), proteins that play a key role in the transduction of the Wg/Wnt signal from the cell surface to the nucleus.
• Drosophila InaD, a photoreceptor scaffolding protein that assembles multiple signal transducing proteins at the membrane via its 5 PDZ domains.
• Bacillus subtilis stage IV sporulation B protein (spOIVB).
• Bacterial gspC proteins, involved in protein export via type II secretion systems.
• Bacterial periplasmic serine proteases high-temperature requirement A (htrA) and tail-specific protease (tsp).
• Escherichia coli and Haemophilus influenzae hypothetical protein YaeL, probably associated with the cytoplasmic membrane.

A profile was developed that covers the minimal PDZ domain and thus lacks about 30 C-terminal residues (which form the β finger involved in PDZ-PDZ interactions).