PROSITE documentation PDOC50835

Ig-like domain profile

The Ig-like domain is probably the most widespread domain, at least in animals. This domain can be considered as an heterogeneous group built on a common fold. Proteins containing an Ig-like domain differ in their tissue distribution, amino acid composition, and biological role [1,2,3].

All Ig-like domains appear to be involved in binding functions. The ligands range from small molecules (antigens, chromophores), to hormones (growth hormone, interferons, prolactin), up to giant molecules (muscle proteins) [3]. Binding sites are localized either in the loop regions (the most variable parts of the immunoglobulins) or in strands. For instance, distinct areas of the sheets are used to bind the ligands of the MHC, CD8, CD4, and PapD molecules or of the growth hormone receptor (GHR). These binding sites may be formed by a single chain (CD2, CD4), by homodimers (GHR, CD8), or by heterodimers [3].

Classical Ig-like domains are composed of 7 to 10 β strands, distributed between two sheets with typical topology and connectivity described as a Greek key β-barrel (see <PDB:3HLA>). The general shape of Ig-like domains is well conserved, but they can differ significantly in their size, owing to high variability of the loops. While a classical domain contains about 100 residues, smaller ones (74-90 residues) have been observed in several Ig-related molecules (CD2, CD4). Large decorations within loops, sometimes including extra domains, are found in hemocyanin (238 amino acids) [4] and transcription factor NFkappaB (201 amino acids) [5]. The schematic representation of the structure of a typical Ig-like domain is shown below:

               ----------======D======>--
              |                          |
              |    ------<=====E=======--
              |   |
              |   |    --======B======>------
              |   |   |                      |
              |   |    --<=====A=======      |
              |   |                          |
              |   |                          |
              |   |      <=====G=======--    |
              |   |                      |   |
              |    ------======F======>--    |
              |                              |
               ----------<=====C=======------

     '=>': indicates the direction of the beta-strands 'A' to 'G'.

Ig-like domains can be classified according to the numbers of β strands [1,3]:

C1-type: classical Ig-like domain, described in the schematic representation. Sheet I: ABED, sheet II: CFG. Domain C1 is found exclusively in the molecules involved in the immune system:

• Immunoglobulins (Ig).
• Major Histocompatibility Complex (MHC) molecules.
• T-cell receptors (TcR).

C2-type: strand D is deleted and replaced by strand C' directly connected to strand E. Sheet I: ABE, sheet II: C'CFG:

• Second domain of the vascular cell adhesion molecule-1.
• Neural cell adhesion molecule 2.
• Vascular endothelial growth factor receptor 3.
• Fibroblast growth factor receptor 4.
• Interleukin-6 receptor α chain.

V-type:extra strands C' and C" between strand C and D. Sheet I: ABED, sheet II: C"C'CFG:

• Variable domain of the immunoglobulin heavy chain.
• T-cell surface glycoprotein CD8 α chain.
• Viral Hemagglutinin.
• Programed cell death protein 1.
• Neurocan core protein.
• Myelin protein zero.

H-type:extra strand C' between strand C and D. Sheet I: ABE, sheet II: CFG. Strand C'/D links sheet I and II:

• N-terminal domain of cellulase c.
• C-terminal domain of galactose oxidase.

We developed a profile based on structural alignments that covers the whole domain and recognizes Ig-like domains closely related to immunoglobulin. Our profiles does not recognise H-type or DNA binding Ig-like domains, because of their high divergence.