Multiple sequence alignment of the b′-like domains of the human protein disulphide isomerase family. The alignment was constructed from a large number of single and multiple alignments and adjusted by hand, taking secondary structure predictions into account. Amino acids identical or similar to each other in four or more proteins are highlighted. Residues in which mutation inhibits the interaction of ERp57 with calreticulin (Russell et al, 2004) are indicated with an asterisk below the alignment, whereas those in which mutation inhibits the interaction of protein disulphide isomerase (PDI) with the peptide substrate Δ-somatostatin (Pirneskoski et al, 2004) are indicated with an asterisk above the alignment. At six of the seven sites implicated in ERp57–calreticulin interactions, ERp57 and ERp72 share identity or similarity with each other but not with the PDI/PDIp/PDILT/ERp27 cluster.

Schematic overview of the human protein disulphide isomerase family. Thioredoxin-like domains are represented by rectangles with the active-site sequence added for catalytic domains (yellow). The catalytically inactive b′ domain is shown in blue and other catalytically inactive domains are in green. The linker region between b′ and a′ domains is coloured in grey, and transmembrane regions are shown in red. Signal sequences are not shown. Note that the catalytic domain of ERp18 contains a putative 23-amino-acid insert between β3 and α3 (Alanen et al, 2003) that is not shown in the figure.