CAPÍTULO II: MARCO TEÓRICO
2.2. FUNDAMENTACIÓN TEÓRICA
2.2.2. Gestión del talento Humano
For some time the PDI protein was considered to be the only enzymatic catalyst of thiol-disulfide exchange reactions in the ER. However, it is now known that PDI is just one protein in an ever-growing family. The PDI family, as it became known, is named after its archetypal and most abundant member, the PDI protein (Hatahet and Ruddock 2009). The PDI family of proteins are a subfamily of the thioredoxin superfamily.
PDI family members are believed to be the only enzymes which catalyse complex rate limiting isomerisation reactions (Weissman and Kim 1993). Proteins are classified as members of the family based on their structure, sequence similarity and ER localisation, rather than by physiological function. All contain at least one domain with a thioredoxin-like fold, which can be catalytic or non-catalytic in function. They also contain a cleavable N-terminal ER targeting signal sequence. Most PDI family members are ER localised with the aid of a C- terminal KDEL ER retention signal, or a similar variant of this motif. Since all PDI family members have distinct features and differing tissue distribution, they are all likely to have distinct physiological functions, although these may partially overlap (Hatahet and Ruddock 2009). This is supported by evidence that yeast PDI family members are not functionally interchangeable (Norgaard, Westphal et al. 2001).
Different organisms have different subfamilies of PDI. Whilst some clearly have functions similar to orthologs in other species, such as human PDI and yeast Pdi1p, many appear to be unique to a subset of organisms. Only 5 PDI family members have been identified inS. cerevisiae, whereas 20 have currently been defined in humans (Hatahet and Ruddock 2009), see Figure 1.9.
Figure 1.9: Illustration of human PDI family members. Yellow, catalytic domains likely to have a thioredoxin-like fold, with their active-site sequence inset; blue, non-catalytic domains likely to have a thioredoxin-like fold; red, transmembrane regions (Hatahet and Ruddock 2009).
Although detailed characterisation has yet to be undertaken on many PDI family members, it appears that only a small subset combine a generalised substrate binding domain with a catalytic domain, a combination that has been shown to be essential for efficient isomerisation by PDI (Darby, Penka et al. 1998). Indeed, many members of the PDI family (e.g. ERp27) do not have any active sites (Alanen, Williamson et al. 2006). Hence, perhaps the classification or nomenclature of this family of ER resident folding catalysts needs to be re- evaluated.
The only two human PDI family members that show the same domain architecture as PDI, i.e. abb’a’c, are ERp57 and PDIp. They all have two catalytically active a and a’ domains with similar active site residues, see Figure 1.9. Other than PDI itself, ERp57 is the most characterised of all the PDI family members (Ferrari and Soling 1999).
DeSilva, Notkins et al. 1997), while PDILT expression is highly specific to the testes after puberty and is the only PDI family member to show developmental control (van Lith, Hartigan et al. 2005; van Lith, Karala et al. 2007).
An interesting subset of the PDI family is the thioredoxin-related transmembrane proteins, which span the membrane of the ER. Five human transmembrane PDI family members have been proposed and their transmembrane regions identified, see Figure 1.9 (named with TMX prefix). All have type I transmembrane protein ER localisation signals (Appenzeller-Herzog and Ellgaard 2008), although only three of these have been published: TMX (Matsuo, Akiyama et al. 2001), TMX2 (Meng, Zhang et al. 2003) and TMX3 (Haugstetter, Blicher et al. 2005). Hence, there is a lot of scope to increase understanding of these ER resident, membrane bound proteins.
Another interesting human PDI family member is ERdj5. It is the only family member to contain a DnaJ domain (labelled J domain in Figure 1.9) (Cunnea, Miranda-Vizuete et al. 2003). This domain associates with BiP (an Hsp70 chaperone), thus implying a role for ERdj5 in ER-associated degradation (ERAD) (Hosoda, Kimata et al. 2003). Recent studies have revealed details of this role, which requires both the DnaJ domain and the redox activity of the catalytic domains for reductase activity of misfolded ER proteins (Dong, Bridges et al. 2008; Ushioda, Hoseki et al. 2008).
Until 2006, calsequestrin was the closest protein to PDI to have a full three dimensional structure (Wang, Trumble et al. 1998). It is a 40 kDa protein consisting of three thioredoxin-like domains and is involved in the regulation of Ca2+ ion channels in the sarcoplasmic reticulum of muscle cells (Kawasaki and Kasai 1994).
Although no structure for full length human PDI has been achieved, a full length structure of PDI fromSaccharomyces cerevisiae (yeast) has been solved by X-
ray crystallography (Tian, Xiang et al. 2006). Like human PDI, yeast PDI (Pdi1p) incorporates all four thioredoxin domains, see Figure 1.10.
Figure 1.10: Alignment of human and yeast PDI domains. a) Domains showing boundary residues and residue numbers for mature human PDI, aligned with the homologous regions of yeast PDI. Residues flanking the active sites in the a and a’ domains are highlighted. b) Crystal structure of yeast PDI (Pdi1p), PDB ID 2B5E. Active sites are highlighted green. Figure from (Gruber, Cemazar et al. 2006).
Sequence similarity between PDI and Pdi1p is high, so the structure has provided valuable insight into its human ortholog. The hydrophobic cleft is speculated to be large enough to accommodate a folded protein of approximately 100 residues, a figure which is important to consider when studying the interaction of PDI and substrate proteins (Tian, Xiang et al. 2006). More recently, three crystal structures of human PDI family members have been
the human PDI family in that as well as having a non-catalytic thioredoxin-like domain it also has a non-thioredoxin, entirely helical domain at its C-terminus, termed the D-domain (Liepinsh, Baryshev et al. 2001). It is very similar in structure and function to its ortholog Wind protein, a PDI-related protein found in Drosophila melanogaster(Lippert, Diao et al. 2007).
A crystal structure of ERp57 was recently solved in a complex with the substrate protein tapasin, to a 2.6 Å resolution (Dong, Wearsch et al. 2009). This is particularly interesting, given that ERp57 is a human PDI family member that shares the same domain architecture as PDI. Since its structure is solved with a bound natural substrate, this provides useful information about the nature of enzyme-substrate binding for all PDI family members. ERp57 is known to act in a complex to catalyse disulfide bond isomerisation specifically in N-glycosylated protein substrates (Russell, Ruddock et al. 2004). A knockout of ERp57 was found to be lethal in embryonic mice (Garbi, Tanaka et al. 2006). However, knockout of ERp57 in individual cells showed that very few proteins were affected (Solda, Garbi et al. 2006), indicating that ERp57 is involved in the folding of very few proteins or that other PDI family members are able to compensate for it.
A crystal structure of ERp44 was recently solved to 2.6 Å resolution (Wang, Wang et al. 2008). The protein consists of three thioredoxin-like domains (a, b and b’), with a C-terminal extension (Figure 1.9). The structure shows that the three thioredoxin domains form a clover shape, with the C-terminal extension occluding the substrate binding site and partially shielding the active site. It is thought that the C-terminus may act to mediate substrate access to the enzyme (Wang, Wang et al. 2008). ERp44 was initially identified as a ER-located binding partner for Ero1 (Anelli, Alessio et al. 2002). However, despite having a RDEL ER retention motif (Raykhel, Alanen et al. 2007), ERp44 is an unusual PDI family member in that it has since been reported to be co-localised to the ER, ER-Golgi intermediate compartment and cis-Golgi (Gilchrist, Au et al. 2006).
Interestingly, its only active site has an unusual motif, CRFS, since it lacks the C-terminal cysteine. It has been proposed that the absence of a second cysteine allows mixed disulfides between ERp44 and its substrate to persist for much longer, thus facilitating its proposed function to assist non-native protein transport from the ER (Anelli, Ceppi et al. 2007).
The only structural data available from full length human PDI comes from low resolution small angle X-ray scattering (SAXS) (Li, Hong et al. 2006). More recently, the same technique was used to characterise ERp72 (Kozlov, Maattanen et al. 2009).
Many PDI family proteins interact with their substrate primarily through a conserved binding pocket located in the non-catalytic b’ domain (Ellgaard and Ruddock 2005). In some of these proteins, the binding site may have become specialised for the binding of particular substrates. For example, ERp57 is a PDI family member that forms a complex with either of the chaperone proteins calnexin or calreticulin (Russell, Ruddock et al. 2004). This complex then acts to catalyse disulfide bond isomerisation specifically in N-glycosylated protein substrates (Coe and Michalak 2010). The b domain seems to have a less crucial role in substrate binding and is thought to play a mainly structural role in multi-domain PDI proteins.