The common screening test for CDG is the isoelectric focusing of the serum protein, transferrin (Tf). Tf has two N-glycosylation sites, both occupied with biantennary or triantennary disialylated glycans, resulting in negatively charged species. In the case of type I CDG, a subpopulation of Tf molecules lacking one or both N-glycan chains leads to a reduction in the negatively charged species (carbohydrate deficient Tf). In type II CDG the N-glycosylation sites are occupied but the glycosylation is incomplete, resulting in different glycoforms with 0, 1, 2, 3, 4 or 5 sialic acids. The Tf serum test is fast, cheap and generally reliable. Recently HPLC and mass spectrometric analysis of carbohydrate deficient Tf by either electrospray ionization or matrix-assisted laser desorption/ionization, have proved to be efficient analytical methods. Often a number of protein variants of transferrin are present in the serum which can affect the isoelectric point of the molecule (Ohno et al. 1992). There are some disorders, such as uncontrolled galactosemia (Charlwood et al. 1998; Sturiale et al. 2005), fructose intolerance (Adamowicz et al. 1996; Jaeken et al. 1996), and heavy alcohol consumption (del Castillo Busto et al. 2005) which can mimic CDG profiles leading to false positive results. To avoid these pitfalls of diagnosis a multiple test can be used to screen CDG patients. Here the Tf isoelectric focusing screen is combined with MALDI-TOF analysis for glycosylation site occupancy and analysis of the glycan structures on Tf glycopeptides (Wada 2006). To identify the subtype of CDG, the most time-consuming step, the identification of the defective gene in each patient is needed. This is usually done by assaying for the altered enzymatic activity.
Since the nature of the O-glycosylation defects are so diverse, testing only one specific glycoprotein would not lead to a diagnosis. So far, the only screening method that is available for the detection of defects in O-mannosylation, is by immunohistochemical staining of the α-dystroglycan with monoclonal antibodies (Wopereis et al. 2006). Recently a novel immunopurification-mass spectrometry method was developed to confirm the suspected glycosylation defects, occurring in Peters Plus syndrome (Hess et al. 2008). Patients with this syndrome have biallelic truncating mutations in the gene encoding β3Glc-T (Lesnik Oberstein et al. 2006), that is responsible for the glucosylation of O-fucosylated TSR repeats (Kozma et al. 2006; Sato et al. 2006). This approach highlights the importance of analysing discrete protein domains, TSR in this case, to detect changes in rare forms of glycosylation.
5.3. Thrombospondin Type 1 Repeats
Thrombospondin type 1 repeats (TSRs) were first identified in the multi-domain extracellular matrix protein thrombospondin 1 (TSP-1) (Lawler et al. 1986). TSRs are conserved in evolution with 31
Drosophila, 28 C. elegans and 141 human proteins. These proteins often consist of multiple modules
with the TSRs secreted extracellularly. The TSRs are functionally important in the regulation of extracellular matrix organization, cell-cell interaction, axonal and cell guidance (Tucker 2004).
There are only two thrombospondin TSR containing proteins, TSP-1 and TSP-2, that each of them has three tandem TSRs located between the vWC and the EGF-like domains. TSRs are not only found in TSPs but also in other multidomain proteins, including F-spondin, SCO-spondin, UNC-5, semaphorin-5, thrombospondin-related anonymous protein (TRAP), thrombospondin-related sporozoites protein (TRSP) and circumsporozoite protein (CSP) from the Plasmodiumfalciparum
parasite, properdin, complement proteins C6, C7, C8, C9 and the ADAMTS (A Disintegrin And Metalloproteinase with ThromboSpondin type 1 repeats) family (Adams et al. 2000).
Figure 9. Domain structure of some TSR-superfamily members
The length of the protein corresponds to the amino acid scale at the top of the figure. The TSR domains are indicated by red diamonds. The cross-hatched boxes represent the transmembrane domains. vWC, von Willebrand factor type C domain; 2, calcium-binding EGF-like domain; Reelin, Reelin domain; Sp, Spondin domain; 7-SPAN, seven transmembrane spanning domain; vWA, von Willebrand factor type A domain; CT, C-terminal cysteine knot-like domain; DEATH, death apoptosis domain; Di, disintegrin-like domain; FI, factor I membrane attack complex; GPS, latrophilin/CL-1-like G-protein-coupled receptor proteolytic site; IB, insulin growth factor-binding protein domain; Ig, immunoglobulin domain; L-R, LDL receptor class A; MACPF, membrane attack complex/perforin domain; MPase, metalloprotease domain; PL, plexin repeat; Pro, reprolysin family propeptide; SEMA, semaphorin domain; Su, Sushi (SCR) domain; TEV, domain that interacts with PDZ-containing proteins; TSC, thrombospondin C-terminal domain; TSN, thrombospondin N-terminal domain; TSR3, thrombospondin type 3 domain; ZU5, domain present in ZO-1 and UNC-5.
Adapted from Tucker, 2004 Sp
TSRs are ~60 amino acids long with around 12 conserved residues (highlighted in Figure 10). The crystal structure of TSR2-TSR3 from human TSP-1 revealed that TSRs fold as three anti-parallel strands (Tan et al. 2002) (Figure 11). The B and C strands form limited regular β-sheets, whereas the A strand has a unique rippled conformation and contains the conserved sequence motif WXXWXXW. The tryptophan side chains form a layer with the conserved arginines guanidinium groups from the B strand filling in the spaces between the indole moieties. The top and bottom of this stacked core is capped by disulphides.
Figure 10. Multiple sequence alignment of the two major groups of TSRs
The disulphide patterns of groups 1 and 2 are drawn schematically as yellow lines on the left-hand side. For group 1, two hydrogen bonds between the jar handle and the N-terminus are drawn as red lines. In the alignment the paired cysteines are also shown in yellow. The putative recognition motif for O- fucosylation (in bold) is indicated under the alignment.
Interestingly, it appears that there are two main groups of TSR-containing proteins based on their disulphide bond pairings (Figure 10). The main difference occurs at the top of the TSR layered structure. For TSRs in group 1, which includes members of TSP, BAI, ADAMTS and properdin, the top cysteine bridge is formed by cysteines from the end of the B strand to the beginning of the C strand (Cys3-Cys4). In contrast, TSRs in group 2, which includes F-spondin and TRAP, the top
disulphide is formed at the start of the A and C strands (Cys1-Cys4). This suggests that the N-terminus in group 2 TSRs is stabilised by this disulphide bridge rather than the jar handle found in group 1 TSR (Figure 10). At the bottom of the TSR domain in both groups 1 and 2, the AB loop is stabilised by a
Modified from Tan et al., 2002
Group 2 Group 1
disulphide bond to the carboxyl terminus of the C strand. In addition, the most N-terminal tryptophan is poorly conserved in TSRs from group 2.
The NMR solution structures of rat F-spondin TSR1 and TSR4, as well as TRAP-TSR from
P. falciparum have also been solved (Paakkonen et al. 2006; Tossavainen et al. 2006). There is a
similar fold between these group 2 TSR containing proteins and the group 1 TSR2-TSR3 from TSP-1, with the overall core and AB loop superimposing well. Surprisingly there is some increase in flexibility in the BC loop region of F-spondin TSR4 compared to F-spondin TSR1 (Paakkonen et al. 2006).