NMR STRUCTURAL STUDIES OF THE FIRST CATALYTIC HALF-DOMAIN OF UBIQUITIN ACTIVATING ENZYME − Rakyat Biologi

Structure of the FCCH domain

The solution structure of the mouse FCCH subdomain presented in this work agrees closely (RMSD of structured part Cα backbone atoms superposition equals 1.142 Å) with the structured part of the corresponding subdomain determined for the yeast E1 (Lee and Schindelin, 2008) (Fig. 4). A full turn around the 6-stranded β-barrel (in the direction of the hydrogen bonds) corresponds to a 10-residue shift around the barrel, typical for 6-stranded β-barrels (Murzin et al., 1994). The DALI scan (Holm and Rosenström, 2010) of the Protein Data Bank using FCCH as a search model identified β-barrel proteins with strand number n = 6 and shear number S = 10 as the most similar (Table 2). The selected proteins are involved in different cellular processes with no obvious connection to the Ub system. Therefore, the hits most likely show only structural similarity without functional implications.

Relaxation studies of the FCCH subdomain

The central fragment of the FCCH (217 – 293) is characterized by low backbone mobility typical for structured proteins, while unstructured C and N termini show much higher mobility (Figs. 5 and 7). Several residues located in loops are characterised by chemical exchange. Moreover, according to the crystal structure, the side chains of residues Arg202 (equivalent of Arg239 in mouse FCCH), Gly204 (equivalent of Gly241 in mouse FCCH) and Glu206 (equivalent of Glu243 in mouse FCCH) of yeast Ub-E1 form hydrogen bonds with Ub (cf. Fig. 8). All three residues are conserved and situated within the loop between b2 and b3 strands (Leu235 - Gly245 region), in which residues exhibit elevated R_ex values, the hallmark of conformational exchange in the micro- to millisecond time scale. If the increased mobility could facilitate the adaptation of an appropriate interface conformation, the FCCH may have a role in the regulation of Ub binding to the E1 enzymatic machinery in the first stages of the Ub activation process, in agreement with an earlier suggestion that the FCCH may help to distinguish ubiquitin from other ubiqutin-like proteins (Viquez et al., 2012). This finding encouraged us to investigate the interaction of FCCH and Ub in solution. However, the FCCH titration with ubiquitin showed no visible chemical shift changes in the 2D ¹H/¹⁵N HSQC spectra of FCCH. Hence, the FCCH in isolation (i.e. without the context of full length E1) does not bind to Ub in solution (Cavanagh et al., 2007).

Sequence alignment of the FCCH subdomain

FCCH domains of yeast, mouse and human Ub-E1s were aligned using MUSCLE (Edgar, 2004) (Fig. 4B). Conserved residues are highlighted in black. The sequences share 58% of identity. Such high conservation in evolutionarily very distant species indicates significant selective pressure on this domain. Conserved residues are approximately equally distributed on the outside of the barrel, therefore no obvious protein-protein interfaces can be delineated from the structure of the FCCH fragment alone. In the yeast E1–Ub crystal structure FCCH forms one of the walls of the Ub adenylation pocket (Fig. 8). In the crystal, the side chains of Arg202 and Glu206 form hydrogen bonds with ubiquitin. In mouse E1 FCCH, both residues are situated within the Glu237 - E243 region, where residues exhibit conformational exchange in the micro to millisecond time scale. However, these residues are not among the most conserved residues in eukaryotes.

Inspection of the yeast E1 crystal structure (PDB id: 3CMM)(Lee and Schindelin, 2008) revealed that some contacts between FCCH and SCCH domains were present (Fig. S.2). FCCH and SCCH domains are evolutionary conserved and structurally similar. The backbone RMSD values between mouse catalytic half-domains and their yeast counterparts are 1.142 Å and 0.890 Å for FCCH and SCCH domain, respectively. These observations prompted us to test whether the mouse FCCH and SCCH domains interact in solution when not linked covalently by the adenylation domain. However, chemical shift perturbation in the ¹H/¹⁵N HSQC spectrum of the FCCH/SCCH equimolar mixture was not observed. This finding points to a lack of interaction between these two catalytic half-domains in solution within the mM concentration range (Cavanagh et al., 2007). This result could either indicate that interactions are only observable when the two domains are covalently linked as in the context of full-length E1, or could indicate that the interactions that are found in the crystal might be due to crystal packing.