Human Putative Ubiquitin Conjugating Enzyme

To begin studying the biological role and molecular function of the human ubiquitin conjugating enzyme, FLJ11011, we have determined its high resolution crystallographic structure. We show that FLJ11011 folds into the canonical UBC fold but with a prominent variation: domain swapping. The first alpha helical element and the first beta strand element (first 33 residues of the protein) are necessary and conserved components of the UBC fold and are being swapped.

This swapping generates a homodimer; the only homodimeric ubiquitin conjugating enzyme system known. The first beta strand (11 residues) interacts in a continuous 19-residue, two-stranded, beta sheet with the second beta strand of the other molecule (8 residues). The dimerization appears to be very stable as suggested by analysis of a number of subconstructs.
For example, dimeric gel-filtration peaks were detected unless constructs did not contain a significant portion of the amino terminal helix (amino terminal truncations resulted in monomeric peaks, low yields, and aggregation). And, carboxy terminal truncations did not affect oligomerization but instead decreased solubility. Strong interactions include the following core-forming residues of the first 2 secondary structure elements: M4, L8, L12, L15, M23, L25, and V30. In addition to the swapped domain, a large surface area of interaction occurs, including the active site loops and a symmetric tetrad, formed by M72 and K59. These observations further suggest that the relative orientation of the domains is fixed.

The presence of a cysteine (C91) in the putative active site suggests that FLJ11011 is a catalytically competent conjugating enzyme. The dimer orientation puts the putative ubiquitin binding sites (centered on alpha helix 2) on opposite sides of the dimer, with no possibility of ubiquitin/ubiquitin interaction. The regions of the ubiquitin binding sites that come nearest to each other are the reactive cysteines, which are 24.8 Angstroms apart. The orientation of the dimer restricts access to the canonical active site: the dimer interface could block the binding of typical substrates. On the other hand, two novel clefts that include the reactive cysteines are generated between the paired molecules.

The putative E3 binding sites (centered on the very amino terminus of the first alpha helix) occur on the same face of the FLJ11011 dimer. FLJ11011 may be recognized by dimeric E3 ligases, and could suggest a role for FLJ11011 in polyubiquitylation, chain elongation, branching, or cross-linking. FLJ11011 has a unique sequence, with no significant identity with other members of the 30 or so known human ubiquitin conjugating enzymes; its biological role may also be unique.