Ubiquitination is a fundamental biochemical process, which controls numerous aspects of protein function, such as degradation, protein-protein interaction and subcellular localization [1]. The attachment of the 8 kDa protein ubiquitin (Ub) to proteins involves three classes of enzyme, an E1 ubiquitin-activating enzyme, an E2 ubiquitin-conjugating enzyme, and an E3 ubiquitin ligase. The C-terminus of Ub first forms a thioester bond with the catalytic cysteine of the E1 in an ATP-dependent manner. Ub is then transferred from the E1 to the catalytic cysteine of the E2. Finally, the E3 binds both the Ub-charged E2 and substrate to catalyze transfer of the C-terminus of Ub to a substrate lysine to form an isopeptide bond, resulting in substrate monoubiquitination. Substrates can be ubiquitinated on numerous lysines, resulting in multiubiquitination [2, 3]. In addition, some E2/E3 combinations can then utilize lysines on the substrate-conjugated ubiquitin, to catalyze further cycles of ubiquitination, resulting in substrate polyubiquitination [1, 3]. Ub contains seven lysines, which can be utilized during polyubiquitin chain formation, and in most cases a specific lysine is utilized by a particular E2/E3 pair [3]. The ability to generate diverse substrate-ubiquitin structures is important for targeting proteins to different fates. For example, monoubiquitination can regulate DNA repair and gene expression [4]. Polyubiquitination through Ub K48 generally targets proteins for proteasomal degradation, while K63-linked Ub chains can regulate kinase activation, DNA damage tolerance, signal transduction and endocytosis [4] (Figure 1).

The mechanisms that control lysine selection in substrates are not clearly understood. Structural aspects of E2/E3 s and how they bind the substrate are believed to be important. Therefore, a "positioning model" posits that the E3 positions the substrate toward the E2~Ub thioester bond to select particular lysines during substrate ubiquitination [3, 5]. However, some substrates are known to be ubiquitinated on numerous lysines, e.g. the budding yeast CDK inhibitor Sic1 [2]. It is believed that binding of the substrate through several binding motifs to the E3 results in various binding geometries, leading to ubiquitination on numerous lysines. Other studies have suggested that substrate lysine selection flexibility may be achieved by release of the ubiquitin-charged E2~Ub from E3, to transfer Ub to different substrate lysines, as proposed by the'hit and run' model [6]. During polyubiquitin chain extension structural features in E2 s position Ub lysines to generate Ub chains via a specific lysine [7, 8]. For example, the Mms2/Ubc13-Ub complex assembles so that K63 of Ub attacks the Ubc13-Ub thioester bond during Ub chain formation [9]. Similarly, an acidic loop region in the E2 Cdc34 is thought to position K48 of Ub for attack of the Cdc34~Ub thioester bond [10]. Other E2 s, such as human UbcH5, can utilize several Ub lysines (K11, K48 and K63), indicating less structural constraint and that other mechanisms may also contribute to Ub lysine selectivity [8].

Apart from higher order structures of E2 s and E3 s contributing to lysine selection in substrates and Ub, it is not clear if amino acid determinants within the catalytic region of E2 s and those surrounding substrate and Ub lysines play a role in lysine selectivity. At a catalytic level, ubiquitination of substrate or Ub lysines occurs by nucleophilic attack of the lysine residue on the E2~Ub thioester bond. Recent studies with the human anaphase promoting complex (APC/C) RING E3 and its E2, UbcH10, have identified a sequence motif adjacent to acceptor lysines, termed the TEK-box, which is important for ubiquitination [7], suggesting that amino acid determinants near the lysine residue may play a crucial role in lysine selection. Initially, insights into the importance of amino acids surrounding acceptor lysines and the E2 catalytic core have come from studies of the SUMO-conjugating E2, Ubc9. Ubc9 attaches the ubiquitin-like protein SUMO onto substrate lysines in an analogous fashion to that used by ubiquitin-conjugating E2 s. Structural studies of Ubc9 complexed with its substrate, RanGAP1, show that amino acids in proximity of the catalytic cysteine make important contacts with amino acids surrounding RanGAP1 K526, which is sumoylated. Hence, Ubc9 Y87 and A129 make van der Waals contacts with L525, S527 and E528 of RanGAP1, which are proximal to the sumoylated K526, while Ubc9 D127 is within hydrogen-bonding distance of sumoylated RanGAP1 lysine 526 [11, 12] (Figure 2). These interactions facilitate sumoylation through optimal alignment and pK suppression for nucleophilic activation of the attacking RanGAP1 K526 [11, 12].

Since previous studies indicate that amino acids in the SUMO E2 catalytic site interact with amino acids surrounding the attacking lysine to help position it for nucleophilic attack [12], we explored if analogous sites contribute to lysine specificity in ubiquitin-conjugating E2 s. We utilized the yeast RING E3 ligase Skp1-Cdc53/cullin-F box protein (SCF) and its cognate E2, Cdc34 to assess ubiquitination of their physiological substrate, Sic1 [3]. In S. cerevisiae, G₁-S phase cell cycle progression depends on SCF^Cdc4/Cdc34-mediated polyubiquitination of Sic1 via K48-linked Ub chains, leading to its proteasomal degradation [2, 13, 14].

Ubiquitin attachment to specific lysine residues in protein substrates and itself is important for generating diverse substrate-ubiquitin structures, providing versatility to this pathway in targeting proteins to different fates. The mechanisms of lysine selection, however, remain unclear. While it has been appreciated that higher order structural features of E2 s and E3 s are important for lysine selection during ubiquitination, our recent studies showed that compatibility between key residues within the E2 catalytic region and those proximal to acceptor lysines in the substrate or Ub provides a further level of specificity control. The interplay of both mechanisms ultimately defines the efficiency of ubiquitination of a lysine. We speculate that E2 s may have evolved divergent structures in their catalytic regions to modulate lysine specificity. This likely contributes to the mechanisms for different E2/E3 combinations generating structural diversity by controlling if a protein is monoubiquitinated, multiubiquitinated or polyubiquitinated. Further biochemical and structural studies are required to define all of the important determinants in the catalytic region of Cdc34 and residues proximal to acceptor lysines to comprehensively define the interaction interface of the Cdc34-substrate complex. In addition, since we propose that this mechanism is generally important for lysine selectivity during ubiquitination, it will be important to characterize other E2 family members for lysine specificity.