Post-translational covalent attachment of ubiquitin-like proteins (UBLs) is a predominant eukaryotic regulatory mechanism. More than a dozen UBLs—such as ubiquitin, NEDD8, ISG15, and SUMO—covalently modify myriad substrates. Different UBLs alter the functions of their target proteins in different ways, such as by changing the target's half-life, conformation, localization, enzymatic activity, and/or intermolecular interactions. Defects in UBL pathways have been associated with numerous diseases, including cancers, neurodegenerative disorders, and viral infections. Thus, we are determining structural and biochemical mechanisms by which enzymes transfer UBLs.
UBLs are attached to protein targets by a series of molecular "handoffs" involving an E1 activating enzyme, an E2 conjugating enzyme (or Ubc), an E3 ligase, and the target. First, at the apex of each UBL's cascade, a dedicated E1 enzyme activates its UBL and ultimately catalyzes UBL transfer to an E2's catalytic cysteine to generate a reactive, thiolester-linked E2~UBL covalent product. The E2~UBL complex typically associates with an E3, which facilitates ligation of the UBL to a target. A vast number of enzymes are associated with UBL cascades. The ubiquitin pathway is the most expansive. In humans, for example, ubiquitination is thought to involve two E1s, tens of E2s, hundreds of E3s, and thousands of protein targets with diverse biological functions.
Our research is focused on understanding (1) basic enzymatic mechanisms underlying E1-E2-E3-mediated UBL attachment to targets, (2) how UBLs are attached selectively, and (3) how UBL modifications alter target functions.
NEDD8: A Ubiquitin-like Protein That Regulates Ubiquitination
Much of our work has focused on the UBL NEDD8. Despite nearly 60 percent sequence identity to ubiquitin, NEDD8 has a completely distinct function: NEDD8 switches on a family of ~300 ubiquitin E3s (see below). NEDD8-activated ubiquitin ligation is thought to mediate ~20 percent of all ubiquitin-proteasomal degradation. We are both defining components of the NEDD8 pathway and also studying the NEDD8 enzyme cascade to follow a UBL from beginning to end: How does NEDD8 enter the cascade via E1, move from enzyme to enzyme to target, and regulate its targets to promote their ubiquitin E3 activities? Sequences of NEDD8 E1-E2-E3 enzymes resemble those for several other UBLs. Our goals are to understand both general mechanisms of UBL transfer and NEDD8-mediated regulation.
Mechanisms of UBL Activation and Conjugation
E1s are fascinating enzymes that efficiently select the correct UBL for its pathway, chemically activate the UBL's carboxyl terminus, and coordinate the UBL with the correct downstream pathway, through four catalytic steps. We established the molecular basis for several of these E1 activities by determining crystal structures of the NEDD8 E1 in various complexes with NEDD8, MgATP, and E2s. Our structural studies provided the first detailed insights into the noncovalent E1-E2 interactions involving domains common to all E1s and E2s. Our data indicated that E1 and E3 binding to E2s is mutually exclusive, with a single E2 surface shuttling back and forth between E1 for UBL loading and E3s for UBL ligation to a target. A fundamental implication of this is that polyubiquitination involves multiple E2-E3 binding events. These results also fueled our interest in understanding how E1s bind free E2s and release E2~UBL products, and how E3s bind E2~UBLs and release free E2s.
To address how an E1 might bind a free E2 and release an E2~UBL product, we determined a crystal structure comprising an E1~NEDD8-NEDD8 intermediate bound to a NEDD8 E2 (UBC12) before NEDD8 transfer. The data suggested that the fundamental feature of UBL cascades—transferring the UBL's thiolester linkage between successive conjugation enzymes—drives consecutive steps in UBL cascades by altering enzyme conformations and interactions.
Specificity of NEDD8 Ligation
How is specificity established for a particular UBL cascade? We deciphered "codes" for selective E1-UBL, E1-E2, and E2~UBL interactions in the NEDD8 pathway, and also features hindering ubiquitin conjugation. Using this information, we discovered a second, uncharacterized NEDD8 E2 (UBE2F) and changed the view of NEDD8ylation. The best-recognized function of a NEDD8 E2 is to modify and activate cullins (e.g., CUL1, 2, 3, 4A, 5). Cullins partner with a RING protein (RBX1 or 2) to nucleate the largest ubiquitin E3 superfamily, the cullin-RING ligases (CRLs). Previously, NEDD8 ligation to cullins was thought to be nonspecific. However, we defined distinct NEDD8 conjugation pathways for different cullins, with UBE2F/RBX2 modifying CUL5. Specificity is dictated by RBX RING/NEDD8 E2 pairings. To our knowledge, this was the first demonstration that RING domains dictate E2 specificity. This has broad implications for how the large family of RING E3s recruit E2s. Furthermore, because CUL5 is hijacked by many viruses, including HIV, our identification of UBE2F/RBX2-mediated CUL5 NEDD8 modification may have therapeutic applications.
Discovery of a "Dual E3" Mechanism for NEDD8 Ligation
The prevailing view was that a thiolester-linked E2~UBL intermediate interacts with one E3 for UBL transfer to a target. We identified a novel mechanism, in which one E2 functions simultaneously with two E3s for optimal UBL ligation. We started by asking this question: What are E3s for NEDD8? It seemed likely that RBX proteins serve as NEDD8 E3s, given our findings that RBX1 and RBX2 RING domains interact with distinct NEDD8 E2s and that RING domains are known to mediate E3 activity. However, a different protein, called Dcn1, had previously been reported as being a NEDD8 E3. We dissected the roles of RBX1 and Dcn1 and found that both are E3s. RBX1 functions like a conventional RING E3 by recruiting the cullin and promoting NEDD8 transfer from the UBC12 active site. Dcn1 functions only in the presence of RBX1's RING domain, to impart specificity for UBC12~NEDD8 rather than an E2~ubiquitin complex, to reduce nonspecific UBC12~NEDD8 reactivity, and to bring UBC12's active site to the cullin target. Based on structural data and modeling, we propose that the second E3 (Dcn1) restricts the otherwise flexible RBX1 RING-UBC12~NEDD8 to a catalytically competent orientation. The two E3s synergize to massively increase catalytic efficiency of UBL transfer to a target.
Regulation, Assembly, and Mechanisms of the Largest Classes of Ubiquitin E3s
Two main classes of E3s are primarily responsible for directing a wide range of ubiquitin modifications to specific targets. RING E3s stimulate ubiquitin transfer from E2s to targets, whereas HECT E3s utilize a covalent HECT~ubiquitin intermediate. We are studying how these classes of ubiquitin E3s work.
NEDD8 activation of cullin-RING ligases: conformational control of RING E3-mediated UBL ligation. The largest E3 subfamily consists of the modular, multiprotein CRLs. For prototypic CRLs in the SCF family, CUL1 binds to Skp1, which in turn binds to an F-box protein substrate receptor. Each of the 69 mammalian F-box proteins is thought to recruit a distinct cohort of substrates to CUL1-RBX1 for ubiquitination. Here, RBX1 is a ubiquitin E3: RBX1 binds a ubiquitin-conjugated E2 and promotes ubiquitin transfer to hundreds of targets to regulate processes including the cell cycle, signaling, transcription, and circadian rhythms. Other cullin-RING complexes function in a parallel manner, assembling with their own sets of substrate receptors to generate ~300 putative ubiquitin E3s in humans.
Previous structural studies of SCFs and other CRLs presented major topological conundrums: a >50-Å gap separating the E2 active site and substrate, and a view that CRL structures are rigid, raised the question of how an E2 active site could be juxtaposed with the substrate for ubiquitin transfer. It was also unclear how an RBX1-bound E2 could add ubiquitin to the growing end of a polyubiquitin chain.
We considered that prior structures were of inactive forms of CRLs because they lacked NEDD8. NEDD8 ligation to a cullin was known to stimulate ubiquitin ligase activity, through unknown mechanisms. Our data indicated that NEDD8 ligation favors striking CRL conformational changes to impart multiple catalytic geometries to an associated E2. We predict that other RING E3s are also regulated by conformational control mechanisms.
Structural mechanisms for assembly of selected CRLs. The largest CRL subfamily involves CUL3 binding to substrate adaptors in the BTB class. BTB proteins have a BTB domain, which both binds CUL3 and dimerizes, and a protein-interaction domain such as MATH or Kelch that binds a substrate for ubiquitination. We provided insights into the assembly of CUL3-based CRLs through characterization of the BTB protein SPOP. Through crystal structures and biophysical and biochemical analyses, we identified a SPOP-binding consensus (SBC) motif and determined the structural basis for recognition of SBCs from diverse substrates. Our crystal structures suggested that phosphorylation of the SBC motif would impede binding to SPOP, and we confirmed this through biophysical studies. Furthermore, our results suggested that structural flexibility within the SPOP dimer may allow avid interaction with structurally diverse substrates containing two or more SBCs. Overall, our data elucidated mechanisms controlling recruitment and ubiquitination of SPOP substrates and will serve as a basis for future efforts to understand assembly of BTB-based and other dimeric CRLs.
Mechanism for ubiquitin transfer from an E2 to a HECT E3. E3s in the HECT family form the second largest class of ubiquitin ligases and play critical roles in cellular regulation and diseases, including cancers, hypertension, and retroviral infections. HECT E3s transfer ubiquitin via a distinct mechanism: the catalytic HECT domain binds a thiolester-linked E2~ubiquitin intermediate; ubiquitin is transferred first to the HECT domain catalytic cysteine and then from the HECT E3 to the target. To gain insights into the first steps in this process, we determined the crystal structure of a complex between the HECT domain of NEDD4L and the E2 UbcH5B bearing a covalently linked ubiquitin at its active site (UbcH5B~ubiquitin). Extensive noncovalent interactions between UbcH5B and ubiquitin with the HECT domain lead to an overall compact structure, with ubiquitin's carboxyl terminus sandwiched between UbcH5B and HECT domain active sites. The structure suggested a model for E2-to-HECT E3 ubiquitin transfer, in which interactions between a donor ubiquitin and an acceptor domain constrain upstream and downstream enzymes for conjugation.
A fundamental implication is that binding to E2~ubiquitin favors a HECT domain conformation that is active for E2-to-E3 ubiquitin transfer. This is fueling our efforts to understand factors influencing enzyme conformations and properties that drive ubiquitin through E1-E2-E3 enzyme cascades and onto targets.
These studies are partially supported by ALSAC/St. Jude Children's Research Hospital. Grants from the National Institute of General Medical Sciences provide support for the work on cullin-RING and HECT E3s.
As of April 25, 2016