Zinc finger 1 of the RING E3 ligase, RNF125, interacts with the E2 to enhance ubiquitylation.

Inflammation is essential for healthy immune function, wound healing, and resolution of infection. RIG-I is a key RNA sensor that initiates an immune response, with activation and termination of RIG-I signaling reliant on its modification with ubiquitin. The RING E3 ubiquitin ligase, RNF125, has a critical role in the attenuation of RIG-I signaling, yet it is not known how RNF125 promotes ubiquitin transfer or how its activity is regulated. Here we show that the E3 ligase activity of RNF125 relies on the first zinc finger (ZF1) as well as the RING domain. Surprisingly, ZF1 helps recruit the E2, while residues N-terminal to the RING domain appear to activate the E2∼Ub conjugate. These discoveries help explain how RNF125 brings about the termination of RIG-I dependent inflammatory responses, and help account for the contribution of RNF125 to disease. This study also reveals a new role for ZF domains in E3 ligases.

From seeds to trees: how E2 enzymes grow ubiquitin chains.

Modification of proteins by ubiquitin is a highly regulated process that plays a critical role in eukaryotes, from the construction of signalling platforms to the control of cell division. Aberrations in ubiquitin transfer are associated with many diseases, including cancer and neurodegenerative disorders. The ubiquitin machinery generates a rich code on substrate proteins, spanning from single ubiquitin modifications to polyubiquitin chains with diverse linkage types. Central to this process are the E2 enzymes, which often determine the exact nature of the ubiquitin code. The focus of this mini-review is on the molecular details of how E2 enzymes can initiate and grow ubiquitin chains. In particular, recent developments and biochemical breakthroughs that help explain how the degradative E2 enzymes, Ube2s, Ube2k, and Ube2r, generate complex ubiquitin chains with exquisite specificity will be discussed.

Ubiquitin and a charged loop regulate the ubiquitin E3 ligase activity of Ark2C.

A large family of E3 ligases that contain both substrate recruitment and RING domains confer specificity within the ubiquitylation cascade. Regulation of RING E3s depends on modulating their ability to stabilise the RING bound E2~ubiquitin conjugate in the activated (or closed) conformation. Here we report the structure of the Ark2C RING bound to both a regulatory ubiquitin molecule and an activated E2~ubiquitin conjugate. The structure shows that the RING domain and non-covalently bound ubiquitin molecule together make contacts that stabilise the activated conformation of the conjugate, revealing why ubiquitin is a key regulator of Ark2C activity. We also identify a charged loop N-terminal to the RING domain that enhances activity by interacting with both the regulatory ubiquitin and ubiquitin conjugated to the E2. In addition, the structure suggests how Lys48-linked ubiquitin chains might be assembled by Ark2C and UbcH5b. Together this study identifies features common to RING E3s, as well elements that are unique to Ark2C and related E3s, which enhance assembly of ubiquitin chains.

Identification of Ubiquitin Variants That Inhibit the E2 Ubiquitin Conjugating Enzyme, Ube2k.

Transfer of ubiquitin to substrate proteins regulates most processes in eukaryotic cells. E2 enzymes are a central component of the ubiquitin machinery, and generally determine the type of ubiquitin signal generated and thus the ultimate fate of substrate proteins. The E2, Ube2k, specifically builds degradative ubiquitin chains on diverse substrates. Here we have identified protein-based reagents, called ubiquitin variants (UbVs), that bind tightly and specifically to Ube2k. Crystal structures reveal that the UbVs bind to the E2 enzyme at a hydrophobic cleft that is distinct from the active site and previously identified ubiquitin binding sites. We demonstrate that the UbVs are potent inhibitors of Ube2k and block both ubiquitin charging of the E2 enzyme and E3-catalyzed ubiquitin transfer. The binding site of the UbVs suggests they directly clash with the ubiquitin activating enzyme, while potentially disrupting interactions with E3 ligases via allosteric effects. Our data reveal the first protein-based inhibitors of Ube2k and unveil a hydrophobic groove that could be an effective target for inhibiting Ube2k and other E2 enzymes.

The Structure and Ubiquitin Binding Properties of TRAF RING Heterodimers.

Tumour necrosis factor (TNF) receptor associated factor (TRAF) family members share a common domain architecture, but play non-redundant physiological roles in cell signalling. At the N terminus, most TRAFs have a RING domain, followed by a series of Zinc finger (ZF) domains. The RING domain of TRAF6 dimerizes, and the RING homodimer together with the first ZF assembles ubiquitin chains that form a platform which facilitates activation of downstream kinases. The RING dimer interface is conserved amongst TRAF proteins, suggesting that functional heterodimers could be possible. Here we report the structure of the TRAF5-TRAF6 RING heterodimer, which accounts for the stability of the heterodimer as well as its ability to assemble ubiquitin chains. We also show that the RING domain of TRAF6 heterodimerizes with TRAF3 and TRAF2, and demonstrate that the linker helix and first ZF of TRAF2 can cooperate with TRAF6 to promote chain assembly. Collectively our results suggest that TRAF RING homo- and hetero-dimers have the potential to bridge interaction of nearby TRAF trimers and modulate TRAF-mediated signalling.

Isolation and Characterization of Ice-Binding Proteins from Higher Plants.

The characterization of ice-binding proteins (IBPs) from plants can involve many techniques, a few of which are presented here. Chief among these methods are tests for ice recrystallization inhibition, an activity characteristic of plant IBPs. Two related procedures are described, both of which can be used to demonstrate and quantify ice-binding activity. First, is the traditional "splat" assay, which can easily be set up using common laboratory equipment, and second, is our modification of this method using superhydrophobic coated sapphire for analysis of multiple samples in tandem. Thermal hysteresis is described as another method for quantifying ice-binding activity, during which ice crystal morphology observations can be used to provide clues about ice-plane binding. Once ice-binding activity has been evaluated, it is necessary to verify IBP identity. We detail two methods for enriching IBPs from complex mixtures using ice-affinity purification, the "ice-finger" and "ice-shell" methods, and we highlight their advantages and limitations for the isolation of plant IBPs. Recombinant IBP expression, necessary for detailed ice-binding analysis, can present challenges. Here, a strategy for recovery of soluble, active protein is described. Lastly, verification of function in planta borrows from standard protocols, but with an additional screen applicable to IBPs. Together, these methods, and a few considerations critical to success, can be used to assist researchers wishing to isolate and characterize IBPs from plants.

Ubiquitin Variant Inhibitors Meet the Deubiquitinase USP15.

Deubiquitinases (DUBs) are important regulators of cellular function and selective inhibitors are required to reveal their biological role and therapeutic potential. In this issue of Structure, Teyra et al. (2019) report the development of DUB USP15 inhibitors that provide a starting point for the analysis of USP15 function.

A bidentate Polycomb Repressive-Deubiquitinase complex is required for efficient activity on nucleosomes.

Attachment of ubiquitin to lysine 119 of Histone 2A (H2AK119Ub) is an epigenetic mark characteristic of repressed developmental genes, which is removed by the Polycomb Repressive-Deubiquitinase (PR-DUB) complex. Here we report the crystal structure of the Drosophila PR-DUB, revealing that the deubiquitinase Calypso and its activating partner ASX form a 2:2 complex. The bidentate Calypso-ASX complex is generated by dimerisation of two activated Calypso proteins through their coiled-coil regions. Disrupting the Calypso dimer interface does not affect inherent catalytic activity, but inhibits removal of H2AK119Ub as a consequence of impaired recruitment to nucleosomes. Mutating the equivalent surface on the human counterpart, BAP1, also compromises activity on nucleosomes. Together, this suggests that high local concentrations drive assembly of bidentate PR-DUB complexes on chromatin-providing a mechanistic basis for enhanced PR-DUB activity at specific genomic foci, and the impact of distinct classes of PR-DUB mutations in tumorigenesis.

The RING domain of RING Finger 11 (RNF11) protein binds Ubc13 and inhibits formation of polyubiquitin chains.

The Really Interesting New Gene (RING) Finger protein 11 (RNF11) is a subunit of the A20 ubiquitin-editing complex that ensures the transient nature of inflammatory responses. Although the role of RNF11 as a negative regulator of NF-κB signalling is well-documented, the molecular mechanisms that underpin this function are poorly understood. Here, we show that RNF11 binds both Ubc13 and the Ubc13~ubiquitin conjugate tightly and with similar affinity, but has minimal E3 ligase activity. Remarkably, RNF11 appears to bind Ubc13 so tightly that it outcompetes the E1 and an active E3 ligase. As a consequence, RNF11 may regulate the activity of E3s that rely on Ubc13 for ubiquitin chain assembly by limiting the availability of Ubc13 and its conjugate.

Structural insights into K48-linked ubiquitin chain formation by the Pex4p-Pex22p complex.

Pex4p is a peroxisomal E2 involved in ubiquitinating the conserved cysteine residue of the cycling receptor protein Pex5p. Previously, we demonstrated that Pex4p from the yeast Saccharomyces cerevisiae binds directly to the peroxisomal membrane protein Pex22p and that this interaction is vital for receptor ubiquitination. In addition, Pex22p binding allows Pex4p to specifically produce lysine 48 linked ubiquitin chains in vitro through an unknown mechanism. This activity is likely to play a role in targeting peroxisomal proteins for proteasomal degradation. Here we present the crystal structures of Pex4p alone and in complex with Pex22p from the yeast Hansenula polymorpha. Comparison of the two structures demonstrates significant differences to the active site of Pex4p upon Pex22p binding while molecular dynamics simulations suggest that Pex22p binding facilitates active site remodelling of Pex4p through an allosteric mechanism. Taken together, our data provide insights into how Pex22p binding allows Pex4p to build K48-linked Ub chains.

The activity of TRAF RING homo- and heterodimers is regulated by zinc finger 1.

Ubiquitin chains linked through lysine63 (K63) play a critical role in inflammatory signalling. Following ligand engagement of immune receptors, the RING E3 ligase TRAF6 builds K63-linked chains together with the heterodimeric E2 enzyme Ubc13-Uev1A. Dimerisation of the TRAF6 RING domain is essential for the assembly of K63-linked ubiquitin chains. Here, we show that TRAF6 RING dimers form a catalytic complex where one RING interacts with a Ubc13~Ubiquitin conjugate, while the zinc finger 1 (ZF1) domain and linker-helix of the opposing monomer contact ubiquitin. The RING dimer interface is conserved across TRAFs and we also show that TRAF5-TRAF6 heterodimers form. Importantly, TRAF5 can provide ZF1, enabling ubiquitin transfer from a TRAF6-bound Ubc13 conjugate. Our study explains the dependence of activity on TRAF RING dimers, and suggests that both homo- and heterodimers mediated by TRAF RING domains have the capacity to synthesise ubiquitin chains.

Regulation of E2s: A Role for Additional Ubiquitin Binding Sites?

Attachment of ubiquitin to proteins relies on a sophisticated enzyme cascade that is tightly regulated. The machinery of ubiquitylation responds to a range of signals, which remarkably includes ubiquitin itself. Thus, ubiquitin is not only the central player in the ubiquitylation cascade but also a key regulator. The ubiquitin E3 ligases provide specificity to the cascade and often bind the substrate, while the ubiquitin-conjugating enzymes (E2s) have a pivotal role in determining chain linkage and length. Interaction of ubiquitin with the E2 is important for activity, but the weak nature of these contacts has made them hard to identify and study. By reviewing available crystal structures, we identify putative ubiquitin binding sites on E2s, which may enhance E2 processivity and the assembly of chains of a defined linkage. The implications of these new sites are discussed in the context of known E2-ubiquitin interactions.

Structure and Function of the RING Domains of RNF20 and RNF40, Dimeric E3 Ligases that Monoubiquitylate Histone H2B.

Monoubiquitylation of histone H2B is a post-translational mark that plays key roles in regulation of transcription and genome stability. In humans, attachment of ubiquitin to lysine 120 of histone H2B depends on the activity of the E2 ubiquitin-conjugating enzyme, Ube2B, and the really interesting new gene (RING) E3 ligases, RING finger protein (RNF) 20 and RNF40. To better understand the molecular basis of this modification, we have solved the crystal structure of the RNF20 RING domain and show that it is a homodimer that specifically interacts with the Ube2B~Ub conjugate. By mutating residues at the E3-E2 and E3-ubiquitin interfaces, we identify key contacts required for interaction of the RNF20 RING domain with the Ube2B~Ub conjugate. These mutants were used to generate a structure-based model of the RNF20-Ube2B~Ub complex that reveals differences from other RING-E2~Ub complexes, and suggests how the RNF20-Ube2B~Ub complex might interact with its nucleosomal substrate. Additionally, we show that the RING domains of RNF20 and RNF40 can form a stable heterodimer that is active. Together, our studies provide new insights into the mechanisms that regulate RNF20-mediated ubiquitin transfer from Ube2B.

The molecular basis of lysine 48 ubiquitin chain synthesis by Ube2K.

The post-translational modification of proteins by ubiquitin is central to the regulation of eukaryotic cells. Substrate-bound ubiquitin chains linked by lysine 11 and 48 target proteins to the proteasome for degradation and determine protein abundance in cells, while other ubiquitin chain linkages regulate protein interactions. The specificity of chain-linkage type is usually determined by ubiquitin-conjugating enzymes (E2s). The degradative E2, Ube2K, preferentially catalyses formation of Lys48-linked chains, but like most E2s, the molecular basis for chain formation is not well understood. Here we report the crystal structure of a Ube2K~ubiquitin conjugate and demonstrate that even though it is monomeric, Ube2K can synthesize Lys48-linked ubiquitin chains. Using site-directed mutagenesis and modelling, our studies reveal a molecular understanding of the catalytic complex and identify key features required for synthesis of degradative Lys48-linked chains. The position of the acceptor ubiquitin described here is likely conserved in other E2s that catalyse Lys48-linked ubiquitin chain synthesis.

Use of E2~ubiquitin conjugates for the characterization of ubiquitin transfer by RING E3 ligases such as the inhibitor of apoptosis proteins.

Ubiquitylation of proteins is a versatile posttranslational modification that can serve to promote protein degradation, or it can have nondegradative roles, such as mediating protein-protein interactions. The Inhibitor of APoptosis (IAP) proteins are important regulators of pathways that control cell death, proliferation, and differentiation. A number of IAP family members are RING E3 ubiquitin-protein ligases, which promote direct transfer of ubiquitin from charged E2 enzymes, or E2~ubiquitin (E2~Ub) conjugates, to substrate proteins. This results in the attachment of nondegradative ubiquitin signals to other proteins, or the autoubiquitylation and degradation of IAPs. Modulating ubiquitin transfer by IAPs is the focus of a number of drug development initiatives and these studies require a detailed understanding of ubiquitylation. Here, we describe preparation of stable E2~Ub conjugates that can be used in biochemical and biophysical experiments to examine RING domain function. In the last 2 years, the availability of these conjugates has helped unveil a molecular understanding of the process of ubiquitin transfer by IAPs. The approaches described here will be suitable for studying other RING E3 ligases.

Isolation and characterization of ice-binding proteins from higher plants.

The characterization of ice-binding proteins from plants can involve many techniques, only a few of which are presented here. Chief among these methods are tests for ice recrystallization inhibition activity. Two distinct procedures are described; neither is normally used for precise quantitative assays. Thermal hysteresis assays are used for quantitative studies but are also useful for ice crystal morphologies, which are important for the understanding of ice-plane binding. Once the sequence of interest is cloned, recombinant expression, necessary to verify ice-binding protein identity can present challenges, and a strategy for recovery of soluble, active protein is described. Lastly, verification of function in planta borrows from standard protocols, but with an additional screen applicable to ice-binding proteins. Here we have attempted to assist researchers wishing to isolate and characterize ice-binding proteins from plants with a few methods critical to success.

Antifreeze protein from freeze-tolerant grass has a beta-roll fold with an irregularly structured ice-binding site.

The grass Lolium perenne produces an ice-binding protein (LpIBP) that helps this perennial tolerate freezing by inhibiting the recrystallization of ice. Ice-binding proteins (IBPs) are also produced by freeze-avoiding organisms to halt the growth of ice and are better known as antifreeze proteins (AFPs). To examine the structural basis for the different roles of these two IBP types, we have solved the first crystal structure of a plant IBP. The 118-residue LpIBP folds as a novel left-handed beta-roll with eight 14- or 15-residue coils and is stabilized by a small hydrophobic core and two internal Asn ladders. The ice-binding site (IBS) is formed by a flat beta-sheet on one surface of the beta-roll. We show that LpIBP binds to both the basal and primary-prism planes of ice, which is the hallmark of hyperactive AFPs. However, the antifreeze activity of LpIBP is less than 10% of that measured for those hyperactive AFPs with convergently evolved beta-solenoid structures. Whereas these hyperactive AFPs have two rows of aligned Thr residues on their IBS, the equivalent arrays in LpIBP are populated by a mixture of Thr, Ser and Val with several side-chain conformations. Substitution of Ser or Val for Thr on the IBS of a hyperactive AFP reduced its antifreeze activity. LpIBP may have evolved an IBS that has low antifreeze activity to avoid damage from rapid ice growth that occurs when temperatures exceed the capacity of AFPs to block ice growth while retaining the ability to inhibit ice recrystallization.

Ice restructuring inhibition activities in antifreeze proteins with distinct differences in thermal hysteresis.

Antifreeze proteins (AFPs) share two related properties: the ability to depress the freezing temperature below the melting point of ice (thermal hysteresis; TH); and the ability to inhibit the restructuring of ice into larger crystals. Since the 'hyperactive' AFPs, which have been more recently discovered, show an order of magnitude more TH than previously characterized AFPs, we have now determined their activities in ice restructuring inhibition (IrI) assays. IrI activities of three TH-hyperactive AFPs and three less TH-active AFPs varied over an 8-fold range. There was no obvious correlation between high TH activity and high IrI activity. However, the use of mutant AFPs demonstrated that severe disruption of ice-binding residues diminished both TH and IrI similarly, revealing that that the same ice-binding residues are crucial for both activities. In addition, bicarbonate ions, which are known to enhance the TH activity of AFPs, also enhanced their IrI activity. We suggest that these seemingly contradictory observations can be partially explained by differences in the coverage of ice by TH-hyperactive and non-hyperactive AFPs, and by differences in the stability of AFP-bound ice under supercooled and recrystallization conditions.

Compound ice-binding site of an antifreeze protein revealed by mutagenesis and fluorescent tagging.

By binding to the surface of ice crystals, type III antifreeze protein (AFP) can depress the freezing point of fish blood to below that of freezing seawater. This 7-kDa globular protein is encoded by a multigene family that produces two major isoforms, SP and QAE, which are 55% identical. Disruptive mutations on the ice-binding site of type III AFP lower antifreeze activity but can also change ice crystal morphology. By attaching green fluorescent protein to different mutants and isoforms and by examining the binding of these fusion proteins to single-crystal ice hemispheres, we show that type III AFP has a compound ice-binding site. There are two adjacent, flat, ice-binding surfaces at 150° to each other. One binds the primary prism plane of ice; the other, a pyramidal plane. Steric mutations on the latter surface cause elongation of the ice crystal as primary prism plane binding becomes dominant. SP isoforms naturally have a greatly reduced ability to bind the prism planes of ice. Mutations that make the SP isoforms more QAE-like slow down the rate of ice growth. On the basis of these observations we postulate that other types of AFP also have compound ice-binding sites that enable them to bind to multiple planes of ice.

Identification of the ice-binding face of a plant antifreeze protein.

The antifreeze protein of Lolium perenne, a perennial ryegrass, was previously modeled as a beta-roll with two extensive flat beta-sheets on opposite sides of the molecule. Here we have validated the model with a series of nine site-directed steric mutations in which outward-pointing short side-chain residues were replaced by tyrosine. None of these disrupted the fold. Mutations on one of the beta-sheets and on the sides of the protein retained 70% or greater activity. Three mutations that clustered on the other flat surface lost up to 90% of their antifreeze activity and identify this beta-sheet as the ice-binding face.