UBLs (ubiquitin-like proteins) are a major class of eukaryotic post-translational modifiers. UBLs are attached to numerous cellular proteins and to other macromolecules, thereby regulating a wide array of cellular processes. In this chapter we highlight a subset of UBLs and describe their regulatory roles in the cell.
PTM (post-translational modification) of lysine residues on proteins by UBLs [Ub (ubiquitin)-like proteins] is an important mechanism regulating many eukaryotic biological processes in numerous ways . UBLs are small (∼8–20 kDa) proteins composed of a globular β-grasp domain and a flexible C-terminal tail (Figure 1) [2,3]. Regulation of substrates by covalent attachment of UBLs enables specific and complex regulation of cellular processes, in part because UBLs have large chemically diverse surfaces that are themselves subject to a variety of PTMs. For instance, some UBLs become ligated to themselves in diverse ways. The resulting poly-UBL chains signal different outcomes than the monomeric forms of the same UBL, greatly expanding the potential for a UBL to encode specific biological information. Modification of substrates by UBLs typically regulates biological function by either altering the conformation of a target or altering the intermolecular interactions of the target.
Most UBLs are translated as precursor proteins that must be cleaved by a UBL-specific protease to a mature form ending in a C-terminal glycine residue which can be conjugated to substrates. The mature UBL is activated in an ATP-dependent reaction, first forming a UBL-adenylate which allows the subsequent formation of a labile thiolester bond between the catalytic cysteine residue of a UBL-specific activating (E1) enzyme and the UBL C-terminus [3,4]. The activated UBL is then transferred down a cascade of conjugating (E2) and ligase (E3) enzymes that work together to typically attach the C-terminus of the UBL on to a lysine residue on the substrate (Figure 2). For UBLs with the ability to form polymers, the same process can conjugate an incoming UBL on to a preceding UBL . Importantly, UBL modification can be reversed by removal from their targets by specific proteases [5,6].
Evolutionary precursors of UBLs come from prokaryotic biosynthetic pathways [2,3]. MoaD and ThiS are structural homologues of UBLs (Figure 1) that are involved in molybdopterin and thiamin biosynthesis pathways respectively. Like UBLs, MoaD and ThiS are adenylated at their C-termini. Unlike eukaryotic UBLs, MoaD and ThiS do not post-translationally modify proteins or other macromolecules. Instead, they are themselves transiently modified, carrying sulfur atoms at their C-termini, and promoting sulfur transfer in their pathways [2,3].
Ub is translated as either Ub-ribosomal fusion proteins or head-to-tail linear Ub fusion proteins, the latter being essential for stress tolerance . DUBs (deubiquitinating enzymes) cleave the precursors, producing identical 76-residue Ub proteins ending in the sequence glycine–glycine.
Ub can be conjugated to lysine residues or N-terminal amino groups of substrate proteins by cascades of enzymes referred to as E1, E2 and E3, which have been described in detail elsewhere [3,4,8,9]. Mammals contain two Ub-activation enzymes (E1s), which form a thiolester linkage between the E1 catalytic cysteine residue and Gly76 of Ub in a reaction that consumes ATP. Next, Ub is transfered via a transthiolation reaction to the catalytic cysteine residue of one of ∼30 Ub-conjugating (E2) enzymes found in mammals . Finally, mammals contain ∼600 Ub ligase enzymes (E3) which interact with E2 enzymes and substrates to facilitate Ub transfer on to amino groups of substrate lysine residues or N-termini. Mammals have ∼85 DUBs, which can remove Ub from substrates [5,6].
Ub attachment to substrates enables regulation of many cellular processes. Substrates can be monoubiquitinated, wherein a covalent bond is formed between Gly76 of Ub and an amino group from the substrate. Substrates can also be polyubiquitinated by conjugation of Ub to an amino group on a preceeding Ub to form chains. Ub chain formation via one of the seven lysine residues in Ub (Lys6, Lys11, Lys27, Lys29, Lys33, Lys48 and Lys63) or the N-terminus (Met1) generates a diverse array of Ub polymers, which encode distinctive biological information (Figure 3) . Specific Ub chains have different solution properties . For instance, Lys48-linked chains adopt a more closed configuration, whereas Lys63 chains are more open . Because UBL tails are flexible, even the same Ub polymer can display different relative orientations for adjoined Ub globular domains in complexes with different polyubiquitin-binding proteins . Finally, branched Ub chains and free Ub chains have been reported [14–16], although their functions in vivo are not fully understood.
Biological information encoded by the array of Ub modifications is decoded by proteins containing different types and combinations of UBDs (Ub-binding domains). To date ∼20 different UBDs have been identified which recognize Ub in a wide variety of ways . The most common mode of Ub binding is achieved through interaction with a hydrophobic surface around Ile44 of Ub. Domains recognizing this surface include helical bundles, such as the UBA (Ub-association) domain, and single-helix motifs, such as the UIM (Ub-interaction motif). Some UBA domains interacting with high affinity through the Ile44-centred hydrophobic patch can bind both monoubiquitin and polyubiquitin chains . However, many UBA domains bind free Ub with low affinity while recognizing specific polyubiquitin linkages through additional interactions exploiting the unique geometry of specific chains . An example is the UBA domain of the proteasome adaptor protein hHR23a binding two Ile44-centred hydrophobic patches in the context of the closed conformation of Lys48-linked polyubiquitin chains . A similar mechanism of linkage-specific binding by a single domain is illustrated by TAB2 and TAB3 (transforming growth factor β-activated kinase 1 binding proteins 2 and 3) NZF (nuclear protein localization 4 zinc-finger) domains, which bind both Ile44-centred hydrophobic surfaces of adjacent Ub molecules within a Lys63-linked chain [20,21]. Many Ub-binding proteins have combinations of UBDs and interact with specific Ub chains through the correct placement of domains to recognize specific chains . Finally, some proteins directly bind the isopeptide linkage. The DUB AMSH-LP [associated molecule with the SH3 (Src homology 3) domain of signal-transducing adaptor molecule-like protein] surrounds the Lys63 isopeptide linkage within its active site, ensuring it only cleaves Lys63-linked polyubiquitin chains .
Consequences of ubiquitination
Targeting substrates for proteasomal degradation
The best understood function of Ub is its ability to target substrate proteins for degradation by a multisubunit protease known as the 26S proteasome . The proteasome is composed of two subcomplexes: a catalytic CP (core particle) and an RP (regulatory particle) that recruits ubiquitinated proteins, unfolds them and feeds them into the CP by an ATP-dependent mechanism.
The majority of proteasome substrates are targeted either directly or indirectly via polyubiquitin chains and processively proteolysed in the CP. The RP or Ub receptor proteins that associate with the proteasome recognize polyubiquitinated substrates, which are deubiquitinated by DUBs, such as the RP component Rpn11, en route to destruction by the CP. Although Lys48 polyubiquitin chains are the most abundant linkage in vivo and were the first proteasome targeting signal identified, other linkages can target substrates for proteasome degradation [14,24–26]. For instance, Lys11-linked polyubiquitin chains are produced by an E3 enzyme called APC (anaphase-promoting complex), which targets important cell-cycle proteins for proteasomal degradation [27,28].
Regulation of the NF-κB (nuclear factor κB) pathway by ubiquitination
NF-κB signalling cascades illustrate how diverse Ub modifications can regulate a pathway. For instance, ligand binding by TNFRs [TNF (tumour necrosis factor) receptors] results in assembly of signalling complexes that include a variety of E3s, DUBs, Ub-binding proteins and Ub chain types [29,30]. Recognition of specific Ub chains, including Lys63- and Met1-linked chains, by downstream kinase complexes leads to activation of the IKK [IκB (inhibitor of NF-κB) kinase] [8,30–32]. IKK activation leads to Lys48-linked ubiquitination and destruction of the inhibitor IκB and activation of the transcription factor NF-κB to promote cell survival and inflammation .
Diverse functional effects of monoubiquitination
Monoubiquitination of different substrates in different contexts can have different effects by either altering protein–protein interactions or regulating the activity of substrates. For instance, monoubiquitination of histones can modulate chromatin and transcription [33,34]. Monoubiquitination also plays important roles regulating endocytosis and endosomal sorting [35–37]. Finally, monoubiquitination of proteins that interact with Ub or other UBLs can regulate the functions of these proteins .
NEDD8 (neural-precursor-cell-expressed developmentally down-regulated 8)
NEDD8 is the most similar UBL to Ub (human NEDD8 is 58% identical with human Ub) [39,40]. After translation of NEDD8, specific proteases produce a mature form ending in the sequence glycine–glycine. NEDD8 is activated by a specific heterodimeric E1 enzyme (NAE1-UBA3), which does not activate Ub. NEDD8 is transferred from its E1 to one of its two known E2s (UBE2F or UBE2M), before transfer to targets by appropriate E3 enzymes . NEDD8 is primarily conjugated to and regulates the activity of cullin proteins, which are subunits of the largest class of Ub E3 enzymes . Cullin modification by NEDD8 involves a ‘dual E3’ mechanism with a RING (really interesting new gene) E3 in the Rbx family and a Dcn1 E3 working synergistically .
Attachment of NEDD8 activates cullins for ubiquitination of substrates by favouring an apparent conformational change that frees the cullin RING domain and prevents the binding of the cullin inhibitor CAND1 (cullin-associated and NEDDylation-dissociated 1) . Cullins are inactivated by the removal of NEDD8, which is carried out by the CSN5 subunit of the COP9 (constitutive photomorphogenesis 9) signalosome .
The regulation of cullin E3 activity by NEDDylation is crucial for the control of many important proteins in a wide variety of contexts. Dysregulation of cullin activity is linked to many diseases, in particular cancer . NEDD8 may also be conjugated to substrates in addition to cullins, but additional work is needed to clarify the extent and importance of other potential NEDD8 substrates 
SUMO (small ubiquitin-related modifier)
SUMO is a UBL with numerous essential functions in eukaryotes which is present as multiple isoforms in mammals [44–47]. SUMO proteins are ∼17% identical with Ub and have an additional N-terminal ∼20 residues. After translation, SUMO proteins are proteolysed to yield mature SUMO with a C-terminal glycine–glycine motif. SUMO is then activated by a specific heterodimeric E1 enzyme (SAE1-UBA2) and transferred to the only known E2 for SUMO, Ubc9. Unlike Ub, which is not conjugated to lysine residues within specific primary sequence motifs, SUMO is often attached to a lysine residue within SUMO consensus motifs, which are recognized by Ubc9 [44,45]. Although Ubc9 can conjugate SUMO to some substrates directly, several E3 enzymes are known and facilitate substrate SUMOylation [44,45].
In mammals, SUMO2 and SUMO3 are 95% identical with one another and are often referred to as SUMO2/3. SUMO2/3 is ∼50% identical with SUMO1 and possesses a SUMO consensus motif in the N-terminal region which is not present in SUMO1 and enables the formation of poly-SUMO2/3 chains [44,45,48]. SUMO can be removed from its target proteins via the action of SUMO-specific proteases .
SUMO is recognized by SIMs (SUMO-interacting motifs), which are short motifs composed of hydrophobic residues flanked by acidic residues. SIMs interact with SUMO by extending the SUMO β-sheet, such that the hydrophobic residues in the SIM bind a hydrophobic pocket on SUMO [44,45]. SUMO binding by SIMs can be subject to complex regulation, resulting from the presence of multiple SIMs in proteins and from phosphorylation sites in SIMs which regulate their SUMO binding. A number of proteins involved in SUMOylation cascades contain SIMs, some of which display SUMO isoform-specific interactions [44,45]. For example, SIMs have been identified in the UBA2 subunit of the SUMO E1 enzymes, SUMO E3 enzymes, SUMO substrates and in SUMO-targeted Ub ligases .
SUMOylation regulates many essential eukaryotic processes, with many of these being nuclear. For instance, the enzyme TDG (thymine DNA glycosylase) initiates repair of DNA G/T mismatches by binding and removing the thymine base from the DNA backbone, producing a potentially harmful abasic site to which it tightly binds. Release of TDG must be co-ordinated with binding of the next repair enzyme. This process involves modification of TDG with SUMO, which binds a SIM within TDG to produce a conformational change that releases TDG from DNA [47,50].
Atg12 and Atg8
Macroautophagy (autophagy) is a lysosomal degradation pathway in which cytosol and organelles to be degraded are engulfed within a double-membrane structure called an autophagosome. Subsequent fusion of autophagosomes with lysosomes enables lysosomal hydrolases to degrade the sequestered cytosolic materials [51,52]. Autophagy is mediated by >30 autophagy-related (Atg) proteins that were primarily identified by genetic screens in yeast. Among these are two distinct UBLs (Atg12 and Atg8) that are activated by the same E1 enzyme and are both required for autophagosome formation [53,54].
Unlike most UBLs, Atg12 is translated in a mature form with a C-terminal glycine residue ready for activation by the autophagy-specific E1 enzyme Atg7. Atg12 is transferred to Atg5 by the E2 enzyme Atg10. The Atg5–Atg12 complex acts as the E3 enzyme for the second UBL required for autophagy, Atg8 [51,53].
Atg8 (mammals have multiple Atg8 homologues) is unique among UBLs in that it is the only UBL known to be ligated to a lipid rather than to another protein. Atg8 is expressed with a C-terminal extension that is cleaved by Atg4 to expose its C-terminal glycine residue. Atg7 then activates Atg8 for transfer to its E2 enzyme Atg3. An E3 complex containing the aforementioned Atg12–Atg5 conjugate and Atg16 enhances Atg8 ligation to its only known target, the amino group on the lipid PE (phosphatidylethanolamine). Atg8 can be removed from PE by Atg4 [51,53].
One role of Atg8 ligation to PE is to mediate delivery of specific cargo to autophagosomes and lysosomes. Atg8 is recognized by a specific cargo receptor (Atg19) to deliver a specific enzyme, ApeI, a vacuolar hydrolase, to the yeast vacuole . Similar roles have been proposed for higher eukaryotic Atg8 orthologues and other cargo adaptors in diverse processes such as mitophagy (removal of mitochondria), xenophagy (removal of micro-organisms) and the clearance of aggregated proteins from cells . Atg8 has also been proposed to be directly involved in membrane fusion events that lead to autophagosome formation [55,56]. Trafficking machinery, such as Rab proteins, and tethering and fusion complexes also appear to have important roles in membrane fusion . Finally, a distinctive N-terminal pair of α-helices on Atg8 associates with microtubules. Although microtubules are not required for autophagy in yeast, microtubule- and dynein-dependent trafficking of autophagosomes and lysosomes to the microtubule-organizing centre may involve Atg8 homologues and facilitate autophagy in mammalian cells . Further studies will be required to understand the detailed functions of Atg8 in autophagy.
Cross-talk among UBL pathways has important biological implications. For instance, NEDD8 modification activates cullin E3 Ub ligases. Furthermore, more than one UBL can modify the same protein or even the same lysine residue, enabling complex regulation of cellular processes. For example, the DNA polymerase processivity factor PCNA (proliferating-cell nuclear antigen) can be modified on a single lysine residue with either SUMO or Ub: SUMO modification controls recombination events during replication, whereas ubiquitination controls DNA repair [44–47].
Finally, it is important to emphasize that UBL modification of proteins is highly dynamic, with UBLs being conjugated to and deconjugated from protein targets via the co-ordinated action of ligases and proteases, which frequently interact . Interestingly, conjugation and deconjugation functions can be combined in single polypeptide Ub editing proteins. For instance, the enzyme A20 functions within the NF-κB pathway, where it removes Lys63-linked polyubiquitin chains from an important substrate and replaces them with Lys48-linked chains to target the substrate for degradation by the proteasome .
We expect that many examples of biological regulation through co-ordination among Ub and other UBL pathways remain to be discovered and we look forward to exciting developments in the future.
• PTM by UBLs is a major eukaryotic mechanism for regulating diverse cellular processes.
• UBLs are structurally similar, being composed of an N-terminal globular core and a C-terminal tail.
• UBLs modify their substrates by the attachment of their C-terminal tails to lysine residues or the N-termini of proteins.
• UBLs are directed to their targets via a cascade of three enzymes: an E1 (activating enzyme) an E2 (conjugating enzyme) and, in most cases, an E3 (ligase enzyme).
• Most UBLs can be removed from their targets by the action of specific proteases.
• Some UBLs have increased functionality due to their ability to form chains.
• Cross-talk among UBL pathways enables complex regulation of cellular processes.
We apologize to our many colleagues whose work we were unable to cite due to space constraints. We thank M. Calabrese for discussion and editing. Work of the authors is supported by ALSAC (American Lebanese Syrian Associated Charities), a St. Jude Cancer Center grant, the National Institutes of Health [grant number R01GM077053 (to B.A.S.)], and the Howard Hughes Medical Institute. B.A.S. is an Investigator of the Howard Hughes Medical Institute.
- © The Authors Journal compilation © 2012 Biochemical Society