elegans *

Ubiquitin is a highly conserved 76 amino acid polypeptide, which is covalently attached to target proteins to signal their degradation by the 26S proteasome or to modify their function or localization. Regulated protein degradation, which is associated with many dynamic cellular processes, occurs predominantly via the ubiquitin-proteasome system. Ubiquitin is conjugated to target proteins through the sequential actions of a ubiquitin-activating enzyme, ubiquitin-conjugating enzymes, and ubiquitin-protein ligases. The nematode Caenorhabditis elegans has one ubiquitin-activating enzyme, twenty putative ubiquitin-conjugating enzymes, and potentially hundreds of ubiquitin-protein ligases. Research in C. elegans has focused on the cellular functions of ubiquitin pathway components in the context of organismal development. A combination of forward genetics, reverse genetics, and genome-wide RNAi screens has provided information on the loss-of-function phenotypes for the majority of C. elegans ubiquitin pathway components. Additionally, detailed analysis of several classes of ubiquitin-protein ligases has led to the identification of their substrates and the molecular pathways that they regulate. This review presents a comprehensive overview of ubiquitin-mediated pathways in C. elegans with a description of the known components and their identified molecular, cellular, and developmental functions.


Overview of ubiquitin conjugation
Ubiquitin (Ub) is a ubiquitously expressed and highly conserved 76 amino acid polypeptide (Hershko and Ciechanover, 1998; Figure 1A).The covalent tandem attachment of multiple Ub to a target protein to form poly-ubiquitin chains can mark the protein for degradation by the 26S proteasome.Ub is covalently attached to substrate proteins by the concerted actions of three classes of enzymes (Hershko and Ciechanover, 1998).A ubiquitin-activating enzyme (E1) uses one ATP molecule to bind Ub via a thiolester linkage.The activated Ub is transferred to a ubiquitin-conjugating enzyme (E2), also via a thiolester linkage.The E2 is brought to the substrate by binding a ubiquitin-protein ligase (E3) that binds both the E2 and the substrate.Once bound to an E3, the E2 either directly transfers Ub to the substrate or transfers the Ub through a thiolester linkage to the E3, which then transfers the Ub to the substrate.Multiple rounds of E2 interactions with substrate-bound E3 are required to produce a poly-Ub chain on the substrate.In a few cases, the E2-E3 combination is not capable of adding more than a few Ub, and in this situation a ubiquitin chain assembly factor (E4) is required for the conjugation of additional Ub to form a poly-Ub chain (Koegl et al., 1999).(S.c.).Differences between Ub residues are boxed.Note that C. elegans and humans only differ at one position relative to each other and at three positions relative to budding yeast.The locations of the seven lysines in Ub are marked with the residue numbers provided above the alignment.B) Translation of the C. elegans ubq-1 polyubiquitin gene.The amino acid sequence is presented at 76 amino acids per line so that individual ubiquitin repeats are aligned.Note that repeat 6 has a different amino acid at position 9 of the repeat (yellow box, isoleucine rather than threonine).There are two amino acids after the final ubiquitin repeat.
Ub is conjugated to target proteins or other Ub through a bond between the conserved C-terminus of Ub and the ε-amino group of a lysine residue on the target protein or other Ub (Hershko and Ciechanover, 1998).A minimum of four tandemly-attached Ub are required to allow recognition of the target protein by the 26S proteasome, presumably because a tetramer of poly-Ub assumes a higher order structure that is required for recognition (Pickart, 2000).Ub has seven lysine (Lys) residues (Figure 1A) and Ub can be conjugated to several of these Lys residues.Poly-Ub chains created by conjugation through a Lys-48 linkage targets a substrate for degradation by the 26S proteasome.In contrast, poly-Ub formed by Lys-63 conjugation does not lead to proteasome-mediated degradation, but instead is associated with the regulation of endocytosis or changes in target protein function (Schnell and Hicke, 2003).Similarly, conjugation of a single Ub (mono-ubiquitination) or less than four Ub can affect protein activity, including aspects of transcriptional regulation, protein trafficking, and endocytosis (Schnell and Hicke, 2003).The functions of poly-Ub chains created with Lys-11 and Lys-29 linkages have not been determined (Aguilar and Wendland, 2003).
Ubiquitin-mediated proteolysis is the most important pathway for the degradation of nuclear and cytosolic proteins.Inactivation of the Ub proteolytic pathway inhibits the degradation of the majority of cellular proteins, regardless of whether the proteins have short or long half-lives (Rock et al., 1994).Given the central importance of ubiquitin-mediated protein degradation in a range of cellular processes, it is not surprising that Ub-mediated pathways are important for multiple aspects of C. elegans development and cellular physiology.

26S Proteasome
The 26S proteasome is a conserved chambered protease complex that is present in both the cytoplasm and the nucleus (Wojcik and DeMartino, 2003).It consists of a 20S proteasome, a central core containing proteolytic subunits, and two 19S regulatory complexes that bind to ubiquitinated substrates, cleave off ubiquitin, and then unfold and translocate the substrate into the 20S core (Pickart and Cohen, 2004).C. elegans has 14 conserved subunits that comprise the 20S core, as well as 18 conserved 19S components (Davy et al., 2001).RNAi depletion of proteasome components during larval stages produces larval arrest and lethality while RNAi depletion in adult hermaphrodites produces progeny that arrest at the one-cell stage with defective meiosis I, indicating the central importance of this pathway (Takahashi et al., 2002;Gonczy et al., 2000).

Ubiquitin
There are two loci for ubiquitin (Ub) in C. elegans, ubq-1 and ubq-2.ubq-1 is a polyubiquitin locus (Graham et al., 1989).The predominant splice form of ubq-1 encodes an 838 amino acid peptide that includes 11 tandem Ub sequences (Figure 1B).The polyubiquitin structure of the locus is common to other eukaryotic species (Schlesinger and Bond, 1987).The polyubiquitin protein is post-translationally cleaved into individual Ub peptides by ubiquitin C-terminal hydrolases (Johnston et al., 1999).The Ub peptides of ubq-1 are identical with the exception of repeat 6, which substitutes an isoleucine for a highly conserved threonine at position 9 of the repeat (Figure 1B).The functional significance of this altered Ub peptide is not known.In C. briggsae, this atypical Ub repeat is not present, instead the orthologous polyubiquitin locus comprises ten Ub repeats that are all identical to the predominant C. elegans Ub sequence.
The second Ub locus, ubq-2, includes an intact canonical Ub fused to the L40 ribosomal large subunit protein (Jones and Candido, 1993).This fusion gene is broadly conserved in eukaryotes (Schlesinger and Bond, 1987).The ubq-2 locus contains the only copy of the L40 ribosomal subunit in the C. elegans genome.In yeast, the hybrid protein is rapidly cleaved to form Ub and the L40 ribosomal subunit (Finley et al., 1989).The transient presence of Ub in the fusion protein promotes the incorporation of the L40 subunit into ribosomes (Finley et al., 1989).Once cleaved, the Ub is functional for covalent attachment to proteins (Ozkaynak et al., 1987).
RNAi of ubq-1 or ubq-2 produces a one-cell stage arrest during the meiotic divisions, similar to inactivation of the proteasome (Gonczy et al., 2000;Piano et al., 2000).The relative importance of ubq-1 vs ubq-2 is not known, as RNAi is expected to inactivate both genes due to their extensive homology (Tijsterman et al., 2002).

Ubiquitin-activating enzyme (E1)
As in other eukaryotes, there is only a single ubiquitin-activating enzyme in C. elegans, UBA-1.Disruption of UBA-1 activity would be expected to completely inactivate the Ub proteolytic pathway.However, while RNAi of uba-1 produces an embryonic arrest, it is not as penetrant as RNAi for ubq-1 or particular proteasome components, perhaps because, as an enzyme, it is more resistant to effects of depletion (Maeda et al., 2001;Kamath et al., 2003;Simmer et al., 2003;Piano et al., 2000;Gonczy et al., 2000).

Ubiquitin-conjugating enzymes (E2s)
There are 22 proteins with homology to ubiquitin-conjugating enzymes (UBCs) in C. elegans, with an additional three ubiquitin E2 variants (UEVs) that lack the critical cysteine residue in the catalytic site (Jones et al., 2002).The C. elegans UBCs are numbered ubc- 1-3, 6-9, and 12-26; with numbers 4, 5, 10, and 11 skipped.The numbering of UBCs in C. elegans does not match that of S. cerevisiae or humans, and orthologous groupings are presented in Table 1.Note that ubc-9 and ubc-12 designate conjugating enzymes for the Ub-like proteins SUMO (SMO-1) and Nedd8 (NED-8), respectively, and do not conjugate Ub (Jones and Candido, 2000;Jones et al., 2002).The functions of the E2 genes have been studied in systematic RNAi screens.Only two of the 20 E2 enzymes that are specific for Ub-conjugation are essential for embryonic viability, ubc-2/let-70 and ubc-14 (Table 1; Jones et al., 2002).This is surprising given the relatively large number of E3 genes that are associated with embryonic lethal phenotypes (see below).This suggests either that UBC-2 and UBC-14 are the only E2s that function with these essential E3s or that there is significant redundancy of E2 function.In general, very little is known about which C. elegans E2s function with particular E3s.Of the remaining Ub-specific E2 genes, the RNAi depletion of four are associated with post-embryonic phenotypes: ubc-19 RNAi produces unhealthy larvae; ubc-20 RNAi produces an impenetrant L3 or L4 larval arrest; ubc-25 RNAi produces defects in neuromuscular function; and ubc-18 RNAi produces animals that have slightly slower growth and reduced brood sizes but otherwise appear wild-type (Maeda et al., 2001;Jones et al., 2002;Schulze et al., 2003;Fay et al., 2003; Table 1).Interestingly, ubc-18 functions redundantly with lin-35 Rb to promote normal pharyngeal morphogenesis, and the simultaneous inactivation of both genes causes synthetic embryonic lethality (Fay et al., 2003).For the remaining 14 E2s, inhibition by RNAi was not associated with any apparent defects.et al., 2002;Schulze et al., 2003), with updates of other species homolog names.ubc-26 (Y110A2AM.3),was named in this study.The peptide predicted to be conjugated (Ub or Ub-like) is derived from information of homologs in other species when not known in C. elegans.Only the more severe phenotypes are listed.Phenotypes derived from RNAi or mutant analysis are denoted (brackets).WT = wild-type phenotype.References for the phenotypes listed are given.

Ubiquitin-protein ligases (E3s)
There are four major classes of ubiquitin ligases: HECT-domain proteins; U-box proteins; monomeric RING finger proteins; and multisubunit complexes that contain a RING finger protein (Passmore and Barford, 2004).HECT-domain E3s are unique in that Ub is transferred to a conserved cysteine residue of the E3 in a thiolester linkage, and then the E3 transfers the Ub to the substrate (Passmore and Barford, 2004; Figure 2).This function as a covalent intermediary in the transfer of Ub is not found in other classes of E3 proteins.The RING finger motif (Really Interesting New Gene) comprises eight cysteine or histidine residues that bind two Zn 2+ ions in a cross-brace structure (Fang et al., 2003).The U-box is structurally similar to the RING finger motif, but it does not bind Zn 2+ , instead, hydrogen bonds take the place of the Zn 2+ in the structure (Fang et al., 2003).Monomeric RING finger E3s and U-box E3s bind to both the substrate and the E2 enzyme (Figure 2).In multimeric RING finger complexes, the RING finger protein binds the E2 while other proteins in the complex bind the substrate.These multimeric complexes fall into two classes: cullin-based complexes, and the APC/C (anaphase promoting complex/cyclosome), which contains the cullin-like protein APC2 (Vodermaier, 2004).The N-terminus of CUL-3 binds directly to the SRS, which utilizes a BTB/POZ domain to bind to CUL-3.

HECT-domain E3s
There are nine genes in C. elegans that encode proteins with a HECT domain (Table 2).Of these genes, only two have been studied in detail: oxi-1 and wwp-1.oxi-1 was cloned as a gene whose expression increases under oxidative stress (growth in high oxygen concentrations; Yanase and Ishi, 1999).There is no observed RNAi phenotype of oxi-1; however, the RNAi analysis was not performed under high oxygen conditions when oxi-1 would be assumed to be active (Table 2).wwp-1 encodes a conserved protein with HECT and WW domains, and its RNAi depletion causes embryonic lethality with defective morphogenesis after the comma stage (Huang et al., 2000).Large-scale RNAi screens revealed that RNAi depletion of the HECT-domain gene D2085.4produces sterility in the P0 (Maeda et al., 2001).The remaining six HECT-domain genes were not associated with reproducible RNAi phenotypes (Table 2) Only the more severe phenotypes are listed.Phenotypes derived from RNAi or mutant analysis are denoted (brackets).WT = wild-type phenotype.References for the phenotype listed are given.

U-box E3s & E4s
There are four genes in C. elegans that encode proteins with a U-box domain (Table 3).The C. elegans U-box protein CHN-1 is the homolog of mammalian CHIP.CHIP binds to the chaperones Hsp70 and Hsp90 and functions as an E3 to degrade misfolded proteins (Hatakeyama and Nakayama, 2003).C. elegans CHN-1 also binds to the Hsp70 homolog HSP-1, suggesting a similar function (Hoppe et al., 2004).Animals homozygous for a null allele of chn-1 have slightly lower brood sizes at 20°C, but otherwise appear normal.Consistent with a cellular role with heat shock proteins, chn-1 homozygotes are sensitive to heat-stress, exhibiting larval arrest and lethality at higher temperatures (Hoppe et al., 2004; Table 3).C. elegans CHN-1 physically interacts with a second U-box protein, UFD-2 (Hoppe et al., 2004).UFD-2 is the ortholog of budding yeast Ufd2, which functions as an E4 (Koegl et al., 1999).As described above, an E4 enzyme catalyzes the elongation of ubiquitin chains on proteins that already have one or a few conjugated Ub (Koegl et al., 1999).Hoppe et al., found that both C. elegans CHN-1 and UFD-2 can function independently of each other as E3s for the addition of one to three Ub to UNC-45, a myosin-directed chaperone (Hoppe et al., 2004).However, more extensive poly-ubiquitination of UNC-45 in vitro required both CHN-1 and UFD-2, suggesting a novel mechanism in which a combination of E3s can produce E4 activity (Hoppe et al., 2004).The CHN-1-UFD-2 complex was able to function in vitro with UBC-2/LET-70, suggesting that UBC-2/LET-70 is the in vivo E2 (Hoppe et al., 2004).
The third U-box gene in C. elegans is cyp-4/mog-6, which is a homolog of human cyclophilin-60 (hCyp60/CYC4).In humans, hCyp60 has both peptidyl-prolyl cis/trans isomerase activity associated with its C-terminus and E3 activity associated with the U-box in its N-terminus (Hatakeyama and Nakayama, 2003).C. elegans CYP-4 also has both domains and exhibits protein-folding activity indicative of a functional prolyl isomerase (Page et al., 1996).Loss of cyp-4/mog-6 results in a failure of the hermaphrodite germ line to switch from producing sperm to producing oocytes, so that only sperm are produced (Graham et al., 1993).The prolyl isomerase domain of CYP-4/MOG-6 is not required for the sperm/oocyte switch, while the N-terminus, containing the U-box, is required (Belfiore et al., 2004).CYP-4/MOG-6 is also required for embryogenesis (Graham et al., 1993).
The final U-box gene is T10F2.4,which is homologous to yeast and human PRP19.Budding yeast Prp19 functions in spliceosome assembly, and human PRP19 has been shown to possess E3 activity (Blanton et al., 1992;Hatakeyama and Nakayama, 2003).Large-scale RNAi screens revealed that T10F2.4 is required for embryonic viability (Table 3).

Monomeric RING finger proteins
There are 152 RING finger proteins in the C. elegans genome (Table 4).While a majority of RING finger proteins tested in vitro exhibit E3 activity, it is unclear if all RING finger proteins function as E3s in vivo (Fang et al., 2003).There are two classes of RING finger motifs, H2 and HC, based on the placement of His or Cys residues in positions 4 and 5 of the motif (Fang et al., 2003).The three RING finger proteins that are known to be integral components of multisubunit complexes (RBX-1, RBX-2, and APC-11) are of the H2 class and are very small proteins of 110-135 amino acids.A recent survey of RING finger proteins in C. elegans found more to be of the HC class (90 genes) than the H2 class (13 genes; Moore and Boyd, 2004).The majority of RING finger genes of either class encode proteins that are much larger than the multisubunit E3 RING-H2 proteins, as would be expected for proteins that function as monomeric E3s that bind to both the E2 and the substrate.

Gene name Phenotype References
Only the more severe phenotypes are listed.Phenotypes derived from RNAi or mutant analysis are denoted (brackets).Phenotype abbreviations are defined in WormBase.References for the phenotype listed are given.
Twenty one of the RING finger proteins in C. elegans that do not function as components of known multisubunit E3 complexes have been the focus of genetic studies.(see named genes in Table 4).Most of these have not been studied for possible E3 activity.Five of the potentially monomeric RING finger proteins have been implicated as E3s, and are discussed below.RNF-5 is the ortholog of the mammalian E3 RNF5 (Didier et al., 2003).RNAi depletion of rnf-5 causes a disorganization of body wall muscle dense bodies (Broday et al., 2004).RNF-5 binds and negatively regulates the protein level of UNC-95, a LIM-domain protein that is required for the integrity of dense bodies (Didier et al., 2003;Broday et al., 2004).Mutations of the RING finger domain of RNF-5 severely reduce its ability to lower UNC-95 protein levels upon overexpression (Broday et al., 2004).These results suggest that RNF-5 directly targets UNC-95 for ubiquitin-mediated proteolysis.SLI-1 is the ortholog of mammalian c-Cbl.c-Cbl functions as an E3 to ubiquitinate active receptor tyrosine kinases (RTKs) to induce their endocytosis or degradation (Shtiegman and Yarden, 2003).SLI-1 negatively regulates the LET-23 RTK, which mediates vulva differentiation (Yoon et al., 1995).A sli-1 mutant that lacks the RING finger domain has a significantly reduced ability to inhibit LET-23 activity, suggesting that SLI-1 functions as an E3 to regulate LET-23 (Yoon et al., 2000).
RFP-1 was identified as a binding partner for the E2 UBC-1 in a yeast two-hybrid screen (Crowe and Candido, 2004).RNAi depletion of rfp-1 produces an L1-stage larval arrest, while escapers exhibit vulval and egg-laying defects (Crowe and Candido, 2004).Interestingly, RNAi depletion of ubc-1, the predicted E2, does not produce any phenotypes, suggesting that RFP-1 can function with additional E2s (Crowe and Candido, 2004).
BRC-1 and BRD-1 are RING finger proteins that are orthologs of the mammalian breast cancer susceptibility gene BRCA1 and the BRCA1-associating protein BRD-1, respectively (Boulton et al., 2004).In mammals, a complex of BRCA1 and BRD1 exhibits E3 activity in vitro, and mutations in the RING finger of BRCA1 that abolish E3 activity are associated with breast cancer (Ohta and Fukuda, 2004).In C. elegans, BRC-1 also binds BRD-1 (Boulton et al., 2004).RNAi depletion of either gene produces a high incidence of males (Him) phenotype, elevated levels of p53-dependent germ cell death, and chromosome fragmentation after irradiation, suggesting a role in DNA repair (Boulton et al., 2004).Intriguingly, mammalian BRCA1 and BRD1 have been implicated in the ubiquitination of p53 (Dong et al., 2003), suggesting that the p53-dependent germ cell death observed upon BRC-1 or BRD-1 RNAi results from elevated p53 levels.In concordance with this, loss of p53 (cep-1) suppressed the brc-1 and brd-1 germ cell deaths (Boulton et al., 2004).
RPM-1 encodes a protein with RING finger and guanine nucleotide exchange domains.rpm-1 mutants have varied defects in neuron branching, and synapse organization and structure, indicating a critical role in presynaptic development (Schaefer et al., 2000;Zhen et al., 2000).RPM-1 controls presynaptic development by negatively regulating DLK-1, which functions as the initial MAP kinase (MAPK) in a MAPK cascade (Nakata et al., 2005).Co-expression of RPM-1 and DLK-1 in mammalian cells promotes the ubiquitination of DLK-1, and DLK-1 is co-immunoprecipitated by the C-terminus of RPM-1, suggesting that RPM-1 directly binds and mediates DLK-1 ubiquitination (Nakata et al., 2005).These results identified the first ubiquitinated substrate of an RPM-1 family member, which also control synapse formation in Drosophila and mammals (Chang and Balice-Gordon, 2000;Burgess et al., 2004).RPM-1 has also been shown to physically associate with an SCF E3 complex (Liao et al., 2004), and this will be discussed in the CUL-1 section of the review.

Multisubunit RING finger complexes: cullin-based E3s
Cullins are a conserved family of E3 components that were first identified in C. elegans and budding yeast (Kipreos et al., 1996;Mathias et al., 1996).There are six cullins in C. elegans.The crystal structure of a human CUL1-based SCF complex reveals that CUL1 forms a rigid scaffold in which the C-terminus binds the RING-H2 finger protein Rbx1/Roc1, and the N-terminus binds to the adaptor Skp1 (Zheng et al., 2002; Figure 2).Skp1 binds the substrate-recognition subunit (SRS), through an F-box motif of the SRS.The SRS binds substrates and positions them for ubiquitination.The E2 binds the complex through interaction with Rbx1 (Figure 2).Other cullins are also predicted to function as scaffolds with similar tertiary structure (Wu et al., 2003).All cullin-based E3s components interact with multiple SRSs, each of which form distinct E3 complexes that direct the ubiquitination of distinct sets of substrates (Guardavaccaro and Pagano, 2004).Based on the large number of potential SRS genes in metazoan genomes, cullin-based E3s may comprise the most abundant class of E3 in metazoa.
Each class of cullin-based E3 complexes includes a small RING finger protein, either Rbx1/Roc1 or Rbx2/Roc2 (Ohta et al., 1999).C. elegans RBX-1 and RBX-2 share 36% identity.Partial inactivation of the C. elegans rbx-1 gene produces phenotypes that resemble a mixture of different cullin loss-of-function phenotypes (Sasagawa et al., 2003;Moore and Boyd, 2004;data not shown).Complete inactivation of rbx-1 causes a one-cell stage arrest that is more severe than any individual cullin loss-of-function phenotype, and may reflect the simultaneous loss of multiple cullin complexes (Sasagawa et al., 2003;Moore and Boyd, 2004).In contrast, inactivation of rbx-2 by RNAi does not cause any apparent defects (Moore and Boyd, 2004), suggesting a more limited role.
Cullin-based complexes are themselves regulated by the covalent addition of a ubiquitin-like protein, Nedd8, onto the cullin.Nedd8 promotes cullin activity, at least in part, by blocking the binding of the cullin inhibitor CAND1 (Pan et al., 2004).NED-8 is removed from cullins by the COP9 Signalosome complex (CSN; Cope and Deshaies, 2003).In C. elegans, inactivation of either the Nedd8 homolog NED-8 or CSN components produce embryonic phenotypes similar to cul-3 RNAi, suggesting that both neddylation and deneddylation are required for CUL-3 activity (Pintard et al., 2003).In contrast, the early embryonic phenotypes observed in cul-2 mutants are not seen upon inactivation of NED-8 or CSN, suggesting that NED-8 modification is not essential for CUL-2 function (Pintard et al., 2003).The functional requirement for NED-8 conjugation of other cullins has not been studied.

CUL-1-based complexes
CUL-1-based E3 complexes are known as SCF complexes to reflect the components: Skp1 (the adaptor); CUL1/Cdc53; and an F-box protein (the substrate recognition subunit; Figure 2).Distinct SCF complexes are formed by the combination of core components (Skp1, CUL1, and Rbx1) with different F-box proteins.C. elegans has an extremely large number of F-box proteins relative to other metazoa: at least 326 in C. elegans compared to 13 in budding yeast and 68 in humans (Kipreos and Pagano, 2000;Jin et al., 2004).This larger number of F-box protein genes is matched by a larger number of Skp1-related genes in C. elegans (21 SKR genes) relative to only a single Skp1 gene in budding yeast and humans (Nayak et al., 2002;Yamanaka et al., 2002).Seven of the SKRs were found to interact with CUL-1: SKR-1, -2, -3, -7, -8, -9, and -10 (Nayak et al., 2002;Yamanaka et al., 2002).The observation of multiple SKRs that can bind CUL-1, along with the large number of F-box proteins, suggests that C. elegans will contain many distinct SCF complexes.The cellular functions associated with known SCF complexes are discussed below.
SCF LIN-23 : LIN-23 is an F-box protein with WD-repeats.lin-23 mutants have hyperplasia in all somatic lineages caused by a failure of dividing blast cells to cease cell division at the appropriate time (Kipreos et al., 2000).This cell cycle exit defect is also the primary phenotype associated with cul-1 mutants and with skr-1 and skr-2 RNAi, suggesting that the closely related SKR-1 and SKR-2 proteins function as adaptors in the SCF LIN-23  complex (Kipreos et al., 1996; Nayak et al., 2002).LIN-23 has a separate function to promote the proper outgrowth of axons (Mehta et al., 2004).
SCF SEL-10 : SEL-10 is an F-box protein with WD-repeats that negatively regulates signaling by the Notch family member LIN-12 (Hubbard et al., 1997).SEL-10 was shown to physically bind the LIN-12 receptor (a Notch family member), suggesting that it directly mediates LIN-12 ubiquitination (Hubbard et al., 1997).This pathway was subsequently shown to be conserved in mammals, with the SEL-10 ortholog responsible for the ubiquitin-mediated degradation of Notch (Wu et al., 2001;Oberg et al., 2001).SEL-10 also binds and negatively regulates the presenilin SEL-12 (Wu et al., 1998).Studies with human cells have subsequently shown that the degradation of presenilin by SCF SEL-10 is conserved (Li et al., 2002).Finally, SEL-10 binds and facilitates the degradation of FEM-1 and FEM-3, which function in the sex determination pathway to promote male development (Jager et al., 2004).
SCF FSN-1 : FSN-1 is an F-box protein with a SPRY domain that is required for regulating neuromuscular junction (synapse) formation in different classes of neurons (Liao et al., 2004).Co-immunoprecipitation (co-IP) analysis indicates that FSN-1 physically associates with CUL-1 and SKR-1, suggesting that it functions as the SRS for an SCF complex (Liao et al., 2004).FSN-1 binds and negatively regulates the level of SCD-1/ALK (Liao et al., Ubiquitin-mediated pathways in C. elegans SCF FSN-1 complex for regulating synapse formation (Liao et al., 2004).
FSN-1 has been shown by co-IP analysis to physically associate with the large RING finger protein RPM-1, and it has been proposed that RPM-1 functions as the RING finger component of the SCF FSN-1 complex (Liao et al.,  2004).However, there is evidence that suggests that RPM-1 does not function analogously to RBX-1 in an SCF FSN-1 complex.In particular, rpm-1 null mutants are not functionally equivalent to fsn-1 null mutants, and RPM-1 requires the presence of FSN-1 to associate with SCF components, which would not be expected for a component that binds directly to CUL-1 (Liao et al., 2004).

CUL-2-based complexes
CUL-2-based E3 complexes have a structure very similar to that of SCF complexes (Wu et al., 2003; Figure 2).CUL-2 complexes employ the Skp1-related protein elongin C as an adaptor in combination with elongin B, which contains a ubiquitin-like domain (Kim and Kaelin, 2003).Substrate recognition subunits bind elongin C through a BC-box/VHL-box motif (Kamura et al., 2004).The C. elegans cul-2 gene is the only metazoan cul-2 ortholog whose functions have been analyzed genetically.cul-2 mutants have a large number of phenotypes reflecting diverse cellular functions.1) CUL-2 is required for the G1-to-S phase transition in germ cells, which is caused, at least in part, by a failure to negatively regulate the levels of the CDK-inhibitor CKI-1 (Feng et al., 1999).
2) CUL-2 is required for the meiosos II metaphase to anaphase transition and meiosis II exit.The failure/delay of the metaphase II to anaphase II transition is correlated with a failure to degrade the cell cycle regulator cyclin B1, while the delay in meiosis exit is correlated with a failure to degrade cyclin B3 (Liu et al., 2004;Sonneville and Gonczy, 2004).3) CUL-2 is required for proper anterior-posterior (A-P) polarity.In cul-2 mutants, A-P polarity is often reversed due to the perduring meiotic spindle acting as a catalyst for the ectopic placement of the PAR-2 polarity protein onto the anterior cortex (Liu et al., 2004;Sonneville and Gonczy, 2004).Additionally, CUL-2 restricts the localization of PAR-2 in regions distant from microtubule-organizing centers (Liu et al., 2004;Sonneville and Gonczy, 2004).4) CUL-2 is required for defects in mitotic chromosome condensation and mitotic progression (Feng et al., 1999).5) CUL-2 is required to prevent cytoplasmic extensions/blebbing in the early embryo (Feng et al., 1999).6) A CUL-2 complex containing the SRS ZIF-1 is required for the degradation of five CCCH Zn finger polarity proteins (PIE-1, POS-1, MEX-1, MEX-5, and MEX-6) in non-germ cell embryonic lineages (DeRenzo et al., 2003).
In mammals, CUL-2 functions in a complex with the von Hippel-Lindau tumor suppressor protein (VHL) as the SRS to target the degradation of hypoxia inducible factor-1α (HIF-1α; Kim and Kaelin, 2003).In C. elegans, VHL-1 also promotes HIF-1 degradation, so it is likely that C. elegans VHL-1 functions as a component of a CUL-2 complex (Epstein et al., 2001).

CUL-3-based complexes
CUL-3 has a slightly different structure from that of CUL-1 and CUL-2-based complexes in that a single BTB/POZ-domain protein functions as both the substrate recognition subunit and adaptor, i.e., the BTB protein binds CUL-3 and the substrate (van den Heuvel, 2004; Figure 2).C. elegans contains over 100 BTB-domain proteins, indicating the possibility for multiple CUL-3-based complexes (Furukawa et al., 2003;Xu et al., 2003).C. elegans was the first organism in which a functional CUL-3-based E3 complex and its substrate were identified.A C. elegans CUL-3 complex that contains the BTB protein MEL-26 was shown to degrade the microtubule-severing katanin MEI-1 (Pintard et al., 2003;Furukawa et al., 2003;Xu et al., 2003).MEI-1 is degraded in the one-cell embryo after meiosis (Pintard et al., 2003).Presumably, the microtubule-severing activity of MEI-1 is required during meiosis to restrict the size of the meiotic spindle, but must be destroyed to allow the larger mitotic spindle to form after meiosis.In cul-3 RNAi animals or mel-26 mutants, mitotic aster microtubules are disorganized and shorter compared to wild type, and this is associated with defects in spindle positioning and elongation, and cytokinesis (Kurz et al., 2002;Pintard et al., 2003;Dow and Mains, 1998).

CUL-4-based complexes
The structure of CUL-4-based E3 complexes has not been fully worked out.In mammals, the CUL-4 complex includes the DDB1 protein, which appears to be capable of functioning either as an adaptor or as a substrate recognition subunit (Wertz et al., 2004;Hu et al., 2004).In C. elegans, CUL-4 has a central role in the regulation of DNA replication by restricting replication to only once per cell cycle (Zhong et al., 2003).RNAi depletion of cul-4 produces an L2-stage larval arrest in which blast cells undergo unrestrained re-replication and attain elevated DNA contents up to 100 C (Zhong et al., 2003).CUL-4 is required for the degradation of the replication licensing factor CDT-1 during S phase (Zhong et al., 2003).The degradation of CDT-1 precludes it from reloading the MCM complex onto replication origins, thereby preventing the re-initiation of DNA replication at origins during the same cell cycle (Zhong et al., 2003;Feng and Kipreos, 2003).The degradation of CDT-1 by CUL-4 was subsequently shown to be conserved in Drosophila and mammals (Higa et al., 2003).

Multisubunit RING finger complexes: APC/C
The APC/C is a conserved multisubunit E3 complex that functions during meiosis, mitosis, and G1 phase (Yeong, 2004).C. elegans contains nine core APC/C components, as well as two accessory components, FZY-1 (also known as CDC20) and FZR-1 (CDH1; Yeong, 2004).C. elegans and budding yeast were the first organisms in which a role for the APC/C in chromosome separation during meiosis I was demonstrated (Salah and Nasmyth, 2000;Yeong, 2004).Mutations of APC/C FZY-1 components produce failures of chromosome separation during mitosis and meiosis I of both oocyte and sperm lineages (Furuta et al., 2000;Davis et al., 2002;Shakes et al., 2003;Kitagawa et al., 2002).Inactivation of APC/C FZY-1 components causes a one-cell arrest at metaphase of meiosis I that is similar to what is observed upon inactivation of proteasome components, suggesting that APC/C FZY-1 mediates this initial requirement for ubiquitin proteolysis in the embryo (Furuta et al., 2000;Davis et al., 2002;Shakes et al., 2003;Kitagawa et al., 2002;Gonczy et al., 2000).
C. elegans APC/C is required for the degradation of IFY-1, the proposed Securin that functions to release Separase, SEP-1, which is a conserved protease that separates sister chromatids at anaphase (Siomos et al., 2001;Kitagawa et al., 2002).However, the degradation of Securin cannot be the sole essential function of APC/C FZY-1 in promoting meiosis I, as experimentally-inducing a loss of chromosome cohesion in APC/C mutants (which should bypass the requirement for Securin degradation) does not rescue the meiosis I arrest (Davis et al., 2002).The APC/C-mediated release of active SEP-1 has been suggested to directly affect anterior-posterior polarity in the one-cell stage embryo by ensuring that the paternal pronucleus/centrosome complex remains in tight association with the posterior cortex of the embryo, where it promotes the cortical association of the PAR-2 polarity protein (Rappleye et al., 2002).Hypomorphic alleles of APC/C or RNAi depletion of SEP-1 are associated with a lack of embryonic polarity due to a failure of PAR-2 to localize to the posterior cortex (Rappleye et al., 2002).However, another study has suggested that the polarity defects are secondary consequences of a failure of meiosis and do not imply a direct regulation of polarity by APC/C or SEP-1 (Shakes et al., 2003).
Genetic experiments have implicated APC/C FZR-1 in the negative regulation of cyclins during G1 phase (Fay et al., 2003).In mammals, the Rb protein negatively regulates the transcription of cyclins in G1 phase (Peters, 2002).Weak alleles of fzr-1 have no overt phenotypes by themselves, but produce a synthetic hyperplasia phenotype when combined with homozygous lin-35 Rb mutant alleles.Additionally, overexpression of cyclins A and E produce more extensive hyperplasia in a fzr-1 mutant background, suggesting that APC/C FZR-1 degrades S phase and mitotic cyclins during G1 phase, as occurs in other metazoa (Fay et al., 2003).More complete inactivation of fzr-1 by RNAi reveals severe pleiotropic effects on cell proliferation and development (Fay et al., 2003).Biochemical and genetic experiments indicate that APC/C functions with the E2 UBC-2 to promote both meiosis and mitosis (Frazier et al., 2004).
The APC also has a non-cell cycle function to regulate the abundance of GLR-1 glutamate receptors in ventral cord nerve cells (Juo and Kaplan, 2004).The endocytosis of GLR-1 is induced by the covalent attachment of one or a few Ub to GLR-1 (Burbea et al., 2002).APC/C promotes GLR-1 endocytosis; however, APC/C does not directly ubiquitinate GLR-1, and the critical target of APC/C in regulating GLR-1 endocytosis is not yet known (Juo and Kaplan, 2004).

Figure 1 .
Figure 1.A) Alignment of individual ubiquitin polypeptides from C. elegans (C.e.), H. sapiens (H.s.), and S. cerevisiae(S.c.).Differences between Ub residues are boxed.Note that C. elegans and humans only differ at one position relative to each other and at three positions relative to budding yeast.The locations of the seven lysines in Ub are marked with the residue numbers provided above the alignment.B) Translation of the C. elegans ubq-1 polyubiquitin gene.The amino acid sequence is presented at 76 amino acids per line so that individual ubiquitin repeats are aligned.Note that repeat 6 has a different amino acid at position 9 of the repeat (yellow box, isoleucine rather than threonine).There are two amino acids after the final ubiquitin repeat.

Figure 2 .
Figure 2. A) Structural model of HECT-domain E3 complex.The mechanism for the conjugation of Ub to a substrate by a HECT-domain E3 is shown.The E2 binds the N-terminal lobe of the HECT-domain E3 (top) and transfers Ub to the C-terminal lobe via a thiolester linkage (middle).The C-terminal lobe swivels on a hinge-loop and catalyzes the transfer of Ub to the substrate protein (bottom).B) Model of U-box or monomeric RING finger E3s.The U-box or RING finger domains of the E3 are directly involved in binding the E2.C) Model of SCF complexes.The N-terminus of CUL-1 binds the adaptor Skp1 (SKR proteins in C. elegans), while the C-terminus binds the RING finger protein Rbx1, which binds the E2.The substrate recognition subunit (SRS) binds to Skp1 through an F-box motif.D) Model of CUL-2-based E3 complexes.The N-terminus of CUL-2 binds to the adaptor elongin C (ELC-1), which is in complex with elongin B (ELB-1).The SRS binds to elongin C through a BC-box motif.E) Model of CUL-3-based E3 complexes.The N-terminus of CUL-3 binds directly to the SRS, which utilizes a BTB/POZ domain to bind to CUL-3.

Table 3 .
U-box-domain encoding genes in C. elegans.Only the more severe phenotypes are listed.Phenotypes derived from RNAi or mutant analysis are denoted (brackets).WT = wild-type phenotype.References for the phenotype listed are given.

Table 4 .
RING finger encoding genes in C. elegans.