Two functionally distinct E2/E3 pairs coordinate sequential ubiquitination of a common substrate in Caenorhabditis elegans development

Abstract

Ubiquitination, the crucial posttranslational modification that regulates the eukaryotic proteome, is carried out by a trio of enzymes, known as E1 [ubiquitin (Ub)-activating enzyme], E2 (Ub-conjugating enzyme), and E3 (Ub ligase). Although most E2s can work with any of the three mechanistically distinct classes of E3s, the E2 UBCH7 is unable to function with really interesting new gene (RING)-type E3s, thereby restricting it to homologous to E6AP C-terminus (HECT) and RING-in-between-RING (RBR) E3s. The Caenorhabditis elegans UBCH7 homolog, UBC-18, plays a critical role in developmental processes through its cooperation with the RBR E3 ARI-1 (HHARI in humans). We discovered that another E2, ubc-3, interacts genetically with ubc-18 in an unbiased genome-wide RNAi screen in C. elegans These two E2s have nonoverlapping biochemical activities, and each is dedicated to distinct classes of E3s. UBC-3 is the ortholog of CDC34 that functions specifically with Cullin-RING E3 ligases, such as SCF (Skp1-Cullin-F-box). Our genetic and biochemical studies show that UBCH7 (UBC-18) and the RBR E3 HHARI (ARI-1) coordinate with CDC34 (UBC-3) and an SCF E3 complex to ubiquitinate a common substrate, a SKP1-related protein. We show that UBCH7/HHARI primes the substrate with a single Ub in the presence of CUL-1, and that CDC34 is required to build chains onto the Ub-primed substrate. Our study reveals that the association and coordination of two distinct E2/E3 pairs play essential roles in a developmental pathway and suggests that cooperative action among E3s is a conserved feature from worms to humans.

Keywords: CDC34; Cullin-RING ligase; RBR E3 ligase; UBCH7; ubiquitin.

Fig. 1.

RNAi screen identified a synthetic genetic interaction between ubc-18 and ubc-3. (A) Quantitation of synthetic interaction between ubc-3 and ubc-18, measured as percent survival to adulthood. The pL4440 empty vector was included as the negative control, and lin-35 served as a positive control (8). Table 1 contains a list of other synthetic interactions identified in the RNAi screen. Error bars indicate the SD of at least three independent experiments. (B) Synthetic lethality between ubc-18(ku354) and ubc-3(RNAi) is associated with a Pharynx unattached (Pun) phenotype. In each image, the pharynx basement membrane is traced with the yellow dashed line. (Top) Wild-type L1 grown on ubc-3(RNAi). (Bottom) ubc-18(tm5426); ubc-3(RNAi) L1. (C) Quantification of the synthetic Pun phenotype. The percentage of Pun L1 observed in animals grown on the indicated RNAi for ubc-18(ku354) (checkered bars) and pha-1(e2123) mutant L1s (striped bars) is shown. The N2 WT control (wt, black bars) did not produce Pun progeny under any of the experimental conditions. Error bars indicate the SD of at least three independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001 relative to L4440 controls. The absence of asterisks implies no significant difference relative to L4440 controls.

Fig. S1.

Synthetic genetic interactions between ubc-18(tm5426), and ubc-3(RNAi). Viability of animals grown on ubc-3(RNAi) was assessed for WT (N2, black bars) and ubc-18(tm5246) animals as in Fig. 1B. Error bars indicate the SD of at least three independent experiments.

Fig. S2.

Fraction of Pun L1, as in Fig. 1C. The percentage of Pun F1 progeny from animals grown on the indicated RNAi for ubc-18(tm5426) (white bars) and pha-1(e2123ts) (striped bars) is shown. Experiments with pha-1(e2123ts) were conducted at 15 °C. The isogenic wild-type control (N2; black bars) did not form Pun progeny under any of the experimental conditions. Error bars indicate the SD of at least three independent experiments.

Fig. S3.

Synthetic genetic interactions with skr-1/2. Viability of WT, N2 (black bars), ubc-18(ku354), and ubc-18(ku354); sup-36(e2217) grown on the indicated RNAi was measured. Error bars indicate the SD of at least three independent experiments.

Fig. 2.

ubc-18 interacts synthetically with genes encoding SCF E3 ubiquitination machinery. (A) Quantification of synthetic lethal interactions with ubc-18. The percentage of F1 progeny of RNAi-fed animals that survive to adulthood is shown for WT (wt) and ubc-18(tm5426) strains (black and white bars, respectively). The rrf-3(pk1426) mutant served as the wild-type control. This mutant is wild type at the ubc-18 locus and reported to be RNAi-sensitive (49). Error bars are the SD between technical replicates in one representative experiment. The candidate screen was performed only once for genes that were not affected. Synthetic interactions with skr-1/2, cul-1, and cand-1 were observed in at least three independent experiments. (B) ubc-18 does not genetically interact with cul-2(or209ts). The percentage of F1 progeny of RNAi-fed animals that survived to adulthood is shown for cul-2(or209ts). Error bars are the SD between two independent experiments. (C) cand-1 genetically interacts with lin-35 and skr-2/1. The percentage of F1 progeny viable to adulthood is shown for N2 WT (wt, black bars) and cand-1(tm1683) mutant animals (gray bars) grown on each RNAi. Error bars represent the SD of at least three independent experiments.

Fig. 3.

SUP-36 has SKP1-like properties. (A) The sup-36(e2217) null allele suppresses the synthetic genetic interactions between ubc-18 and SCF components. (B) Yeast two-hybrid experiments with either CUL-1 or CUL-6 fused to the Gal4 DNA-binding domain and SKR-1, SKR-2, or SUP-36 fused to the Gal4 DNA-activating domain. (Left) Growth on selective media (−His, −Leu, −Trp) indicates interactions of the two proteins tested. As shown previously, CUL-1 binds to both SKR-1 and SKR-2. SUP-36 binds to CUL-1 (Top), but not to CUL-6 (Bottom). (Right) Yeast were also spotted on nonselective media (−Leu, −Trp) to control for growth.

Fig. 4.

SUP-36 protein levels are regulated by UBCH7/HHARI (UBC-18/ARI-1) and CDC34/CUL-1 (UBC-3/CUL-1). (A) Representative images for SUP-36::GFP expression levels measured in comma to twofold embryos under indicated RNAi conditions. (B) Quantification of embryos in the comma to twofold stage [L4440 vector, n = 27; ubc-18(RNAi), n = 45; ubc-3(RNAi), n = 67; ubc-18(RNAi);ubc-3(RNAi), n = 95]. Boxplots show quartiles, with mean indicated by the thick black line, and whiskers show 95% confidence intervals. Statistical significance was determined via Welch’s unequal variance t tests. Measurements for embryos in early embryogenesis and twofold to threefold stage are provided in Fig. S5A. (C) In vitro immunoprecipitation (IP) using recombinant T7-SUP-36 (IP: T7) and recombinant, untagged N8-CUL-1. Results were visualized by Western blot analysis with indicated antibodies. Replicates are shown in Fig. S5B. (D) In vitro ubiquitination of SUP-36. Each reaction contained T7-SUP-36 and other reaction components as indicated. All time points are post-ATP addition, except lane 1 (0 min). Longer time courses are shown in Fig. S6A. (E) The T7-SUP-36 in vitro ubiquitination assay was performed using Lys-less Ub, which is unable to form chains. Unmodified SUP-36 and single-ubiquitinated and double-ubiquitinated T7-SUP-36 are indicated by arrows. All time points are 10 min after ATP addition; longer time courses are shown in Fig. S6 A and B.

Fig. S4.

HHARI_RBR does not form poly-Ub chains. Autoubiquitination assays in which the GST-tagged E3 (HHARI) served as a proxy substrate were performed with either WT Ub (Left) or Lys-less Ub (Right). The construct used, GST-HHARI_RBR, lacks the Ariadne domain and thus is constitutively active (32). Lys-less Ub is unable to form Ub-chains; therefore, the higher molecular weight bands observed are numerous multiple mono-Ub modifications. Products were visualized by Western blot analysis against the GST-tag on HHARI.

Fig. S5.

(A) SUP-36::GFP expression levels were measured in early embryogenesis and twofold to threefold embryos under indicated RNAi conditions. (Left) Quantification of SUP-36 expression in early embryogenesis. L4440 vector control, n = 15; ubc-18(RNAi), n = 19; ubc-3(RNAi), n = 41; ubc-18(RNAi);ubc-3(RNAi), n = 32. (Right) Quantification of SUP-36 levels at the twofold to threefold stage. L4440 vector control, n = 30; ubc-18(RNAi), n = 65; ubc-3(RNAi), n = 72; ubc-18(RNAi);ubc-3(RNAi), n = 60. Statistical significance was determined using Welch’s unequal variance t test, with P < 0.05. (B) Three replicates of in vitro IP using recombinant T7-SUP-36 (IP: T7) and recombinant, untagged N8-CUL-1. The SUP-36^N109K/F110K double mutant at the predicted SUP-36/CUL-1 interface retains binding to N8-CUL-1, in contrast to results reported for the analogous mutation in human SKP1 (13). Results were visualized by Western blot analysis with the indicated antibodies. The intensities of elution bands were quantified using ImageJ. Ratios of intensities (for either CUL-1 or SUP-36, as indicated) in mutant SUP-36 elutions compared with WT SUP-36 elutions are shown. (C, Left) A SUP-36 homology model was generated using Phyre2 (50) and then overlaid with the structure of human SKP1 bound to CUL-1 (Protein Data Bank ID code 1LDK; SUP-36 is shown in red, SKP1 in yellow, and CUL-1 in blue). SUP-36 residues N109 and F110 overlay closely with N108/Y110 of human SKP1 at the SKP1/CUL-1 interface (13). (C, Right) Region of clustal omega sequence alignment containing SUP-36 residues N109 and F110 (red) and the aligned sequences of (worm) SKR-1 and (human) SKP-1.

Fig. S6.

(A) SUP-36 in vitro ubiquitination assay similar to the assay shown in Fig. 4D, but with longer time points displayed. Each reaction contained T7-SUP-36 and other reaction components as indicated below the gel. All time points are after ATP addition, except lane 1 (0 min). In the presence of CUL-1, UBCH7/HHARI modify SUP-36 with one and two single Ubs (lanes 5–7), whereas no product is detected when CDC34 is the sole E2 (lanes 8–10). All four components (UBCH7/HHARI and CDC34/CUL-1) are required for the formation of poly-Ub chains on SUP-36 (lanes 11–13). (B) Longer time points of the Lys-less Ub assay shown in Fig. 4E. SUP-36 is modified with one or two single Ubs in the presence of all four components. The same pattern is observed when CDC34 is omitted, confirming that CDC34 is required for Ub chain formation (in the presence of WT Ub), and that the multiple SUP-36 bands seen in Fig. 4D (lanes 8–9) and A (lanes 5–7) arise from addition of multiple mono-Ubs. Unmodified SUP-36 and mono- and double-Ub T7-SUP-36 are indicated by arrows. (C) Precharged UBCH7∼Ub was added to reactions containing HHARI, N8-CUL-1, and T7-SUP-36. The time points shown are after UBCH7∼Ub addition. (The 0 s time point was taken immediately before UBCH7∼Ub addition.) The results were visualized by Western blot analysis against T7 (SUP-36). (D) The same assay as in C, but using the SUP-36^N109K/F110K mutant at the predicted SUP-36/CUL-1 interface.

Fig. 5.

Proposed model for UBC-18/UBC-3-dependent ubiquitination of SUP-36. (A) A sequential model in which UBC-18/ARI-1 acts coordinately with UBC-3/CUL-1 to regulate the proteasomal degradation of SUP-36 is consistent with genetic and biochemical data (this work and others). In the model, CUL-1 binds SUP-36 and binds and activates ARI-1, which, together with UBC-18, monoubiquitinates SUP-36. The E2 UBC-3, with its E3 CUL-1, then builds poly-Ub chains using mono-Ub-SUP-36 as its substrate, thereby promoting the degradation of SUP-36. (B) SUP-36 protein levels increase when ubc-18, ari-1, or ubc-3 is depleted by RNAi (this work and ref. 19), and SUP-36 levels are even higher when both ubc-18 and ubc-3 are depleted, leading to inhibition of pharyngeal morphogenesis (this work). (C) A model in which SUP-36 inhibits formation of a canonical SKR-1/CUL-1 complex, and this inhibition is released when SUP-36 is ubiquitinated by UBC-18/ARI-1 and UBC-3/CUL-1, leading to its degradation, is consistent with these observations: (i) a functional SKR-1/CUL-1 complex is required for pharyngeal morphogenesis; (ii) the synthetic phenotype of ubc-18;skr-1/2(RNAi) can be rescued by a sup-36 null allele; and (iii) the SKP1-like SUP-36 can bind to CUL-1, and its degradation is dependent on UBC-18/ARI-1/UBC-3.

Fig. S7.

Several potential parallel pathways that are consistent with genetic data presented herein. We cannot rule out the formal possibility that another E3 with polyubiquitinating activity may perform the elongation step following the action of UBC18/ARI-1 on SUP-36, or that UBC-18 and UBC-3 use their different functions to act in parallel pathways. Combinations of alternative models 1–3 with either model 4 or 5 also would be consistent with the genetic results.