Ubiquitin binding by a CUE domain regulates ubiquitin chain formation by ERAD E3 ligases

Ubiquitin Binding by a CUE Domain Regulates Ubiquitin Chain Formation by ERAD E3 Ligases

Katrin Bagola,

Maximilian von Delbru¨ck,

Gunnar Dittmar,

Martin Scheffner,

Inbal Ziv,

Michael H. Glickman,

Aaron Ciechanover,

and Thomas Sommer

*

1Max-Delbru¨ck-Center for Molecular Medicine, Robert-Ro¨ssle-Strasse 10, D-13122 Berlin, Germany

2Humboldt-University zu Berlin, Institute for Biology, Invalidenstrasse 43, D-10115 Berlin, Germany

3Department of Biology, Konstanz Research School Chemical Biology, University of Konstanz, D-78457 Konstanz, Germany

4Department of Biology, Technion-Israel Institute of Technology, 3200 Haifa, Israel

5Cancer and Vascular Biology Research Center, The Rappaport Faculty of Medicine and Polak Cancer Center, Technion-Israel Institute of Technology, Haifa 31096, Israel

*Correspondence:tsommer@mdc-berlin.de

SUMMARY

Ubiquitin-binding domains (UBDs) differentially recognize ubiquitin (ub) modifications. Some of them specifically bind mono-ub, as has been shown for the CUE domain. Interestingly, so far no signifi- cant ubiquitin binding has been observed for the CUE domain of yeast Cue1p. Cue1p is receptor and activator of the ubiquitin-conjugating enzyme Ubc7p. It integrates Ubc7p into endoplasmic reticu- lum (ER) membrane-bound ubiquitin ligase com- plexes, and thus, it is crucial for ER-associated protein degradation (ERAD). Here we show that the CUE domain of Cue1p binds ubiquitin chains, which is pivotal for the efficient formation of K48-linked polyubiquitin chains in vitro. Mutations that abolish ubiquitin binding by Cue1p affect the turnover of ERAD substrates in vivo. Our data strongly imply that the CUE domain facilitates substrate ubiquityla- tion by stabilizing growing ubiquitin chains at the ERAD ubiquitin ligases. Hence, we demonstrate an unexpected function of a UBD in the regulation of ubiquitin chain synthesis.

INTRODUCTION

Ubiquitin-binding domains (UBDs) interact with the small 76 amino acid polypeptide ubiquitin that is covalently conjugated to proteins as a posttranslational modification (Hicke et al., 2005). By means of an enzymatic cascade, involving ubiquitin- activating enzymes (E1), ubiquitin-conjugating enzymes (E2s), and ubiquitin ligases (E3s), the C-terminal glycine residue of ubiquitin is covalently attached to a lysine residue of a client protein. Besides this monoubiquitylation of a protein, polyubi- quitylation is a result of further conjugation of multiple ubiquitin molecules via one of the seven lysine residues within ubiquitin.

Protein ubiquitylation functions as signaling mechanism (Hershko and Ciechanover, 1998; Haglund and Dikic, 2005) that triggers different events such as the assembly of protein complexes and the sorting of ubiquitylated proteins for trans- port or for proteolysis by the 26S proteasome (Hurley et al., 2006; Dikic et al., 2009). To this end, the diverse ubiquitin sig- nals are differentially recognized and decoded by UBDs. So far, only few UBDs were shown to be directly involved in the ubiquitylation process (Chen et al., 2006) and to facilitate protein monoubiquitylation (Shih et al., 2003; Polo et al., 2002;

Hoeller et al., 2007).

The UBDs UBA (ubiquitin-associated) (Madura, 2002) and CUE (Ponting, 2000; Shih et al., 2003) interact with ubiquitin via a similar three

helix bundle, whereas other UBDs bind ubiquitin via a different tertiary structure (Randles and Walters, 2012;

Searle et al., 2012). The affinity for both mono- or polyubiquity- lated substrates and differently linked ubiquitin chains can vary greatly (Raasi et al., 2005; Dikic and Do¨tsch, 2009; Husnjak and Dikic, 2012).

In

ubiquitin-conjugation to ER degradation), containing a CUE domain, recruits the soluble ubiquitin-conjugating enzyme Ubc7p to the membrane-bound RING ubiquitin ligases Hrd1p and Doa10p (Biederer et al., 1997), which are key factors for ER-associated protein degradation (ERAD) (Smith et al., 2011; Hirsch et al., 2009). Although ligase complex formation and activation of Ubc7p depend on the C-terminal part of Cue1p (Bazirgan and Hampton, 2008; Kostova et al., 2009), neither a significant binding of ubiquitin (Shih et al., 2003) nor another functional impact could be described for its CUE domain.

Here we present that the CUE domain of Cue1p binds lysine 48-linked ubiquitin chains. This function is a prerequisite for the efficient formation of higher-molecular-weight ubiquitin chains by Ubc7p and the E3 ligases in vitro and is necessary for the degradation of membrane-bound ERAD substrate proteins in vivo. Thus, our findings demonstrate a regulation of polyubiquitylation activity of ubiquitin ligase complexes by a UBD.

528

Erschienen in: Molecular Cell ; 50 (2013), 4. - S. 528-539

https://dx.doi.org/10.1016/j.molcel.2013.04.005

RESULTS

Cue1p Enhances the Activity of Both ERAD Ligases

We set out to examine ubiquitin chain formation by Ubc7p, Cue1p, and Hrd1p or Doa10p by an in vitro ubiquitylation assay. To this end, all proteins, except the ubiquitin-activating enzyme and ubiquitin, were expressed in and purified from

(Figure 1A and see Figure S1A and Table S1 online).

Since Cue1p and both ubiquitin ligases are integral mem- brane proteins, we expressed the soluble cytosolic regions (Figure 1A, shown in brackets). The fragments of Hrd1p and Doa10p harbored the RING domain, whereas Cue1p encom- passed the Ubc7p-binding region and the internal CUE domain. Equal amounts of Ubc7p, Cue1p, or Hrd1p were incubated with Flag epitope-tagged ubiquitin, E1, and ATP.

In addition to free monoubiquitin, the reaction containing Ubc7p was enriched in diubiquitin (Figure 1B). Reactions con- taining Ubc7p and Cue1p or Ubc7p and Hrd1p resulted in the formation of higher-molecular-weight ubiquitin chains, although the ubiquitylation pattern differed slightly. In line with Bazirgan and Hampton (2008), incubation of Ubc7p, Cue1p, and Hrd1p led to a substantial increase in polyubi- quitin chains. Comparable results were obtained with Doa10p instead of Hrd1p (Figure 1C). Hence, the addition of Cue1p

Figure 1. Cue1p Stimulates the Ubiquityla tion Activity of Ubc7p at Both ERAD Ligases (A) Schematic representation of Ubc7p, Cue1p, Hrd1p, and Doa10p. Full length Ubc7p and soluble parts of Cue1p, Hrd1p, and Doa10p, indi cated by brackets, were used for heterologous expression in E. coli and the following in vitro ubiquitylation assays.

(B and C) In vitro ubiquitylation reactions with Ubc7p, Cue1p, and the ERAD ligases Hrd1p or Doa10p. Equal amounts (3.5 mM) of the indicated proteins were mixed together with 7.2mM Flag epitope tagged ubiquitin, E1 enzyme, and ATP in a buffer system followed by a 15 min incubation in a 30C water bath. Ubiquitin was detected by a monoclonal anti Flag antibody.

Small amounts of diubiquitin were detected in all reactions due to preformed Flag ubiquitin conjugates. Presence of Ubc7p, Cue1p, Hrd1p, or Doa10p was verified by protein specific immunoblotting. (B) Reactions were performed with the soluble C terminal part of Hrd1p (amino acids 325 552) containing the RING motif. (C) Reactions with the N terminal soluble part of Doa10p (amino acids 1 125) which com prises the RING.

dramatically stimulated polyubiquitin chain formation by Ubc7p and the E3s.

The synthesized chains were mostly unanchored, since no substrate had been added to the reactions. We de- tected autoubiquitylation of neither the E3 ligases nor Ubc7p. However, some ubiquitin was conjugated to Cue1p (Fig- ure S1B). Still, we could never detect ubiquitylated Cue1p in yeast cell extracts.

ERAD Ligases Assemble Lysine 48-Linked Ubiquitin Chains

Because ubiquitin chains linked via different internal lysines

have been found in membrane preparations of yeast cells (Xu

et al., 2009), we wanted to determine which polyubiquitin link-

age type is required for the degradation of the well-established

ERAD substrate CPY*, a mutant form of carboxypeptidase Y

(Finger et al., 1993). We analyzed the amount of CPY* within

3 hr after cycloheximide treatment (Walter et al., 2001) in cells

overexpressing either ubiquitin mutants in which individual

lysine residues were mutated to arginine (K-R, Figure 2A) (Finley

et al., 1994) or ubiquitin mutants containing only one of its

seven lysine residues (K-only, Figure 2B). As shown in Figure 2A,

CPY* was rapidly degraded upon expression of WT, K6R,

K11R, K27R, or K63R ubiquitin variants (also K29R and K33R,

data not shown), whereas the overexpression of the K48R

ubiquitin strongly delayed CPY* degradation. Regarding the

K-only ubiquitin mutants (Figure 2B), overexpression of wild-

type ubiquitin and the K48-only ubiquitin mutant enabled faster

CPY* degradation, while CPY* turnover was decelerated in the

other mutants. From this we concluded that polyubiquitin chains

linked via lysine 48 facilitate the degradation of the ERAD model substrate.

Considering that CPY* is ubiquitylated by the HRD ligase com- plex (Bordallo and Wolf, 1999; Bays et al., 2001), we performed in vitro ubiquitylation reactions to test whether Ubc7p, Cue1p, and Hrd1p assemble K48-linked ubiquitin chains. A mass spec- trometric analysis (Figure 2C and Figure S2A) revealed exclusive formation of K48-linked ubiquitin conjugates. This suggested that the ubiquitin linkage is likely defined or at least mediated by Ubc7p (see also Metzger et al., 2013). This was further confirmed by in vitro ubiquitylation reactions of Ubc7p and Cue1p with either Hrd1p or Doa10p using different ubiquitin var- iants (K-R and K-only) (Figures S2B–S2E). Hence, K48-linked ubiquitin chains are assembled by both Hrd1p and Doa10p ligase complexes.

Different Parts of Cue1p Contribute to the High Ubiquitylation Activity of the Ligase Complexes

Kostova et al. (2009) had shown that Ubc7p is recruited to the ER membrane by a C-terminal Ubc7p-binding region in Cue1p (Cue1p

) that also activates Ubc7p-dependent polyubiqui- tylation in vitro. We examined this activating effect in more detail in an in vitro ubiquitylation assay containing Ubc7p, Alexa Fluor

Figure 2. K48 Linked Ubiquitin Chain For mation by the HRD Ligase Complex Facili tates Degradation of ERAD Substrates (A and B) Degradation of the ERAD model sub strate CPY* (mutant carboxypeptidase Y, left panel) was analyzed in a cycloheximide (CHX) decay assay. (A) Subsequent to 4 hr CuSO4

2

induced overexpression of plasmids encoding wild type ubiquitin or the different indicated K R ubiquitin mutants, cells were analyzed after 0, 40, 80, or 120 min. The ER membrane bound protein Sec61p (right panel) served as loading control.

(B) Cells with constitutively overexpressed N terminally RGS 8xHis tagged wild type ubiq uitin or K only ubiquitin mutants were lysed at indicated time points.

(C) Quantitative analysis of ubiquitin linkages by mass spectrometry. Different in vitro ubiquitylation reactions with Ubc7p (U7), Cue1p (C1), or Hrd1p (H1) and bovine ubiquitin (ub) were measured in selected reaction monitoring (SRM) in comparison to human ubiquitin standards (ub with identical amino acid sequence). The reaction with only E1 enzyme and ubiquitin served as negative control.

Standard error bars were obtained from three technical replicates. Error bars indicate the stan dard deviation of three technical replicates.

488-labeled ubiquitin, and either wild-

type Cue1p or the short Ubc7p-binding

region (sU7BR, residues 147–203). Chain

formation was monitored in time course

reactions in the absence or presence of

Hrd1p by scanning fluorescence inten-

sities in SDS gels. As shown in Figure 3A,

we observed a similar activity of ubiquitin

chain formation for Cue1p and sU7BR in reactions without

Hrd1p. However, in the presence of Hrd1p, Cue1p stimulated

polyubiquitylation much stronger than did sU7BR. This indicated

that parts of Cue1p alongside the Ubc7p-binding region are

needed to obtain full activity of the enzymes. To pursue this

idea, we generated a longer version of the Ubc7p-binding

construct (lU7BR, residues 117–203) and a Cue1p variant with

point mutations in the CUE domain (Cue1pRGA, leu-ala-pro at

position 76–78 substituted by arg-gly-ala) (see also Figure 4B,

Figure S3A). In the presence of Hrd1p, the sU7BR fragment

caused a moderate stimulation of the ubiquitylation reaction

compared to the reaction without Cue1p or a C-terminally trun-

cated version of Cue1p lacking parts of the Ubc7-binding region

(Cue1p

C, residues 24–174) (Figure 3B) (see also Metzger et al.,

2013). Interestingly, addition of the lU7BR or Cue1pRGA variant

resulted in enhanced chain formation but did not reach the

extent of ubiquitylation observed with wild-type Cue1p. The dif-

ferences in ubiquitin chain formation with either wild-type Cue1p

or the Cue1p variants could be observed independently of the

amount of ubiquitin added to the reaction or an epitope tag

that was fused to ubiquitin (Figures S3B–S3D). Therefore,

regions more N-terminally in Cue1p, most likely the CUE domain,

contributed to ubiquitin chain formation.

Cue1p Binds Ubiquitin Chains via an Unconventional CUE Domain

CUE domains were shown to efficiently bind mono- or polyubi- quitin. By contrast, the CUE domain of Cue1p has only a weak affinity to monoubiquitin (Shih et al., 2003). As determined by multiple sequence alignments, CUE domains of various species contain some highly conserved amino acids (Figure 4A). A MFP motif with invariant proline and a dileucine motif more C-terminal in the CUE domain are required for high-affinity binding of ubiq- uitin (Shih et al., 2003; Kang et al., 2003). Although these motifs are less well conserved in Cue1p, we wanted to reinvestigate a potential ubiquitin binding activity of this protein.

We generated a variety of additional Cue1p constructs, as illustrated in Figure 4B. Besides the minimal consensus CUE domain (CUE

) (Shih et al., 2003), we tested an elongated re- gion, either as wild-type (CUE

) or as a mutated version (CUE

). In addition to the Cue1pRGA variant (see above),

we also deleted the entire CUE domain of Cue1p (

65-106, Cue1p

CUE). These constructs, as well as wild-type Cue1p and the UBA domain of Dsk2p, were heterologously expressed as GST fusion proteins in

and immobilized on glutathione Sepharose. Binding of K48-linked ubiquitin chains was analyzed by detecting the amount of bound ubiquitin chains and the remainder in the supernatant (Figure 4C, Figure S4A). Ubiquitin chains were bound by wild-type Cue1p as well as the UBA domain of Dsk2p that served as a positive control. Interestingly, while binding of diubiquitin was very weak, Cue1p bound longer ubiquitin chains more efficiently. Ubiquitin binding was abro- gated by mutations within the CUE domain. Moreover, the minimal consensus CUE domain did not bind ubiquitin chains, whereas the extended CUE domain exhibited ubiquitin binding activity. Again, point mutations within this extended CUE domain resulted in loss of binding. In line with previous data (Biederer et al., 1997), Cue1p did not bind monoubiquitin, whereas the

(A) In vitro ubiquitylation activity of Ubc7p was examined over a time period of 15 min. Incubation of Ubc7p with Alexa Fluor 488 labeled ubiquitin in the absence or presence of Hrd1p, Cue1p, or the short Ubc7p binding region (sU7BR, amino acids 147 203) of Cue1p. Sample analysis was performed by SDS PAGE, followed by fluorescence scan of the gels.

(B) Time course of in vitro ubiquitin chain formation. Different N or C terminally truncated Cue1p variants, wild type Cue1p, and the mutant form Cue1pRGA were added to reactions with Ubc7p, Hrd1p, and Flag epitope tagged ubiquitin, and chain formation was analyzed at indicated time points (0, 8, 16, or 24 min). Mono and oligoubiquitin (upper panel, 18% SDS gel) as well as higher molecular weight ubiquitin chains (lower panel, 9% SDS gel) were visualized by immunoblotting using an anti Flag antibody. Sample aliquots from time point 0 were also analyzed by western blotting to visualize the amounts of the added Cue1p variants. Since the anti Cue1 peptide antibody recognizes the C terminus of Cue1p, this antibody was not able to bind to the Cue1pDC variant. Amounts of the Cue1p variants can therefore be also compared from a Coomassie stained gel (Figure S3A).

UBA domain of Dsk2 showed weak monoubiquitin binding (Fig- ure S4B). To exclude the possibility that binding of ubiquitin chains was due to an artificial dimerization of two Cue1 proteins via the GST tag, we performed a similar pull-down experiment using Cue1p variants without the GST fusion, which were immo- bilized via their C-terminal His

tag (Figures S4C and S4D).

Results obtained with this setup were comparable to those ob- tained from the GST pull-down experiment. Wild-type Cue1p,

Figure 4. Cue1p Binds Ubiquitin through Its Extended CUE Domain

(A) Alignment of amino acid sequences of CUE domains from various yeast or mammalian (gp78) proteins. Numbers indicate the respective resi dues. The sequence of the Cue1p CUE domain is written in bold, and the LAP motif used for mutagenesis is framed by a black box. Highly conserved amino acids are highlighted in yellow, and those known to be absolutely required for ubiquitin binding are highlighted in red.

(B) Schematic depiction of Cue1p constructs used in the in vitro ubiquitylation assay or in vitro binding studies. All variants were purified from E. coli through an N terminal GST tag that was removed by Prescission Protease cleavage for ubiq uitylation experiments. The C terminal His6tag of some constructs was added to prevent expression and purification of degradation products.

(C) Binding of ubiquitin chains to immobilized GST fusion proteins. GST, GST Cue1p con structs, and GST UBA (Dsk2) were purified from E. coli, immobilized by binding to glutathione Sepharose, and incubated with equal amounts of ubiquitin chains. After binding, TCA precipitated supernatant (S) and material sedimented with the Sepharose (B) were loaded to 18% SDS gels.

(C) Binding of preformed K48 linked ubiquitin chains (ub2 8) was analyzed by anti ubiquitin immunoblotting. The lower panel visualizes the amount of GST fusion protein for each binding sample.

(D) In vitro deubiquitylation of K48 linked ubiquitin chains by Otubain 1 in a 4 min progress. Equal amounts of ubiquitin chains were mixed with either BSA, Cue1p, or Cue1pRGA. Time point 0 shows the ubiquitin chains before addition of Otubain 1. Incubation with Otubain 1 for 360 s on ice. Aliquots were taken at indicated time points.

and Cue1DC showed binding of ubiquitin chains, whereas mutation of the CUE domain impeded this function.

Ubiquitin binding to Cue1p was further verified by a ubiquitin-specific protease protection experiment (Figure 4D). Here, the human K48-specific deubiquitylating enzyme Otubain-1 was added to K48- linked ubiquitin chains in the presence of either wild-type Cue1p or Cue1pRGA.

While ubiquitin conjugates were clearly disassembled in the BSA control and in the presence of Cue1pRGA, hydrolysis of the chains was delayed in the presence of wild-type Cue1p, most likely due to decreased access of the deubiquitinating enzyme to the ubiqiui- tin chains.

We therefore concluded that Cue1p binds ubiquitin chains via

an extended version of the consensus CUE domain, whereas the

Ubc7p-binding region of Cue1p did not to contribute to this

function.

Efficient Formation of Longer Ubiquitin Chains Depends on the Cue1p CUE Domain

Next we wanted to investigate the impact of the CUE domain on ubiquitin chain formation. We therefore tested the in vitro ubiquitylation efficiency of purified wild-type Cue1p and three CUE domain mutants, Cue1pRGA, Cue1p

LAP (

76-78), and Cue1p

CUE (Figure 4B, Figure S3A), in a time-dependent reac- tion with Ubc7p and Hrd1p (Figure 5A). In samples containing wild-type Cue1p, large amounts of high-molecular-weight (hmw) ubiquitin chains were assembled after 8 min. The number and size of chains increased within 24 min of incubation. In contrast, ubiquitin chain formation was apparently delayed in all three CUE domain mutants, as there were fewer chains detectable at the 8 min time point. Additionally, there was a remarkable accumulation of diubiquitin in all samples generated with CUE domain mutants. Although the size of the synthesized ubiquitin chains increased with time, it never reached the chain length observed for wild-type Cue1p. Similar effects were ob- tained in reactions with Doa10p (Figure 5B).

In addition to immunoblotting, we took advantage of the Alexa Flour 488-labeled ubiquitin and quantified a time course of ubiq- uitin chain formation by fluorescence measurement (Figures S5A and S5B). In comparison to wild-type Cue1p, the assembly of ubiquitin chains in reactions with Cue1pRGA and lU7BR was delayed (Figure 5C). As shown above, sU7BR exhibited a dramatically reduced chain synthesis rate. Fluorescence inten- sity profiles revealed that not only the velocity of chain formation but also the ubiquitylation pattern differed in reactions with either wild-type Cue1p, Cue1pRGA, or lU7BR (Figure S6A).

The Cue1p CUE Domain Strongly Promotes Elongation of Diubiquitin

As analyzed by fluorescence measurement, ubiquitylation reac- tions in the presence of Cue1pRGA or lU7BR instead of wild-type Cue1p led to an increased formation of diubiquitin (Figure 5D).

However, in the reaction with Cue1p, the amount of triubiquitin, and even more, tetraubiquitin, exceeded the values measured for Cue1pRGA or lU7BR (Figure 5D), indicating that binding of wild-type Cue1p to di- and oligoubiquitin chains advances a higher efficiency of diubiquitin elongation.

As a consequence, reactions with Cue1p revealed a strongly accelerated hmw chain formation and synthesized larger amounts of hmw chains than did reactions with Cue1pRGA or lU7BR (Figures 6A and 6B). In contrast, fluorescence intensities measured in reactions with Cue1pRGA or lU7BR exhibited increased amounts of ubiquitin chains in a molecular weight range between 46 and 175 kDa. Compared to that, the reaction containing the sU7BR displayed an ubiquitin chain formation with only few chains reaching the hmw range (Figure 6B). These differences in the distribution of fluorescent ubiquitin conjugates between wild-type and mutant Cue1p are illustrated in more detail in profile overlays (Figures S6B–S6D).

Next we tested whether CUE domain and lU7BR can act as in- dependent units. Whereas Cue1p promoted hmw ubiquitin chain formation, CUE

was unable to compensate the reduced chain synthesis in the presence of lU7BR (Figure 6C). Hence, the CUE domain and Ubc7p-binding region have to be on the same molecule to enable efficient polyubiquitylation.

A Functional CUE Domain within Cue1p Facilitates the Degradation of Membrane-Bound ERAD Substrates

Having discovered a function of the CUE domain in vitro, we wondered if this UBD is important for turnover of ERAD sub- strates in vivo. The

alleles that were used for the expression in yeast cells are depicted in Figure 7A. Besides wild-type

and the CUE domain mutants

,

, and

, we generated the

mutant that is defective in Ubc7p recruitment but still contains the CUE domain (Figures 7A and 7E, left panel). At first, we examined the degradation of the membrane-bound Doa10p ligase substrate Ubc6p within 3 hr after cycloheximide treatment (Figure 7B). Ubc6p was rapidly degraded in

cells. Expression of either

or

resulted in a significant delay in Ubc6p proteoly- sis, and Ubc6p was even more stabilized in

or in

cells. The soluble ERAD model CPY* behaved differently.

Based on pulse-chase analysis (Figure 7C), CPY* was degraded equally well in the CUE domain mutants and in cells expressing wild-type

, whereas it was stabilized in

cells.

The differential effect of

mutants on the stability of the examined substrates suggested that the CUE domain could be required for the degradation of a specific subclass of substrates.

We therefore monitored the turnover of fusion proteins harboring the Deg1 degradation signal, which targets these constructs for ubiquitylation via the Doa10p ubiquitin ligase. Deg1-GFP-GFP (Lenk and Sommer, 2000) is a soluble, cytosolic construct, whereas Deg1-VP (Ravid et al., 2006) represents an integral membrane protein. As determined by a cycloheximide decay assay (Figure 7D), both fusion proteins were efficiently degraded within 60 min in

cells, and their turnover was blocked in

or

cells. Intriguingly, they behaved differently upon expression of the CUE domain mutants. In contrast to the cytosolic Deg1-GFP-GFP that was degraded in CUE domain mutants to the same extent as in wild-type

cells, mem- brane-bound Deg1-VP was slightly stabilized upon expression of

or

and in a more significant manner upon expression of

. Additionally, genetic evidence suggests that the short-lived mutant ER membrane protein Sec61-2 (Deshaies and Schekman, 1987; Biederer et al., 1996) was also stabilized in cells expressing

(Figures S7A and S7B). Moreover, we found that the degradation of a myc- tagged version of the Hmg2 protein via the Hrd1 ligase is affected in cells expressing CUE domain mutants (Figures S7C and S7D). We concluded from this that the Cue1p CUE domain might be specifically needed for degradation of a membrane- bound subset of ERAD substrates.

To confirm that the Cue1p proteins were correctly assembled into the E3 ligase complex, we immunoprecipitated 13xmyc- tagged Doa10p from solubilized membranes and examined the amounts of coprecipitated Cue1p and Ubc7p, and other ligase-associated proteins (Figure 7E, Table S2). Wild-type Cue1p and Cue1pRGA, expressed from plasmids, localized to the membrane fraction in amounts comparable to those in endogenous Cue1p. Increased amounts were detected for Cue1p

CUE and Cue1pLL/SS. Recruitment of Ubc7p to the membrane could be detected for all CUE domain mutants.

As expected, Ubc7p was hardly detectable in

cells or

upon expression of

(see also Metzger et al., 2013).

Figure 5. A Functional CUE Domain Increases the Efficiency of Polyubiquitin Chain Formation

(A and B) Analysis of ubiquitin chain formation by Ubc7p, Flag tagged ubiquitin, and various Cue1p constructs monitoring the formation of diubiquitin (lower panel, 18% SDS gel) and high molecular weight ubiquitin chains (upper panel, 9% SDS gel). Flag tagged ubiquitin conjugates were detected by anti Flag immunoblotting. (A) Twenty four minute time course showing polyubiquitylation reactions with Hrd1p in the presence of Cue1p or different CUE domain mutants.

The anti Cue1 blot shows the amount of added Cue1 proteins at time point t = 0. (B) In vitro ubiquitylation reactions with Cue1p or Cue1pRGA and in the absence or presence of either Hrd1p or Doa10p after 15 min incubation.

(C and D) Quantification of a typical in vitro ubiquitylation reaction of Ubc7p, Hrd1p, and different Cue1p variants with Alexa Fluor 488 labeled ubiquitin using ImageQuant software. SDS PAGE of samples taken at the indicated time points followed by measurement of fluorescence intensities. (C) Synthesis rates of all ubiquitin conjugations (ub2to ubn) were analyzed from 18% SDS gels. Time point t = 0 of each reaction was used for normalization of fluorescence intensities measured throughout the time course. (D) Diagrams illustrating the formation of di , tri , or tetraubiquitin in reactions with different Cue1p variants, each over a time period of 25 min. Fluorescence intensities (counts) were normalized by subtraction of values measured at time point t = 0 of the appropriate reaction.

Interestingly, although Ubc7p was precipitated with Doa10p in wild-type cells or the

mutant, the interaction of Ubc7p with the ligase complex of cells expressing

was reduced. This could be due to the shortened Cue1p

CUE protein and a very transient interaction of Ubc7p with the ligase in those cells. A generally impaired interaction was unlikely, since

mutants showed a normal degradation phenotype for both ERAD substrates CPY* and Deg1-GFP-GFP (Figures 7C and 7D). Furthermore, in

mutants, Cdc48p was still recruited to the ligase complex by Ubx2p, which implies that substrate ubiquitylation proceeds at a near-wild-type level.

DISCUSSION

The results presented here unravel a mechanism of ubiquitin chain formation at the ERAD RING ubiquitin ligases. In an in vitro ubiquitylation assay employing Ubc7p, Cue1p and the ligases Hrd1p or Doa10p ubiquitin chains assembled exclusively via lysine 48 of ubiquitin. Consistently, we found that this linkage specificity, mediated by Ubc7p, is required for degradation of the ERAD model substrate CPY* in vivo.

As previously published, the ligase-associated factor Cue1p is not only required for recruitment of Ubc7p to the ER membrane in vivo but also stimulates polyubiquitin chain formation in vitro (Bazirgan and Hampton, 2008; Kostova et al., 2009). A small C-terminal Ubc7p-binding region in Cue1p promotes a basic activation of Ubc7p-dependent ubiquitin conjugation. Structural analysis of the G2BR (Ube2g2-binding region) domain in human gp78, which is homologous to the U7BR in Cue1p, revealed a backside binding mechanism to the corresponding E2 enzyme Ube2g2, a homolog of Ubc7p (Das et al., 2009). This interaction causes a conformational change at the catalytically active center of Ube2g2 and promotes efficient discharge of ubiquitin, which, in consequence, accelerates chain formation (Das et al., 2009).

The enhanced activation of Ubc7p by an enlarged U7BR (lU7BR) is probably caused by a further stabilization of this back- side binding by adjacent residues.

In addition, other parts of the protein, especially the CUE domain, further enhance ubiquitin chain formation. This stimu- lating effect can be explained by the capacity of the CUE domain to bind ubiquitin. In line with a previous report (Shih et al., 2003), we found that the consensus CUE domain within Cue1p binds ubiquitin only weakly. In contrast to other CUE domains, which have a strong monoubiquitin binding activity, an extended CUE domain of Cue1p preferentially binds oligo- and polyubiquitin chains.

The amino acid sequence of the Cue1p CUE domain differs from the consensus motif in some conserved residues that were shown to be required for high-affinity binding of monoubi- quitin (Shih et al., 2003; Kang et al., 2003). Moreover, previous structural analysis revealed an involvement of lysine 48 of ubiq- uitin in contact formation with the CUE domain (Kang et al., 2003), which seemed to exclude polyubiquitylation via lysine 48 in this complex (Kang et al., 2003). Contrarily, our results

Weight Ubiquitin Chains

(A and B) Quantification of a typical in vitro ubiquitylation reaction of Ubc7p, Hrd1p, and different Cue1p variants with Alexa Fluor 488 labeled ubiquitin (see above). (A) Formation rates of ubiquitin chains with a molecular weight above 175 kDa were quantified from 9% SDS gels. Fluorescence intensity values of the different time points were normalized to time point t = 0 min of the appropriate reaction. (B) Distribution of ubiquitin chains of different molecular weight range in reactions with the indicated Cue1p constructs after 25 min incubation. Relevant molecular weight sizes were estimated by protein standards.

(C) Ubiquitin chain formation of Ubc7p and different indicated Cue1p con structs in the presence of Hrd1p and Flag tagged ubiquitin after 15 min incubation at 30C (upper panel, 9% SDS gel; lower panel, 18% SDS gel).

reveal that polyubiquitylation via lysine48 is strongly promoted by the CUE domain of Cue1p. These discrepancies suggest that the elongated interacting region of Cue1p may bind to ubiq- uitin in a different manner.

In vitro ubiquitylation reactions containing a Cue1p variant without functional CUE domain revealed a strong accumulation of diubiquitin and a clearly delayed formation of longer ubiquitin chains. The processivity of ubiquitin conjugation in general was only slightly affected. Taken together, this indicates early disso- ciation of short ubiquitin chains from the ligases. Remarkably,

once ubiquitin moieties were conjugated to chains of three or four molecules, further elongation of these chains seems to occur very quickly. There is a considerable number of hmw ubiq- uitin chains detected within a short incubation time, whereas chains with lower molecular weight are rare. Apparently, a functional Cue1p CUE domain binds to diubiquitin and longer ubiquitin conjugates that are assembled by the ERAD ligase complexes and stabilizes them in a way such that elongation of these chains can efficiently proceed. Dissociation of substrate proteins that were only conjugated with a diubiquitin moiety

(A) Graphical description of Cue1p proteins that were examined in in vivo studies. The corresponding genes were expressed from ARS/CEN plasmids in yeast strains.

(B) Turnover of Ubc6p in the indicatedcue1mutants within 3 hr after cycloheximide treatment of the yeast cells (left panel). The stable ER membrane bound protein Sec61p (right panel) served as loading control. Samples were analyzed by SDS PAGE and immunoblotting using anti Ubc6p and anti Sec61p antibodies.

(C) Quantification of CPY* turnover incue1mutants from three independent pulse chase experiments. Values measured for the indicated time points were normalized (time point 0 = 100%; error bars indicate standard deviation).

(D) Separate cycloheximide decay assays of the cytosolic Deg1 GFP GFP or the membrane anchored Deg1 Flag Vma12 ProtA fusion protein incue1mutants.

The remaining amount of each protein was analyzed 0, 20, 40, and 60 min after addition of cycloheximide to the cells. Proteins were detected by immunoblotting with anti GFP or anti Flag antibody.

(E) Immunoprecipitation of 13xmyc tagged Doa10p from yeast cells by anti myc antibody coupled to protein A Sepharose. Analysis of coprecipitated proteins by SDS PAGE and immunoblotting with protein specific antibodies (anti Ubx2, anti Cdc48;Neuber et al., 2005).

might impede efficient substrate degradation, since ubiquitin chains of four to six molecules in length were claimed to be required for proteasomal degradation of proteins (Thrower et al., 2000; Kang et al., 2007). This would explain the moderate albeit robust effect of mutations in the CUE domain on the turn- over of ERAD substrates.

Surprisingly, this effect seems to be limited to membrane- bound substrate proteins that probably exhibit defects within the membrane-spanning regions and are degraded via the ERAD-M pathway (Hmg2p, Sec61-2p) (Carvalho et al., 2006) or the Doa10p ligase (Ubc6p, membrane-anchored Deg1 fusion).

Presumably, ubiquitylation of these types of proteins makes special demands on the ERAD ligases. In contrast to soluble proteins, which may be targeted to the ERAD ligases during dislocation (Hrd1p) or at the cytosolic surface (Doa10p), mem- brane-bound proteins may be ubiquitylated in a ‘‘hit-and-run’’

mechanism which requires high processivity of the ligase. The binding of partially ubiquitylated substrates by Cue1p may assist in arresting such proteins at the ligases and thereby promote their efficient ubiquitylation.

Thus, our results do not only provide evidence for a functional relevance of ubiquitin chain binding by the Cue1p CUE domain, but also suggest a more general role for UBDs in the assembly of polyubiquitin chains by certain ligase complexes.

EXPERIMENTAL PROCEDURES Miscellaneous

We used standard media and protocols for manipulating yeast andE. colicells (Sambrook et al., 1989). Protocols for cloning and purification of plasmid DNA are as published (Ausubel et al., 1995). SDS PAGE and immunoblotting were done as described elsewhere (Laemmli, 1970).

Plasmids

The plasmids used in this study were generated by amplifying appropriate yeast genomic DNA via oligonucleotides containing selected restriction sites and cloning in the indicated vectors. All constructs were verified by sequencing. Plasmids used in this study are listed inTable S1.

Yeast Strains

Yeast strains were haploid descendants of the wild type strain DF5. Geno types of the yeast strains used in this study are summarized inTable S1. Yeast cells were transformed using the lithium acetate method (Gauss et al., 2005).

We used PCR based methods for C terminal epitope tagging of Doa10p (Longtine et al., 1998). Yeast strains used in this study are listed inTable S2.

Protein Purification

E. coliBL21 cells were transformed with the appropriate plasmids encoding GST fusion proteins and were precultivated in TB medium at 37C. After cool ing to 16C and induction of protein expression by addition of 0.5 mM IPTG, cells were grown for 15 20 hr at 16C. Cells were pelleted and lysed in PBS (containing protease inhibitor) by sonification. GST fusion proteins were precipitated using glutathione Sepharose (4 FastFlow or GSTrap, GE Health care). Removal of the GST tag and, thus, elution of the protein were performed by treatment with PrescissionProtease (GE Healthcare) overnight at 4C.

Proteins, except ubiquitin, were concentrated and stored at 80C. Ubiquitin or its mutant ubS20C was further purified by size exclusion chromatography (HiLoad 16/600 Superdex 75 pg, GE Healthcare). Purity of all generated pro teins was tested by SDS PAGE following Coomassie brilliant blue staining.

Site-Specific Labeling of Ubiquitin

For site specific labeling of the cystein in ubiquitin (ubS20C), Alexa Fluor 488 C5 Maleimide (Invitrogen) solved in dimethyl sulfoxide (DMSO) was utilized. To

reduce thiol groups of cysteine, 60 nmol of the ubiquitin variant was incubated with 120 nmol tris (2 carboxyethyl)phosphine (TCEP) in 400ml PBS (pH 7.5) for 10 min at 25C. The labeling preceded 90 min at 25C in the dark with a 4 fold excess of fluorescent dye. 10 mMbmercaptoethanol (bME) was added to stop the reaction. Desalting steps (Nap5 columns, GE Healthcare) into 20 mM HEPES/NaOH (pH 7.5) removed excess dye andbME. The eluate was concentrated (Amicon Ultra 0.5, 3 kDa cutoff, Millipore), and aliquots were stored at 80C.

In Vitro Ubiquitylation Reaction

Equal amounts (3.5mM) of purified Ubc7p, Cue1p variants, and Hrd1p or Doa10p fragments were mixed with 7.2mM (7.5mM) ubiquitin, 170 nM E1 enzyme (Uba1 from yeast, or His6Ube1 human recombinant, Boston Bio chem), 4 mM ATP (pH 8.0, Sigma), and 0.5 mM DTT in a 50 mM Tris HCl buffer (pH 7.5) containing 2.5 mM MgCl2on ice and subsequently incubated at 30C.

Components present in all reactions of the same experiment were prepared as master mix. In time course experiments, samples from time point t = 0 were taken before incubation at 30C. Flag epitope tagged ubiquitin as well as K R or K only ubiquitin variants were purchased from Boston Biochem. Ubiq uitin from bovine red blood cells was ordered from Sigma, and fluorescently labeled ubiquitin was generated as described below. Reactions were stopped by addition of Laemmli sample buffer. Monoubiquitin and ubiquitin conjugates were detected using anti ubiquitin antibody (polyclonal, ENZO Life Sciences) or anti Flag antibody (monoclonal, SIGMA). Ubc7 (Neuber et al., 2005), Cue1 (Biederer et al., 1997), Hrd1, and Doa10 were detected by polyclonal pro tein specific antibodies.

Quantification of Ubiquitin Chains by Mass Spectrometry

Ubiquitin chains were quantified as described inMirzaei et al. (2010). In short, heavy labeled peptides (Spiketides, JPT peptide technology) resembling the products of a tryptic digest of the ubiquitin branched peptides were spiked into the ubiquitination reactions. The reactions were subjected to a two step proteolytic digest with endopeptidase LysC and Trypsin under strong dena turing conditions (de Godoy et al., 2008). The peptides were separated on a 20 cm in house packed C18 column (75mm inner diameter, 3mm Reprosil, Dr. Maisch GmbH) using a 5% 50% acetonitrile gradient at a flow rate of 250 nl/min and sprayed directly into a Q TRAP 5500 mass spectrometer (AB Sciex). The signals for the different branched peptides were recorded in MRM mode and analyzed using the MultiQuant software package (AB Sciex).

Statistical analysis was performed using the R statistical language package (http://www.r project.org/).

Synthesis of Ubiquitin Chains for GST Pull-Down

Ubiquitin chains were generated in in vitro ubiquitylation reactions with either wild type ubiquitin, Alexa Fluor 488 labeled ubiquitin, or Flag ubiquitin as well as Ubc7p and Hrd1p. Reaction was stopped by addition of 6.5ml apyrase following incubation for 50 min at room temperature.

GST Pull-Down

Equal amounts of ubiquitin chains (z0.8 mg, K48 linked, untagged, Flag tagged, or fluorescently labeled) were incubated with 15 20ml (bed volume) Sepharose coupled GST fusion proteins in a 50 mM Tris binding buffer (pH 7.5) containing 150 mM NaCl and 1 mg/ml BSA for 3 hr at 4C on a rotator mixer. Proteins from the supernatant was precipitated by addition of 10% TCA, washed with 100% acetone, and resuspended in 40ml Laemmli sample buffer.

Sepharose beads were washed three times with binding buffer and then also treated with 40ml Laemmli sample buffer. Samples were analyzed by SDS PAGE following immunoblotting using anti ubiquitin antibody (monoclonal, Covance), anti Flag antibody (see above), anti GST antibody (polyclonal, Rockland) or by scanning the SDS gels with Typhoon FLA9500, excitation wavelength 473 nm, cutoff filter LBP510.

In Vitro Deubiquitylation Reaction

For the in vitro deubiquitylation reaction 7.8 mM K48 linked ubiquitin chains were mixed on ice with 30 mM Cue1p wild type, Cue1pRGA, or BSA in a 50 mM Tris HCl (pH 7.5), MgCl2buffer containing 0.5 mM DTT. An aliquot of each reaction was taken (time point t = 0). After addition of 30mM

His6 Otubain 1 (ENZO Life Sciences) and further incubation on ice, aliquots were taken at indicated time points, and reaction was stopped by immediate addition of urea sample buffer (Plemper et al., 1998), followed by incubation at 35C for 45 min.

Cycloheximide Decay Assay

In the cycloheximide decay experiments, 20 OD600 log phase cells were harvested and resuspended in 4.5 ml prewarmed SD medium containing 100mg/ml cycloheximide (‘‘Actidione,’’ Sigma). Aliquots were taken at the indicated times, degradation was stopped by the addition of NaN3, and total cell lysates or membrane fractions were prepared. Proteins of interest were analyzed by SDS PAGE and immunoblotting using anti Sec61p or anti Ubc6 antibody (Biederer et al., 1997;Walter et al., 2001), anti CPY* antibody (mono clonal, Invitrogen), anti Flag antibody (monoclonal, SIGMA), or anti GFP anti body (polyclonal, Invitrogen).

Pulse-Chase Experiments

Yeast cells were grown at 30C in SD media with the appropriate supplements to 0.8 OD600/ml. Then 18 OD600growing cells were harvested; resuspended in prewarmed, fresh SD medium; and labeled with 80mCi/10 OD600of a mixture of [35S] methionine and cysteine (Perkin Elmer) for 8 min. Cells were sedi mented and resuspended in prewarmed, fresh SD medium supplemented with 0.004% methionine and 0.003% cysteine and the required amino acids.

At indicated time points, aliquots (z6 OD600) were taken, and cellular pro cesses were stopped by addition of 10 mM ice cold NaN3. Cells were pelleted and resuspended in 50 mM Tris HCl (pH 7.5), 1% SDS and disrupted by mixing with glass beads for 3 min. After dilution to 1% Triton, 0.1% SDS 50 mM Tris (pH 7.5), 150 mM NaCl, 0.5 mM EDTA, CPY* was precipitated via a protein specific antibody (monoclonal anti CPY, Invitrogen) coupled to protein A Sepharose. Beads were first washed with buffer containing 1% Triton, 0.1%

SDS 50 mM Tris (pH 7.5), 150 mM NaCl, 0.5 mM EDTA and then with an equiv alent buffer containing 0.02% SDS, followed by incubation with N glycosidase F (Roche) in a buffer withbmercaptoethanol. Samples were loaded onto SDS gels and analyzed by phosphorimaging.

Immunoprecipitation

For immunoprecipitation of Doa10p 13xmyc 50 OD600, exponentially growing yeast cells were harvested, washed with water, and then resuspended in buffer containing 50 mM Tris HCl (pH 7.5), 200 mM NaCl, 1 mM EDTA, 10%

glycerol, and protease inhibitors. Cells were lysed by strongly mixing them with glass beads. After dilution, using the same buffer, preparation of micro somes was performed by centrifugation. The membrane pellet was resus pended in 1% digitonin and solubilized during 1 hr incubation at 4C on a rotator mixer. Remaining membrane fragments were sedimented by centrifu gation. Aliquots of the supernatant were taken as input sample and treated with urea sample buffer (Plemper et al., 1998). The bulk of the supernatant was diluted to 0.5% digitonin and incubated with monoclonal anti myc antibody (Sigma) immobilized to protein A Sepharose (GE Healthcare). After overnight incubation at 4C, Sepharose resin was washed, and proteins were eluted in urea sample buffer.

ACKNOWLEDGMENTS

The authors wish to thank the members of their research groups for discus sions, unpublished data, and materials. The Deutsche Forschungsgemein schaft generously supports research in the laboratory of T.S. (SFB 740, Priority Program 1365). T.S., A.C., and M.H.G. receive joined funding from the German Israel Project Cooperation DIP. A.C. acknowledges the support of the AMRF (Adelson Medical Research Foundation) and an ICRF (Israel Cancer Research Fund) Professorship.

REFERENCES

Ausubel, F.M., Brent, R., Kingston, R.E., Moore, D.D., Seidman, J.G., Smith, J.A., and Struhl, K. (1995). Current Protocols in Molecular Biology (New York: John Wiley & Sons).

Bays, N.W., Gardner, R.G., Seelig, L.P., Joazeiro, C.A., and Hampton, R.Y.

(2001). Hrd1p/Der3p is a membrane anchored ubiquitin ligase required for ER associated degradation. Nat. Cell Biol.3, 24 29.

Bazirgan, O.A., and Hampton, R.Y. (2008). Cue1p is an activator of Ubc7p E2 activity in vitro and in vivo. J. Biol. Chem.283, 12797 12810.

Biederer, T., Volkwein, C., and Sommer, T. (1996). Degradation of subunits of the Sec61p complex, an integral component of the ER membrane, by the ubiq uitin proteasome pathway. EMBO J.15, 2069 2076.

Biederer, T., Volkwein, C., and Sommer, T. (1997). Role of Cue1p in ubiquitina tion and degradation at the ER surface. Science278, 1806 1809.

Bordallo, J., and Wolf, D.H. (1999). A RING H2 finger motif is essential for the function of Der3/Hrd1 in endoplasmic reticulum associated protein degrada tion in the yeast Saccharomyces cerevisiae. FEBS Lett.448, 244 248.

Carvalho, P., Goder, V., and Rapoport, T.A. (2006). Distinct ubiquitin ligase complexes define convergent pathways for the degradation of ER proteins.

Cell126, 361 373.

Chen, B., Mariano, J., Tsai, Y.C., Chan, A.H., Cohen, M., and Weissman, A.M.

(2006). The activity of a human endoplasmic reticulum associated degradation E3, gp78, requires its Cue domain, RING finger, and an E2 binding site. Proc.

Natl. Acad. Sci. USA103, 341 346.

Das, R., Mariano, J., Tsai, Y.C., Kalathur, R.C., Kostova, Z., Li, J., Tarasov, S.G., McFeeters, R.L., Altieri, A.S., Ji, X., et al. (2009). Allosteric activation of E2 RING finger mediated ubiquitylation by a structurally defined specific E2 binding region of gp78. Mol. Cell34, 674 685.

de Godoy, L.M., Olsen, J.V., Cox, J., Nielsen, M.L., Hubner, N.C., Fro¨hlich, F., Walther, T.C., and Mann, M. (2008). Comprehensive mass spectrometry based proteome quantification of haploid versus diploid yeast. Nature455, 1251 1254.

Deshaies, R.J., and Schekman, R. (1987). A yeast mutant defective at an early stage in import of secretory protein precursors into the endoplasmic reticulum.

J. Cell Biol.105, 633 645.

Dikic, I., and Do¨tsch, V. (2009). Ubiquitin linkages make a difference. Nat.

Struct. Mol. Biol.16, 1209 1210.

Dikic, I., Wakatsuki, S., and Walters, K.J. (2009). Ubiquitin binding domains from structures to functions. Nat. Rev. Mol. Cell Biol.10, 659 671.

Finger, A., Knop, M., and Wolf, D.H. (1993). Analysis of two mutated vacuolar proteins reveals a degradation pathway in the endoplasmic reticulum or a related compartment of yeast. Eur. J. Biochem.218, 565 574.

Finley, D., Sadis, S., Monia, B.P., Boucher, P., Ecker, D.J., Crooke, S.T., and Chau, V. (1994). Inhibition of proteolysis and cell cycle progression in a multi ubiquitination deficient yeast mutant. Mol. Cell. Biol.14, 5501 5509.

Gauss, R., Trautwein, M., Sommer, T., and Spang, A. (2005). New modules for the repeated internal and N terminal epitope tagging of genes in Saccharomyces cerevisiae. Yeast22, 1 12.

Haglund, K., and Dikic, I. (2005). Ubiquitylation and cell signaling. EMBO J.24, 3353 3359.

Hershko, A., and Ciechanover, A. (1998). The ubiquitin system. Annu. Rev.

Biochem.67, 425 479.

Hicke, L., Schubert, H.L., and Hill, C.P. (2005). Ubiquitin binding domains. Nat.

Rev. Mol. Cell Biol.6, 610 621.

Hirsch, C., Gauss, R., Horn, S.C., Neuber, O., and Sommer, T. (2009). The ubiquitylation machinery of the endoplasmic reticulum. Nature458, 453 460.

Hoeller, D., Hecker, C.M., Wagner, S., Rogov, V., Do¨tsch, V., and Dikic, I.

(2007). E3 independent monoubiquitination of ubiquitin binding proteins.

Mol. Cell26, 891 898.

Hurley, J.H., Lee, S., and Prag, G. (2006). Ubiquitin binding domains.

Biochem. J.399, 361 372.

Husnjak, K., and Dikic, I. (2012). Ubiquitin binding proteins: decoders of ubiq uitin mediated cellular functions. Annu. Rev. Biochem.81, 291 322.

Kang, R.S., Daniels, C.M., Francis, S.A., Shih, S.C., Salerno, W.J., Hicke, L., and Radhakrishnan, I. (2003). Solution structure of a CUE ubiquitin complex reveals a conserved mode of ubiquitin binding. Cell113, 621 630.

Kang, Y., Chen, X., Lary, J.W., Cole, J.L., and Walters, K.J. (2007). Defining how ubiquitin receptors hHR23a and S5a bind polyubiquitin. J. Mol. Biol.

369, 168 176.

Kostova, Z., Mariano, J., Scholz, S., Koenig, C., and Weissman, A.M. (2009). A Ubc7p binding domain in Cue1p activates ER associated protein degrada tion. J. Cell Sci.122, 1374 1381.

Laemmli, U.K. (1970). Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature227, 680 685.

Lenk, U., and Sommer, T. (2000). Ubiquitin mediated proteolysis of a short lived regulatory protein depends on its cellular localization. J. Biol. Chem.

275, 39403 39410.

Longtine, M.S., McKenzie, A., 3rd, Demarini, D.J., Shah, N.G., Wach, A., Brachat, A., Philippsen, P., and Pringle, J.R. (1998). Additional modules for versatile and economical PCR based gene deletion and modification in Saccharomyces cerevisiae. Yeast14, 953 961.

Madura, K. (2002). The ubiquitin associated (UBA) domain: on the path from prudence to prurience. Cell Cycle1, 235 244.

Metzger, M.B., Liang, Y. H., Das, R., Mariano, J., Li, S., Li, J., Kostova, Z., Byrd, R.A., Ji, X., and Weissman, A.M. (2013). A structurally unique E2 binding domain activates ubiquitination by the ERAD E2, Ubc7p, through multiple mechanisms. Mol. Cell50. Published online May 9, 2013.http://dx.doi.org/

10.1016/j.molcel.2013.04.004.

Mirzaei, H., Rogers, R.S., Grimes, B., Eng, J., Aderem, A., and Aebersold, R.

(2010). Characterizing the connectivity of poly ubiquitin chains by selected reaction monitoring mass spectrometry. Mol. Biosyst.6, 2004 2014.

Neuber, O., Jarosch, E., Volkwein, C., Walter, J., and Sommer, T. (2005). Ubx2 links the Cdc48 complex to ER associated protein degradation. Nat. Cell Biol.

7, 993 998.

Plemper, R.K., Egner, R., Kuchler, K., and Wolf, D.H. (1998). Endoplasmic reticulum degradation of a mutated ATP binding cassette transporter Pdr5

proceeds in a concerted action of Sec61 and the proteasome. J. Biol.

Chem.273, 32848 32856.

Polo, S., Sigismund, S., Faretta, M., Guidi, M., Capua, M.R., Bossi, G., Chen, H., De Camilli, P., and Di Fiore, P.P. (2002). A single motif responsible for ubiq uitin recognition and monoubiquitination in endocytic proteins. Nature416, 451 455.

Ponting, C.P. (2000). Proteins of the endoplasmic reticulum associated degradation pathway: domain detection and function prediction. Biochem.

J.351, 527 535.

Raasi, S., Varadan, R., Fushman, D., and Pickart, C.M. (2005). Diverse polyu biquitin interaction properties of ubiquitin associated domains. Nat. Struct.

Mol. Biol.12, 708 714.

Randles, L., and Walters, K.J. (2012). Ubiquitin and its binding domains. Front.

Biosci.17, 2140 2157.

Ravid, T., Kreft, S.G., and Hochstrasser, M. (2006). Membrane and soluble substrates of the Doa10 ubiquitin ligase are degraded by distinct pathways.

EMBO J.25, 533 543.

Sambrook, J., Fritsch, E.F., and Maniatis, T. (1989). Molecular Cloning: A Laboratory Manual, 2nd edition (New York: Cold Spring Harbor Laboratory Press).

Searle, M.S., Garner, T.P., Strachan, J., Long, J., Adlington, J., Cavey, J.R., Shaw, B., and Layfield, R. (2012). Structural insights into specificity and diver sity in mechanisms of ubiquitin recognition by ubiquitin binding domains.

Biochem. Soc. Trans.40, 404 408.

Shih, S.C., Prag, G., Francis, S.A., Sutanto, M.A., Hurley, J.H., and Hicke, L.

(2003). A ubiquitin binding motif required for intramolecular monoubiquityla tion, the CUE domain. EMBO J.22, 1273 1281.

Smith, M.H., Ploegh, H.L., and Weissman, J.S. (2011). Road to ruin: targeting proteins for degradation in the endoplasmic reticulum. Science334, 1086 1090.

Thrower, J.S., Hoffman, L., Rechsteiner, M., and Pickart, C.M. (2000).

Recognition of the polyubiquitin proteolytic signal. EMBO J.19, 94 102.

Walter, J., Urban, J., Volkwein, C., and Sommer, T. (2001). Sec61p indepen dent degradation of the tail anchored ER membrane protein Ubc6p. EMBO J.20, 3124 3131.

Xu, P., Duong, D.M., Seyfried, N.T., Cheng, D., Xie, Y., Robert, J., Rush, J., Hochstrasser, M., Finley, D., and Peng, J. (2009). Quantitative proteomics reveals the function of unconventional ubiquitin chains in proteasomal degra dation. Cell137, 133 145.