Detecting and Measuring Cotranslational Protein Degradation in Vivo

See allHide authors and affiliations

Science  22 Sep 2000:
Vol. 289, Issue 5487, pp. 2117-2120
DOI: 10.1126/science.289.5487.2117


Nascent polypeptides emerging from the ribosome and not yet folded may at least transiently present degradation signals similar to those recognized by the ubiquitin system in misfolded proteins. The ubiquitin sandwich technique was used to detect and measure cotranslational protein degradation in living cells. More than 50 percent of nascent protein molecules bearing an amino-terminal degradation signal can be degraded cotranslationally, never reaching their mature size before their destruction by processive proteolysis. Thus, the folding of nascent proteins, including abnormal ones, may be in kinetic competition with pathways that target these proteins for degradation cotranslationally.

Nascent polypeptides emerging from the ribosome may, in the process of folding, present hydrophobic patches and other structural features that serve as degradation signals similar to those recognized by the ubiquitin (Ub) system in misfolded or otherwise damaged proteins (1). Whether a substantial fraction of nascent polypeptides is cotranslationally degraded is a long-standing question.

The Ub sandwich technique was developed to detect cotranslational protein degradation by measuring the steady-state ratio of two reporter proteins whose relative abundance is established cotranslationally. The technique requires that the polypeptide to be examined for cotranslational degradation, termed B, be sandwiched between two stable reporter domains, A and C, in a linear fusion protein (Fig. 1A). The three polypeptides are connected by Ub moieties, creating anAUb-BUb-CUb fusion protein. Ub-specific processing proteases (UBPs) cotranslationally cleave such linear Ub fusions at the C-terminal residue of Ub (2–4), generating three independent polypeptides, AUb, BUb, and CUb (5). UBP-mediated cleavage establishes a kinetic competition between two mutually exclusive events during the synthesis ofAUb-BUb-CUb: cotranslational UBP cleavage at the BUb-CUb junction to release the long-lived CUb module or, alternatively, cotranslational degradation of the entire BUb-CUb nascent chain by the proteasome (6) (Fig. 1B). In the latter case, the processivity of proteasome-mediated degradation results in the destruction of the Ub moiety between B and Cbefore it can be recognized by UBPs. The resulting drop in levels of the CUb module relative to levels of AUb, referred to as the C/A ratio, reflects the cotranslational degradation of domain B (Fig. 1B).

Figure 1

The ubiquitin sandwich technique. (A) Organization of a Ub sandwich fusion. The polypeptide assayed for cotranslational degradation, B, is sandwiched between two stable reporter domains A and C. Red arrows indicate the locations of UBP cleavage sites. (B) The principle of the method. The reporter module AUb is the first synthesized and is cotranslationally released from B, thereby providing a measure of the number of nascent Bchains that initially emerge from the ribosome. If degradation of the emerging B domain, indicated by its insertion into the proteasome, is strictly posttranslational, UBP-mediated cleavage at theBUb-C junction releases CUb beforeB is degraded, so the molar yields of CUb andAUb are identical. However, if degradation of Bcan be cotranslational, a substantial fraction ofBUb-CUb may be degraded as a unit. This will result in the molar yield of CUb being lower thanAUb, the difference being a measure of cotranslational degradation.

To verify that UBP-mediated cleavage is cotranslational (3), we carried out in vivo radiolabeling in which the labeling pulse was substantially shorter than the time required for the complete synthesis of AUb-BUb-CUb. Saccharomyces cerevisiae cells expressing the fusion protein {DHFRhaUb} - {MeKβgalUb} - {MeKDHFRhaUb} (Fig. 2A), predicted to require ∼350 s for complete synthesis, were radiolabeled for 45 s (7). Labeling was terminated by addition of cycloheximide, and UBPs were simultaneously inactivated withN-ethylmaleimide (NEM) (3). Under these conditions, nascent chains that are just starting to be synthesized when the pulse begins will incorporate label into the N-terminalA domain but do not elongate to full-length chains (Fig. 2B). Thus, detection of free, labeled AUb and BUb by immunoprecipitation indicated that UBP-mediated cleavage at theAUb-BUb junction was cotranslational (Fig. 2C). No full-length AUb-BUb-CUb fusion was detected (Fig. 2C), indicating that cotranslational cleavage by UBPs was highly efficient.

Figure 2

UBP-mediated cleavage of ubiquitin sandwich fusions is cotranslational. (A) The protein fusions used. Domains A and C are mouse dihydrofolate reductase tagged with the influenza hemagglutinin-derived ha epitope (DHFRha). Domain C carries an N-terminal extension (eK, see text), which makes it electrophoretically distinguishable from A. The different B domains are E. coli βGal, Sindbis virus RNA polymerase (nsP4), andS. cerevisiae Ura3p. Unstable versions of these domains have an N-terminal arginine (R) residue; stable versions have methionine (M). (B) The population of nascent chains produced by a radiolabeling pulse that is shorter than the time of translation of an AUb-BUb-CUb fusion. Stretches of polypeptide containing radiolabel are in red; unlabeled stretches are in black. If UBPs efficiently cleave the nascent chain, free radiolabeled AUb, BUb, and CUb should all be detected. If UBPs can cleave solely the full-length mature protein, only labeled CUb will be observed. (C) The UBP cleavage of Ub sandwich fusions is cotranslational. The release of AUb, BUb, and CUb by UBP cleavage was assayed by immunoprecipitation (7). CUb* denotes both the MeKDHFRhaUb band and a cross-reacting band present in these NEM-treated extracts but not in untreated ones (compare with Fig. 3A). M r, relative molecular mass.

Previous experiments bearing on cotranslational degradation used inhibitors or cell-free systems (8–10). Nascent polypeptide chains might be protected from degradation in vivo, either because they are sterically shielded by chaperones or because their translation time is short compared with the time required for targeting by the degradation machinery. The Ub sandwich technique was used to detect in vivo cotranslational degradation of a 118-kD, β-galactosidase (βgal)–derived polypeptide carrying a strong N-terminal degradation signal, specifically an N-degron (Fig. 2A). The βgal-linked N-degron comprises the destabilizing N-terminal residue arginine (R) and a short, lysine (K)–bearing extension, eK(Fig. 2A), that is the site of multi-Ub chain attachment (11). Ubr1p, the E3 component of the N-end rule pathway, targets ReK-βgal for rapid degradation in vivo [half-life (t 1/2) ∼ 2 min] (2, 12). Changing the N-terminal residue of the protein to methionine (M) inactivates the degradation signal by precluding recognition by Ubr1p. The resulting MeK–βgal is posttranslationally stable (t 1/2 > 20 hours) (2). AUb-BUb-CUb fusion proteins in which domain B was either ReK-βgal or its N- degron–lacking counterpart MeK-βgal (Fig. 2A) were expressed in S. cerevisiae strains containing different levels of Ubr1p (13). The extent of cotranslational degradation was assessed by radiolabeling for 30 min and immunoprecipitation (14) to determine the levels of CUb relative to AUb (the C/A ratio).

The C/A ratio was lower in cells expressing the N-degron–bearing domain B, but only in those strains that also expressed Ubr1p (Fig. 3A). To determine the percentage of nascent chains cotranslationally degraded by the N-end rule pathway, we compared the C/A ratios for ReK-βgal in wild-type (0.52) and Ubr1p-overexpressing strains (0.39) with the ratio found with the ubr1Δ strain (0.86) (Fig. 4A). This comparison indicated that ∼40% of the nascent ReK-βgal chains were cotranslationally degraded in the wild-type strain (Fig. 4B) (15). This fraction increased to ∼55% when Ubr1p was overexpressed from the PGAL1 promoter.

Figure 3

Nascent poly-peptides bearing an N-terminal degradation signal can be degraded cotranslationally. (A) Determination of C/A ratios through immunoprecipitation of in vivo–labeled AUb-BUb-CUb fusion proteins (14). Two variants of βgal were used as domain B, one carrying an N-degron (ReK-βgal; t 1/2 ∼ 2 min) and one lacking this degradation signal (MeK-βgal;t 1/2 > 20 hours). Each pair of lanes corresponds to two independent experiments. (B) Immunoblot analysis of the C/A levels for different B domains (described in Fig. 2A). The identity of the N-terminal residue of each of the B domains is indicated above each lane. F marks the lane in which domain B was the 18-residue FLAG-containing moiety.

Figure 4

The extent of cotranslational protein degradation depends on the presence of a degron, the activity of a degron-specific proteolytic pathway, and the nature and size of the protein. (A) C/A ratios obtained for the differentB domains. (B) Ubr1p-dependent cotranslational degradation (15). Each bar represents a mean value derived from at least four independent experiments; standard errors are indicated. Differences between means were significant to at leastP < 0.05 by the Mann-Whitney test. Bars are as in (A).

The extent of cotranslational degradation of two otherB domains of different sizes was also determined. The mammalian Sindbis virus RNA polymerase, termed nsP4, is a 69-kD protein that naturally bears an N-degron (16), and XeK-Ura3p (X = M or R) is a 34-kD enzyme of theS. cerevisiae uracil biosynthetic pathway that either carries (ReK-Ura3p) or lacks (MeK-Ura3p) an N-degron (17) (Fig. 2A). Although the short-lived R-nsP4 (69 kD) and ReK-βgal (118 kD) were cotranslationally degraded to similar extents, ReK-Ura3p (34 kD), which was also short-lived posttranslationally (17), exhibited much less cotranslational degradation than the other two proteins (Fig. 3B). Radiolabeling and immunoprecipitation (14) showed that ∼50% of the nascent chains of R-nsP4 were cotranslationally degraded by the N-end rule pathway in wild-type cells and ∼55% in cells overexpressing Ubr1p (Fig. 4B). These values were similar to ∼40% and ∼55% cotranslational degradation of ReK-βgal in these strains, respectively (Fig. 4B). In contrast, only ∼20% cotranslational degradation was observed with the 34-kD ReK-Ura3p in either wild-type or Ubr1p-overexpressing strains (Fig. 4B). These results suggest that smaller proteins are less susceptible to cotranslational degradation; however, factors other than translation time likely influence the presentation or accessibility of the degradation signal by the nascent chain, because protein size was not directly proportional to the extent of cotranslational degradation.

These comparisons established the amounts of cotranslational degradation by the N-end rule pathway. The C/A ratios for two of the N-degron–lacking B domains, MeK-βgal (0.84) and M-nsP4 (0.72), were less than the C/A ratio obtained with a very short FLAG epitope–containing sequence that lacks a degradation signal and exhibits a C/A ratio of 0.97, indistinguishable from the theoretical value of 1.0 (Fig. 4A). This suggests that ∼15% of MeK-βgal and ∼25% of M-nsP4 nascent chains might be cotranslationally degraded by a Ubr1p-independent pathway. Premature termination of translation and/or transcription may also contribute to the drop in the C/A ratios for these B domains. However, it is unlikely that premature termination of translation accounts entirely for the difference between the C/A ratios for FLAG and M-nsP4, because the codon adaptation index of the M-nsP4 open reading frame (0.1) is higher than the one for MeK-βgal (0.07), but the drop in C/A ratio is greater with M-nsP4 than with MeK-βgal. The degradation of newly synthesized polypeptides is a major source of peptides presented to the immune system (18, 19). Our observations with the above two B domains suggest that the source of these peptides is cotranslational protein degradation. In this regard, it is interesting that nsP4 is a viral protein that is presented efficiently to the immune system by infected cells.

The extent of cotranslational degradation can be strikingly high: In the case of ReK-βgal in Ubr1p-overexpressing cells, over 50% of nascent polypeptide chains never reach their full size before their destruction by processive proteolysis. Thus, if a nascent chain displays a degron of the Ub system, such a protein becomes a target of kinetic competition between cotranslational biogenesis and cotranslational degradation. Because the folding of a protein molecule begins during its synthesis on the ribosome, a nascent polypeptide may cotranslationally expose degradation signals that become shielded through the folding of the newly formed protein (20,21). Thus, cotranslational protein degradation may represent a form of protein quality control that destroys nascent chains that fail to fold correctly rapidly enough.

  • * To whom correspondence should be addressed. E-mail: avarsh{at}


View Abstract

Navigate This Article