Development of β-Hairpin Peptides for the Measurement of SCF-Family E3 Ligase Activity in Vitro via Ornithine Ubiquitination.

Regulation of the ubiquitin-proteasome system (UPS) to treat select types of cancer has become a popular area of drug discovery research. The FDA approval of proteasome inhibitors Bortezomib and Carfilzomib in the treatment of multiple myeloma has led to an increased need for chemical reporters capable of detecting and quantifying protein ubiquitination and the activity of members of the UPS including E3 ubiquitin ligases and the proteasome in the tumor cells of the patients. One limitation of peptide-based reporters is their rapid degradation in the cellular environment by cytosolic peptidases. Conversely, β-hairpin "protectides" exhibit a pronounced secondary structure that significantly increases their lifetime under cellular conditions. The goal of this work was to develop a family of novel, ornithine-rich protectides that could act as primary degrons serving as substrates for in vitro ubiquitination. The fluorescent peptide-based reporters were demonstrated to be highly resistant to degradation in multiple myeloma cell lysates. The most stable β-hairpin primary degron, containing a single ornithine residue at the N-terminus, OWRWR [Ac-OWVRVpGO(FAM)WIRQ-NH2], demonstrated rapid ubiquitination kinetics and a 20-fold increase in stability when compared with an unstructured primary degron. A screen of E1 and E3 enzyme inhibitors in cell lysates showed that ubiquitination of OWRWR was significantly impaired by inhibitors of the SCF family of E3 ligases. Furthermore, this is the first report demonstrating the use of an ornithine residue on a primary degron as a ubiquitination site. This study serves as a strong foundation for the development of stable, fluorescent, peptide-based reporters capable of quantifying protein ubiquitination and the enzymatic activity of members of the UPS.

Introduction

Regulation of cellular homeostasis, migration, and division is accomplished via a continuous cycle of protein expression, posttranslational modification, and degradation.[1] Degradation of misfolded or inactivated proteins is primarily accomplished by proteasome-mediated degradation, which requires the posttranslational addition of a polyubiquitin chain to a target protein. Ubiquitin is a small 76 amino acid globular protein whose attachment and removal from target proteins is governed by a large family of enzymes collectively referred to as the ubiquitin–proteasome system (UPS).[2] Polyubiquitination is the result of a cascade of enzymes culminating in the formation of an isopeptide bond between the C-terminal carboxylate of an incoming ubiquitin and the ε-amino group of a lysine residue on the target protein; however, ubiquitin has been observed to be conjugated to other amino acids including cysteine, serine, or threonine.[3] Protein ubiquitination starts with an E1 ubiquitin-activating enzyme forming a high energy thioester bond with free ubiquitin, which is recognized and transferred to an E2 ubiquitin-conjugating enzyme. Next, an E3 ubiquitin ligase forms a complex with the E2 enzyme to mediate the transfer of ubiquitin to the target protein. The E3 ligase recognizes and binds to a specific amino acid degradation sequence or degron. These degrons, normally in close proximity to an ubiquitin-accepting lysine residue, impart specificity to protein degradation because each E3 binds to a subset of degrons. The large number of E3 ligases (>600 in humans) permits recognition of a wide variety of degrons including phosphodegrons, oxygen-dependent degrons, and N-terminal degrons.[4] Although it is well-established that degron-mediated ubiquitination of proteins leads to proteasomal degradation, the role of protein monoubiquitination versus polyubiquitination is still unclear. Recent work by Shabek et al. demonstrated that proteins consisting of ∼150 residues required only a single ubiquitin moiety to target them to the proteasome for degradation.[5] Although initially the term degron was applied to a sequence that was responsible for both ubiquitination and proteasome-mediated degradation, a recent study by Guharoy et al.[6] expanded this definition of a degron to include three essential components for ubiquitin-dependent protein degradation by the proteasome. The primary degron is a specific sequence of amino acids recognized by an E3 ubiquitin ligase, the secondary degron is a proximal ubiquitination site lysine, and the tertiary degron is an unstructured sequence at least 20–30 amino acids long and is responsible for the initiation of proteasomal degradation. All three components are necessary for proteasome-mediated degradation; however, a primary degron with a ubiquitination site offers the potential to serve as a reporter for ubiquitination and E3 ligase activity alone.

Regulating proteasome expression and function has become a popular target of drug discovery research in the treatment of cancer, neurodegenerative disease, and cardiac disease.[7] The success of the proteasome inhibitors Bortezomib and Carfilzomib in the treatment of patients suffering from multiple myeloma has inspired research and exploration of the UPS that could one day provide as many “druggable” targets as the kinase/phosphatase proteome.[8] Whereas the throughput of inhibitor discovery for members of the UPS has increased dramatically, the development of novel methods to quantify the enzyme activity in the UPS has lagged behind. The traditional methods for measuring the enzyme activity include western blotting and ELISA, both of which rely on the analysis of bulk cell lysates. Although effective for preliminary discovery, all heterogeneity amongst cells is lost because these methods report a population-averaged result. Thus valuable information about distinct subpopulations, such as clonal subgroups, is unavailable. To overcome this limitation, fluorescent peptide-based reporters have been coupled with analytical methods such as capillary electrophoresis (CE) and microfluidic separation to assay single cells for intracellular enzyme activity.[9] Peptide-based reporters can be rapidly synthesized and easily modified and do not require complex manipulations for the incorporation into single, living cells. Previous work has utilized synthesized peptides to elucidate the underlying biological machinery of the UPS[5,10] and to monitor kinase and phosphatase activity in intact single cells using CE.[11] One limitation of peptide-based reporters is their rapid degradation by intracellular peptidases and proteases. Designing fluorescent peptide-based reporters that report the activity of the UPS by the detection of proteasome-mediated reporter degradation[2] can be challenging. Often, a significant amount of the reporter is degraded by cytosolic proteases before the activity of the UPS can be detected. Previous work by Melvin et al.[12] demonstrated that degron-based peptide substrates were readily degraded in cytosolic lysates, suggesting a need for modifications to make the substrates more stable and resistant to peptidase-mediated degradation to function as an accurate reporter. Recent work demonstrated the utility of protecting motifs by adding the secondary structure to increase the lifetime of the peptide-based reporters in cells.[13] As such, it was hypothesized that a β-hairpin “protectide” conjugated on the N-terminus of an unstructured primary degron could produce a viable fluorescent peptide-based reporter of peptide ubiquitination or proteasome-mediated degradation.

This paper details the discovery and characterization of a family of ornithine-rich, structured, primary degrons that are rapidly ubiquitinated and highly resistant to degradation by intracellular peptidases. The β-hairpin peptide OWOWO [Ac-OWVOVpGO(FAM)WIOQ-NH2] was initially thought to act solely as a “protectide” for a previously established primary degron; however, the protectide itself was discovered to be rapidly ubiquitinated in cell lysates. This finding was novel as ubiquitin conjugation to an ornithine residue had not previously been demonstrated in the literature. The secondary structure of the peptide was verified using circular dichroism (CD) and its degradation resistance in cell lysates was determined using CE. Analysis of the ubiquitination kinetics and identification of the optimal ubiquitination-site ornithine on the OWOWO peptide was performed. Finally, preliminary studies were performed in the presence of E1 and E3 enzyme inhibitors to identify enzyme targets for which the ornithine-rich, β-hairpin primary degron could serve as a potential reporter. The work presented here is an exciting first step toward the development of the next generation of cell-compatible, fluorescent, peptide-based ubiquitination reporters capable of directly measuring E3 ligase and proteasome activity.

Results and Discussion

Ornithine-Rich β-Hairpins Function as Primary Degrons Capable of in Vitro Ubiquitination

A well-folded β-hairpin “protectide” was added to the N-terminus of a previously isolated portable degron developed by Bonger et al.[14] to determine if the protecting motif could increase the stability of the degron while not impeding its ability to be rapidly ubiquitinated. Recently, Melvin et al.[12] have confirmed that this degron could be incorporated into a peptide-based substrate that functioned as a primary degron; however, the rate of ubiquitination was similar to the rate of peptidase-mediated degradation, limiting its use as a ubiquitination reporter. The protectide selected to confer stability, named OWOWO (Figure A), was a modification of a β-hairpin sequence previously characterized by Cline et al.[13b] and utilized by Yang et al.[13a] to increase peptide lifetime in a peptidase solution. The primary degron (RRRG) was demonstrated by Melvin et al. to be rapidly ubiquitinated in HeLa cell lysates when a lysine residue was positioned at the C-terminus.[12] The protectide-conjugated degron (OWOWO-RRRG, Figure A) was ubiquitinated in a time-dependent manner (Figure B, lanes 1–2). Interestingly, the peptide appeared to be multimonoubiquitinated as demonstrated by the presence of several higher molecular weight bands in a sample incubated with methylated ubiquitin (MeUb) (Figure B, lane 1), a modified ubiquitin in which all lysine residues were methylated to prevent polyubiquitin chain formation. Because the peptide substrate contained a single lysine residue at the C-terminus, it was suspected that the ornithine residues located on the OWOWO protectide at positions 1, 4, and 11 were acting as additional ubiquitination sites. The OWOWO peptide alone was incubated with either MeUb (Figure B, lanes 3–4) or wild-type ubiquitin (Figure B, lanes 5–6). In both instances, ubiquitination was observed confirming that the OWOWO peptide was capable of time-dependent ubiquitination and indeed could function as a primary degron. Ubiquitination of ornithines in cytosolic lysates has not been previously demonstrated. Athough unexpected, this is not entirely surprising as the sole difference between ornithine and lysine is the elimination of a CH2 group on the amino acid side chain.

Figure 1

Ubiquitination of ornithine-rich β-hairpin peptides in cell lysates. (A) Names and sequences of peptides studied. Bold amino acids (O—ornithine, K—lysine) correspond to potential ubiquitination sites. Substrates are acetylated at the N-terminus and amidated at the C-terminus. FAM denotes 6-carboxyfluorescein, and p denotes d-proline. (B) Ubiquitination of β-hairpin conjugated to a known degron (OWOWO-RRRG, lane 1–2) compared with β-hairpin alone (OWOWO, lane 3–6) at the indicated time points using either ubiquitin or MeUb. Ubiquitin-conjugated peptides were purified using ubiquitin-binding beads, separated using SDS-PAGE, and visualized using the fluorescein tag. Relative protein sizes (kDa) are compared with the values obtained from a fluorescent protein marker (left of the gel) with mono-, di-, tri-, and tetraubiquitinated species. (C) Optimal ubiquitination site ornithine was identified with either ubiquitin or no lysine ubiquitin. All reactions were incubated at 37 °C for 2 h.

To further characterize this novel primary degron and explore which of the three ornithine residues was the favored ubiquitination site (secondary degron), four new peptides were synthesized with a substitution of arginine for all but one ornithine (OWRWR, RWOWR, and RWRWO) or with no ornithine at all (RWRWR) (Figure A). Arginine was selected because it possesses a positive charge similar to that of ornithine and is unable to undergo ubiquitination. Ubiquitination of this library of potential degrons was performed in the presence of either ubiquitin (Ub) or no-lysine ubiquitin (No K Ub), a ubiquitin mutant in which all of the lysine residues were mutated to arginine to eliminate polyubiquitin chain formation (Figure C). The peptide with the N-terminal ornithine (OWRWR) exhibited the most ubiquitination (Figure C, lanes 5–6). The two other ornithine locations at position 4 (RWOWR) and position 11 (RWRWO) were also ubiquitinated but to a lesser extent (Figure C, lanes 7–10). This suggested that although the ornithines at position 4 and position 11 could be ubiquitinated, the N-terminal ornithine was the preferred ubiquitination site or secondary degron. The absence of any ubiquitination in the peptide with no ornithine residues (RWRWR, Figure C, lanes 3–4) verified that the ornithine residues were indeed the site of ubiquitination. Similar results were obtained in the absence of the proteasome inhibitor MG-132 (data not shown), which confirmed that although these ornithine-rich β-hairpins do function as primary and secondary degrons, they are lacking the tertiary degron needed for proteasome-mediated degradation. This suggests that this motif may be able to function solely as an E3 ligase reporter.

One interesting observation was the presence of higher molecular weight bands at ∼10, ∼18, and ∼26 kDa in OWOWO samples incubated with MeUb (Figure B, lane 3–4), No K Ub (Figure C, even numbered lanes), and wild-type ubiquitin (Figure B,C). The location on the gel of the ∼10, ∼18, and ∼26 kDa bands was similar for peptides incubated with native ubiquitin, No K Ub, or MeUb, suggesting that these bands correspond to peptides with one, two, or three ubiquitins, respectively. The slightly higher location of the di- and triubiquitinated species incubated with MeUb was due to the higher molecular weight of MeUb (8.7 kDa) compared with wild-type ubiquitin or No K Ub (8.6 kDa). There were two possible explanations for the presence of the higher molecular weight bands. The first was that the concentration of endogenous wild-type ubiquitin in the S100 cytosolic lysates was sufficiently high to compete with the supplemented MeUb or No K Ub to form polyUb chains. The second explanation was that the presence of a single ubiquitin on two or three of the ornithines of the peptide (multimonoubiquitination). To explore these possibilities, the OWOWO peptide was incubated with native Ub or No K Ub in the presence or absence of S100 lysates (Figure S1). The same molecular weight bands were observed when the peptide was incubated in S100 lysates in the presence of Ub or No K Ub (Figure S1, lanes 1–2). Additionally, these same molecular weight bands were observed in the absence of exogenous ubiquitin, suggesting the presence of endogenous ubiquitin in the lysates and the possibility of peptide polyubiquitination even when supplemented with No K Ub.

Kinetic Analysis of Time-Dependent Ubiquitination of β-Hairpin Degrons

Once it was determined that the β-hairpin degrons were readily ubiquitinated, the ubiquitination kinetics were measured to compare and contrast with that previously measured for other degron-based peptide substrates. Time-dependent ubiquitination of the peptides in cytosolic lysates was measured in the presence of No K Ub using a ubiquitin pull-down assay. Ubiquitinated species were isolated using agarose beads conjugated to a ubiquitin antibody, separated by gel electrophoresis, and visualized using the fluorescein tag to identify the time-dependent ubiquitination of OWRWR (Figure A) and OWOWO (Figure C). Peptide with greater than one ubiquitin added was observed for both peptides and was likely due to the addition of multiple single ubiquitins (multimonoubiquitination for OWOWO) and polyubiquitination (for OWOWO and OWRWR). Quantification of time-dependent ubiquitination was performed by measuring the fluorescent intensity of the bands corresponding to mono-, di-, and triubiquitinated species (Figure B,D). These values were normalized by the fluorescent intensity of unreacted peptide to calculate time-dependent relative ubiquitination. Next, the reaction kinetics for the stepwise ubiquitination of the peptides was performed using a modified kinetic model similar to that presented by Melvin et al.[12] The model incorporated first-order reaction kinetics to describe the transfer of ubiquitin to an ornithine residue on the peptide in addition to considering unmodified peptide or degraded peptide (Figure E). The model, governed by a set of ordinary differential equations (ODEs), was fit to the quantitative data using a Markov Chain Monte Carlo (MCMC) algorithm to successively minimize the cumulative sum of the squared deviation between the model and the data as previously described by Melvin et al.[12] The kinetic model calculated three kinetic rate constants for mono- (k1), di- (k2), and tri- (k3) ubiquitinated species. One difference between this model and the one reported previously by Melvin et al. was the definition of the rate constant k4. Here, the k4 term characterized the loss of the peptide (or peptide no longer competent to participate in the ubiquitination reaction). In the model, the presence of di- and triubiquitinated species is represented by the presence of two or three ubiquitin species on the peptide, that is, a diubiquitinated species represents both the addition of two ubiquitin monomers to a single ornithine residue or the addition of a single ubiquitin monomer to two different ornithine residues. As expected, the kinetic model achieved a very good fit for all three peptides tested with low values for the cSSD (Figure F, top) and Pearson correlation coefficient values near 1 for most of the species tested (Figure F, bottom). As per convention, a positive Pearson correlation coefficient approaching 1 corresponded to a very good fit.

Figure 2

Quantification of peptide multi-monoubiquitination. Time-dependent ubiquitination of OWRWR (A) and OWOWO (C) peptides. Ubiquitin-conjugated peptides were purified using ubiquitin-binding beads, separated using SDS-PAGE, and visualized using the fluorescein tag. Relative protein sizes (kDa) are compared with the values obtained from a fluorescent protein marker (left of the gel) with mono-, di-, and triubiquitinated species. Experimental data points for mono(closed squares), di(closed circles), and tri(open circles) ubiquitinated species for OWRWR (B) and OWOWO (D). The Y axis depicts the relative ubiquitination (Ci/Co) to permit the comparison of the peptides. The lines indicate the best fit of the data to the kinetic model. (E) Summary of kinetic rate constants determined by the kinetic model. The rate constants correspond to mono- (k1), di- (k2), or tri- (k3) ubiquitinated species along with a term (k4) for peptide unable to participate in the reaction. cSSD is the cumulative sum of the squared difference minimized in the model to achieve the desired fit. Pearson correlation coefficients (r) describe the goodness of fit between the model and experimental data for mono-, di-, and triubiquitinated species.

Peptide monoubiquitination (closed squares) was the dominant reaction product for both the OWRWR and OWOWO peptides (Figure B,D). The low degree of di- and tri-ubiquitination for OWRWR was attributed to the lack of secondary and tertiary ubiquitination sites coupled with the lower amount of endogenous ubiquitin. The observed lack of di- and triubiquitinated OWOWO was ascribed to the N-terminal ornithine being a highly preferred ubiquitination site relative to the other ornithine residues (Figure C). The similar kinetic constants for di- and triubiquitinated species (k2 and k3) for both OWOWO and OWRWR suggested that the presence of the higher molecular weight bands observed for OWOWO was more likely to be due to di- and tripolyubiquitination on the N-terminal ornithine residue instead of multimonoubiquitination of the three potential ornithine residues because the OWRWR peptide cannot be multimonoubiquitinated. To explore the effect of secondary structure on peptide ubiquitination, a scrambled sequence (Scram-OWOWO, Figure A) was synthesized and assayed for ubiquitination in a cell lysate using the ubiquitin pull-down assay (Figure S2). Interestingly, while the di- (closed circles) and tri- (open circles) ubiquitinated species were very low for the structured peptides OWRWR and OWOWO (Figure B,D), the scrambled sequence exhibited a higher degree of both di- and triubiquitination (Figure S2B). Moreover, the rate constants for di- and triubiquitinated species for Scram-OWOWO were 1.5 times greater than those for the structured peptides, suggesting an increased degree of multimonoubiquitination of the three ornithine residues. This was attributed to the lack of secondary structure in the peptide, which likely increased the availability of the internal ornithine residues for multimonoubiquitination.

Structural Characterization of β-Hairpin Degron

CD was used to determine the intrinsic structure of the degrons OWOWO and OWRWR. Two closely related peptides, WKWK[15] and WRWR,[16] have been fully characterized using NMR and CD for WKWK, and shown to be extremely well-folded (>93% in both cases). Thus, we used CD to confirm that the changes made to the degron sequences do not negatively impact the structure. Although it was demonstrated that the secondary structure is not a requirement for the ornithine-rich peptides to be ubiquitinated, it is well-established that secondary structures like β-hairpins dramatically increase the resistance of peptide substrates to proteases and peptidases.[13] A scrambled version of the OWOWO peptide (Scram-OWOWO, Figure A) was also characterized using CD to serve as a negative control as its scrambled sequence should adopt a random coil structure. The resultant CD spectra of Scram-OWOWO confirmed that it was unstructured based on the minima near 195 nm (Figure ). Conversely, the CD spectrum of the OWOWO peptide displayed characteristics of this class of β-hairpin peptides with a minimum near 205 nm (Figure ). Additionally, the peptides exhibit a minimum near 215 and maximum near 225 nm that arise from exciton coupling of the indole rings of the cross-strand tryptophan residues, further supporting the correct fold and register.[17] The CD spectrum of OWRWR similarly illustrated β-hairpin characteristics with a minimum near 205 nm and exciton coupling between the sidechains of the cross-strand tryptophan residues (Figure ). These spectra are consistent with the CD spectrum of the related peptide, WKWK, that has previously been characterized using both NMR and CD and shown to adopt a well-folded β-hairpin.[15] Although the mean residue ellipticities (MRE) of OWOWO and OWRWR differ, neither exhibits a peak at 195 nm, which is the characteristic of the random coil structure. Thus, the differences in MRE may be related to subtle differences in tryptophan conformation and degree of twist in the β-hairpin.[15]

Figure 3

CD spectra of a β-hairpin degron. Experiments were performed using 40 μM peptide in 10 mM sodium phosphate buffer, pH 8 at 25 °C. Scram-OWOWO (blue) was determine to be a random coil while OWOWO (green) and OWRWR (red) peptides exhibited spectra associated with a well-folded β-sheet confirmation.

β-Hairpin Degrons Demonstrate Increased Resistance to Intracellular Peptidases

An important feature of β-hairpin peptides is their increased protease resistance under cytosolic conditions relative to that of unstructured peptides. Previous β-hairpin protectides identified by Cline et al. were highly resistant to peptidase-mediated degradation.[13b] As such, it was essential to determine whether the β-hairpin peptides identified in this study not only acted as a primary degron but also exhibited the peptidase-resistant properties of other β-hairpin protectides. To test this, the OWRWR peptide was incubated with OPM2 lysates at 37 °C to mimic the intracellular environment of multiple myeloma cells. After varying the incubation times, peptide fragmentation was assayed by separating the mixture using CE and identifying peptide fragments from their fluorescence and migration times as previously described by Proctor et al.[18] The OWRWR peptide was resistant to the degradation in the OPM2 lysates (Figure , black squares), with 80% of the full-length peptide remaining after 180 min, yielding a half-life of 630 min for the intact peptide in the cytosolic mixture. Conversely, the scrambled version of the OWRWR peptide (termed Scram-OWRWR, sequence in Figure A) was almost completely degraded in the OPM2 lysates within 60 min (Figure , blue circles), with a half-life of 30 min. Similar to Scram-OWOWO, Scram-OWRWR did not exhibit secondary structure owing to the randomization of the sequence (data not shown). Thus, the well-folded nature of the OWRWR peptide resulted in a pronounced increase in the stability when compared with its unstructured counterpart. Another peptide that has previously been demonstrated to be rapidly degraded (III-67B, sequence in Figure A) was used as an additional control and was found to have a half-life of 4 min in OPM2 lysates (Figure , red triangles).[11a] Taken together, these results indicated that although the secondary structure of the peptide was unnecessary for it to act as portable degron (Figure D), the β-hairpin was necessary to increase the lifetime of the peptide in the cytosolic environment.

Figure 4

Analysis of peptide degradation in OPM2 lysates using CE. Stability of structured and unstructured peptides was evaluated by incubation with 2 mg/mL OPM2 lysates at 37 °C. Samples were removed from the enzymatic mixture at the indicated time points and separated using CE to identify and quantify the area of the peak corresponding to the parent peptide. Percent of intact peptide was calculated by dividing the area of the peak corresponding to the parent peptide at the indicated time points to the area of the parent peptide at the 0 min time point. The structured, β-hairpin degron OWRWR (black squares, t1/2 = 626.2 min) was substantially more resistant to degradation than its unstructured counterpart Scram-OWRWR (blue circles, t1/2 = 27.9 min) as evidenced by a substantially greater amount of intact peptide and a significantly higher half-life. An unstructured, rapidly degraded peptide (III-67B) was used as a positive control to demonstrate the activity of the OPM2 lysates (red triangles, t1/2 = 3.96 min). The zero minute time point was achieved using heat-shocked lysates and served as the baseline for the intact peptide. Errors were calculated using three separate experiments.

Select E3 Ligase Inhibitors Modulate OWRWR Ubiquitination

The use of the cytosolic lysates enabled characterization of the ubiquitination and protease resistance of the β-hairpin primary degrons but provided no information as to which E1, E2, and E3 enzymes might be responsible for peptide ubiquitination. As the ornithine-rich peptides were not derived from previously established degron sequences, it was important to perform preliminary studies to identify which E3 ligase families might be responsible for peptide ubiquitination. To accomplish this objective, the single ornithine-containing β-hairpin (OWRWR) was incubated with HeLa cytosolic lysates that were pretreated with select commercially available E1 enzymes and E3 ligase inhibitors in the presence of an excess of exogenous, full-length ubiquitin. Ubiquitinated species were then purified from the mixture using agarose beads coated with anti-ubiquitin. Five E3 ligase inhibitors were selected including serdemetan,[19] nutlin-3,[20] SKPin C1,[21] SMER3,[22] and thalidomide,[23] in addition to the E1 enzyme inhibitor PYR-41 (Figure ). Each inhibitor was added at a concentration three times that of the established IC50 values to ensure appropriate enzyme inhibition in the cytosolic milieu. Additionally, all reactions were supplemented with a DUB inhibitor (ubiquitin aldehyde) and proteasome inhibitor (MG-132) to prevent deubiquitination and proteasome-mediated degradation. A DMSO vehicle control was also performed to evaluate the effect of the inhibitor compounds (Figure , lane 1). The E1 ubiquitin-activating enzyme inhibitor PYR-41 did not alter ubiquitination of OWRWR (Figure , lane 7) relative to the control, suggesting that a different E1 enzyme was responsible for initiating the ubiquitination cascade acting on OWRWR. E3 ligase inhibitors nutlin-3 (Figure , lane 2), serdementan (Figure , lane 3), and thalidomide (Figure , lane 6) had no apparent effect on the ubiquitination of the OWRWR peptide. Interestingly, the SKPin C1 compound, which inhibits the cullin-RING ubiquitin E3 ligase SCF-Skp2, prominently reduced polyubiquitination of OWRWR but not monoubiquitination (Figure , lane 4). Additionally, the SMER3 (small molecule enhancer of rapamycin 3) compound, which is a specific inhibitor of the E3 ligase SCF-MET30, significantly reduced both the poly- and monoubiquitination of OWRWR (Figure , lane 5). These results suggest that the β-hairpin primary degron identified in this study could act as reporters for the SCF family of E3 ubiquitin ligases. This group of multiprotein E3 ligases contain three core subunits including Skp1 (the bridging protein), cullin (the major structural scaffold of the complex), and RBX1 (the RING finger zinc-binding domain responsible for E2 ubiquitin-conjugating enzyme binding). However, each member of the SCF family contains a different F-box protein (e.g., Skp2 or Met30), enabling these E3 family members to recognize and bind to specific target proteins. The results presented here suggest that the β-hairpin OWRWR was recognized and ubiquitinated by an SCF complex; however, its ubiquitination was not constrained to a single F-box protein as a recognition metric.

Figure 5

E3 ligase inhibitors selectively decrease OWRWR ubiquitination in cell lysates. The single free ornithine-containing OWRWR peptide (10 μM) was incubated in HeLa S100 lysates for 2 h at 37 °C in the presence of commercially available E1 and E3 enzyme inhibitors. Inhibitors were added at the following concentrations (three times the IC50 value for each compound): nutlin-3 (0.27 μM), serdementan (30 μM), SKPin C1 (150 μM), SMER3 (300 μM), thalidomide (90 μM), PYR-41 (30 μM), or a DMSO vehicle control. Ubiquitin-conjugated peptides were purified using ubiquitin-binding beads, separated using SDS-PAGE, and visualized using the fluorescein tag. Relative protein sizes are compared with the values obtained from a fluorescent marker (left side). Fluorescent species at the appropriate molecular weight for mono-, di-, tri-, and tetraubiquitinated peptides are labeled accordingly.

These results demonstrate that OWRWR acts as both a primary degron recognized by the SCF family of E3 ligases and a secondary degron providing a site for mono- or polyubiquitination. Although the SCF-based inhibitors have been demonstrated to inhibit ubiquitination of natural target proteins found in the cell,[21,22] it is possible that they are able to recognize other unnatural primary degron sequences such as OWRWR. Recently, Melvin et al. have identified a nine amino acid long peptide (KGSYGK(FAM)RRR) based on a MDM2 binding region on the tumor suppressor p53, which could be readily ubiquitinated in cytosolic lysates.[24] Interestingly, the ubiquitination of this sequence was also impaired in the presence of SKPin C1 and SMER3. Given that OWRWR is not a naturally occurring sequence, it is possible that the SCF family of E3 ligases may have a lack of specificity when it comes to the ubiquitination of short peptide sequences, perhaps favoring polycationic peptides, as is the case for OWRWR and the MDM2-based primary degrons. As such, this area requires further study to identify the role of short peptide sequences as potential primary degrons for the SCF family of E3 ligases. Nevertheless, the results presented here provide the foundation for a new family of primary degrons that are rapidly ubiquitinated and highly resistant to intracellular degradation.

Conclusions

With the rapid expansion in efforts to inhibit the proteasome, there is a great need to expand the chemical reporters for characterizing the enzymatic activity associated with E3 ligase and proteasome function. To that end, we have identified a 12-residue β-hairpin peptide that acts as both a primary degron for ubiquitination and a “protectide”, demonstrating a 20-fold increase in protease resistance compared to an unstructured peptide. Moreover, a screen of a panel of E3 ligase inhibitors demonstrated that OWRWR is a potential substrate to detect the ubiquitination mediated by members of the SCF family of E3 ligases, an important group of E3 ligases that have been linked to several pathologies.[25] Finally, this is the first example of ornithine residues acting as secondary degrons (e.g., ubiquitination sites) on a ubiquitination reporter, which may provide broader benefits as its incorporation into peptides increases protease resistance. This protectide-based primary degron is thus a promising new tool as a reporter for detecting ubiquitination and potentially characterizing the proteasome activity.

Materials and Methods

Substrate Synthesis and Purification

Peptides were synthesized by automated solid-phase peptide synthesis on a Creosalus TetrasUI peptide synthesizer using Fmoc-protected amino acids on a CLEAR-amide resin purchased from Peptide Internationals. All natural Fmoc-[N]-protected amino acids were acquired from Advanced Chem Tech. Fmoc-Orn(ivDde)-OH was purchased from Chem Impex International. 6-Carboxyfluorescein was obtained from Chem Impex International. Activation of amino acids was performed with 4 equiv HBTU (2-(1H-benzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate) and 4 equiv HOBt (hydroxybenzotriazole) in the presence of 5 equiv DIPEA (N,N-diisopropylethylamine) in DMF (dimethylformamide) or NMP (N-methylpyrrolidone). Peptide deprotection was carried out in 2% DBU (1,8-diazobicyclie[5.4.0]undec-7-ene) and 2% piperidine in DMF for 2 cycles of 15 min each. Each natural amino acid coupling step was performed twice for 30 min or 1 h. For coupling of Fmoc-Orn(ivDde)-OH (4 equiv), standard coupling agents were used for a single coupling of 4 h. All peptides were acetylated at the N-terminus with 5% acetic anhydride and 6% 2,6-lutidine in DMF for 35 min. Deprotection of the Fmoc-Orn(ivDde)-OH side chain was performed with 3% hydrazine monohydrate in DMF for 3 × 3 min. Removal of the ivDde protecting group was confirmed by the Kaiser test. Conjugation of 6-carboxyfluorescein (4 equiv) was done using 4 equiv PyBOP (benzotriazol-1-yl-oxytripyrrolidinophosphonium hexafluorophosphate), HOBt (4 equiv), and DIPEA (8 equiv) in DMF and allowed to react overnight in the dark. Cleavage of the peptide from the resin was performed in 95:2.5:2.5 TFA (trifluoroacetic acid)/TIPS (triisopropylsilane)/dH2O for 3.5 h. TFA was evaporated and cleavage products were precipitated with cold ethyl ether. The peptide was extracted into water and lyophilized. It was then purified by reverse-phase HPLC using an Atlantis C-18 semipreparative column first with a gradient of 0–100% B over 60 min and second with a gradient of 0–100% B over 100 min, where solvent A was 95:5 water/acetonitrile with 0.1% TFA and solvent B was 95:5 acetonitrile/water with 0.1% TFA. After purification the peptide was lyophilized to a powder and identified using ESI or MALDI mass spectrometry.

Cell Culture and Lysate Generation

HeLa S3 cells (ATCC) were maintained in Dulbecco’s modified eagle medium (DMEM) with 10% v/v bovine calf serum (HyClone) and maintained in a 37 °C, 5% CO2 environment. OPM-2 cells (a kind gift from Donald McDonnell) were maintained in RPMI 1640 media supplemented with 12% fetal bovine serum (HyClone), 21.8 mM glucose, 8.6 mM HEPES (pH 7.4), and 1.0 mM sodium pyruvate. All media components were from Cellgro unless otherwise noted. All reagents used in following assays were from Sigma-Aldrich unless otherwise noted. HeLa S100 cytosolic lysates were generated from cells based on the Dignam protocol as previously described.[12] Isolated HeLa S100 cytosolic lysates were quantified with a Nanodrop 2000 (Thermo Scientific), aliquoted, and stored at −80 °C.

Ubiquitin Pull-down Assay

Ubiquitinated reporter was isolated using a ubiquitin pull-down assay to concentrate the reporter and to remove cell lysate debris so that the ubiquitinated reporter was readily detected and quantified on SDS-PAGE gels using the fluorescein appended to the peptide. The ubiquitin pull-down assay was carried out at the indicated times at 37 °C in a total reaction volume of 100 μL containing assay buffer (10 mM Tris-HCl pH 7.6 and 5 mM MgCl2) with 2 mM DTT, 20 μg/mL ubiquitin aldehyde (Boston Biochem), 100 μM MG-132 (EMD Chemicals), 400 μg/mL ubiquitin, methylated ubiquitin, or no lysine ubiquitin (Boston Biochem), 1× ATP energy-regenerating solution (ATP-ERS, Boston Biochem), 4.2 μg of indicated peptide substrate, and 2 mg/mL HeLa S100 cytosolic lysates as the source of E1, E2, and E3 enzymes. At the end of the indicated times, samples were incubated with Control-Agarose beads (LifeSensors), diluted in TBS-T buffer (20 mM Tris-HCl pH 8.0, 150 mM NaCl, and 0.1% v/v Tween-20), for 60 min on a tube rotator at 4 °C. Samples were subsequently centrifuged at 1800×g for 5 min to pellet and remove control beads. The supernatant was transferred to a solution of Agarose-TUBEs (LifeSensors) diluted in TBS-T and incubated overnight on a tube rotator at 4 °C. Ubiquitin-bound beads were washed 5× with 1× TBS-T, and then, the samples were eluted off the bead with 2× tricine sample buffer, heated for 5 min (>90 °C), and then isolated by centrifugation for 5 min at 13 000×g. Samples were loaded onto SDS-PAGE gels (precast 16.5% Mini PROTEAN Tris-Tricine, Bio-Rad) using 1× Tris-tricine running buffer and visualized with a Typhoon Imager (GE Healthcare Life Sciences). Gels were quantified using ImageJ (US National Institute of Health) by comparing the sample intensity to the unreacted parent peptide intensity, which is not depicted in the gels.

Pull-down assays performed in the presence of chemical inhibitors were completed as described above except that the reaction mixture contained 10 μM peptide and was supplemented with 1 μL of Complete ULTRA (Roche) and 1 μL of PhosSTOP (Roche). The HeLa S100 lysates were preincubated for 60 min with the following single inhibitor compounds (all obtained from LifeSensors) at the following concentrations (all three times the listed IC50 value for each compound) before the addition of the remaining reaction mixture components: 30 μM serdementan, 0.27 μM nutlin-3, 150 μM SKPin C1, 300 μM SMER3, 90 μM thalidomide, or 30 μM PYR-41. The reaction mixture was incubated at 37 °C for 2 h and then subjected to ubiquitin-conjugated protein isolation as described above. Samples were separated using SDS-PAGE and visualized with a Typhoon imager.

Circular Dichroism

Peptide concentrations were determined based on the absorbance of FAM at 492 nm. CD spectroscopy data were collected using an Applied Photophysics Chirascan circular dichroism spectrophotometer. Spectra were generated at 25 °C with a wavelength scan (260–185 nm) using 0.5 s scanning in a 0.1 cm cell. All peptides were at a final concentration of 40 μM in 10 mM sodium phosphate buffer (pH 8.02).

Peptidase Degradation Reactions

Peptide degradation studies were carried out in lysates prepared from the OPM2 cell line. Briefly, 1 × 106 cells/mL OPM2 cells were harvested, washed 2×, and pelleted in phosphate buffered saline (PBS; 137 mM NaCl, 10 mM Na2HPO4, 27 mM KCl, and 1.75 mM KH2PO4 at pH 7.4). The cell pellet was re-suspended in an approximately equivalent volume of mammalian protein extraction reagent (MPER, ThermoFisher) to the volume of the cell pellet (∼300–500 μL) and then vortexed for 10 min at room temperature. Following this, the mixture was centrifuged at 14 000×g for 15 min at 4 °C, and the supernatant was transferred to a centrifuge tube and stored on ice until use. Total protein concentration was determined using fluorescamine as previously described.[18] To investigate peptide stability, 10 μM peptide was incubated with lysates diluted to a total protein concentration of 2 mg/mL in an assay buffer (10 mM Tris-HCL, pH 7.6) at 37 °C in the dark. Aliquots of the reaction mixture were removed at set intervals, at which point further peptidase activity was quenched by heating the aliquots at 90 °C for 5 min followed by immediately freezing in liquid nitrogen and then storage at −20 °C until analysis using CE. The 0 min time point measurements were made using lysates that were heat-killed before peptide incubation. Samples were analyzed using a Beckman Coulter ProteomeLab PA800 automated CE-LIF system CE with laser-induced fluorescence (CE-LIF, 488 nm) as previously described.[18] Before analysis using CE, samples were thawed to room temperature and diluted 1:100 in separation buffer (125 mM sodium tetraborate, 3% v/v Tween-20, pH 8.5). Before each analysis, the capillary was washed with 1 M NaOH for 2 min, dH2O for 2 min, and separation buffer for 2 min. Separation was carried out in a fused-silica capillary (30 μm inner capillary and 360 μm inner capillary; Polymicro Technologies) with a total length of 30 cm and an effective length of 20 cm. Samples were injected into the capillary by applying 0.5 psi pressure for 5 s, and separation was initiated by applying a negative voltage of 6.5 kV. Percent intact peptide remaining was calculated by the corrected area under the parent peptide peak divided by the total corrected peak area (the area under all peaks of the electropherogram) for each time point. The identity of the parent peptide peak was confirmed using the parent peptide alone and verified with the t = 0 min electropherogram. The data were analyzed using commercial software (32 Karat, version 8.0, Beckman Coulter). All degradation reactions and analysis was performed in triplicate.

Kinetic Analysis of Substrate Ubiquitination

The analysis of substrate ubiquitination was performed using a series of ODEs to model peptide ubiquitination as previously described.[12] One change in the current version of the model was kinetic rate constant k4 (Figure ). Here, the k4 term characterized loss of peptide (or peptide no longer competent to participate in the ubiquitination reaction). Possible mechanisms for peptide loss include degradation, binding to cytosolic constituents, or sequestration into cellular subcompartments. No K Ub was used as the exogenous source of ubiquitin in all experiments to reduce the possibility of peptide polyubiquitin; however, the presence of endogenous ubiquitin in the HeLa S100 cytosolic lysates prevented the complete elimination of substrate polyubiquitination. The bands identified as mono-, di-, and triubiquitinated peptides were selected based on the approximate molecular weight of the ubiquitinated peptide compared with the standard indicated on the gel. As all three peptides display a similar molecular weight, a band at the same migration distance was selected for all three peptides in each gel. These bands are denoted in Figures and S2. Kinetic rate constants were determined using a MCMC algorithm in MATLAB. Experimental concentration values were determined bywhere Co is the initial concentration of the parent peptide, Ci is the experimental concentration for a sample at a given time, Iexp is the intensity of the sample band determined by ImageJ, and Iparent is the intensity of the unmodified parent peptide (run alongside experimental samples in every gel).