Synthesis and Application of a Versatile Immunoproteasome Activity Probe

Abstract

The immunoproteasome (iCP) has gained significant interest in recent years as it has been discovered to be significantly expressed under inflammatory conditions, as well as playing significant roles in several diseases, such as autoimmune disorders, viral infection, and cancer. Selective inhibitors have been generated as a method to overcome the off-target effects of current proteasome inhibitor therapeutics. However, selective probes that allow for monitoring this protein complex remain limited, hindering our understanding of the iCP. Current probes are non-selective, not commercially available, or require difficult synthesis. Here, we describe the modular synthesis and application of an iCP-selective probe. The modular nature of the synthetic strategy can enable the incorporation of different fluorophores and covalent warheads, demonstrating the versatility of this probe.

Graphical Abstract

The ONX-0914-alkyne probe described here allows for the conjugation to a variety of fluorophores through click chemistry. It can then be used to visualize immunoproteasome activity in gel-based, microscopy, and flow cytometry experiments.

Introduction

Proteasome mediated degradation is essential for maintaining cellular homeostasis, and defects to this process can adversely impact cellular pathways such as cell cycle regulation, apoptosis, antigen presentation, and more.^1-3 Apart from the standard proteasome core particle (sCP), there is another isoform of the proteasome, named the immunoproteasome (iCP), that is expressed significantly in certain immune cells, as well as induced in response to proinflammatory cytokines such as interferon-γ (IFN-γ) or tumor necrosis factor.^4-6 Upon exposure to IFN-γ, the constitutive catalytic subunits β1, β2, and β5 in the sCP are exchanged with the immuno-subunits-β1i, β2i, and β5i respectively (also known as inducible subunits LMP2, MECL-1, and LMP7, respectively). The altered catalytic activity of the iCP produces peptide products that are better presented onto MHC-I complexes for immune system regulation. Given the essential proteolytic nature of the sCP and iCP, several inhibitors have been developed for hematological cancers, and more recently, selective inhibitors have been under investigation for several other iCP related diseases.^7,8 These covalent inhibitors react with the beta subunits, irreversibly inhibiting the proteolytic activity. With the advent of targeted protein degradation technology, there is a renewed interest in leveraging the proteasome for therapeutics beyond inhibition.^2,9 To better understand the nature of this protein complex, there is a need for tools that allow direct monitoring of immunoproteasome activity in live cells.⁷

The activity of the proteasome is often monitored using peptide substrates or substrate mimics, often employing fluorogenic peptides.⁸ The most commonly used are tripeptides or tetrapeptides containing 7-amino-4-methylcoumarin (AMC) at the C-terminus, which has weakened fluorescence when linked to the peptide, but increases in fluorescent around 440 nm when cleaved from the peptide by the proteasome (Figure 1A). Although valuable for in vitro and kinetic assays, the commercially available AMC probes are weakly cell permeable, limiting their use in cell-based assays. To overcome this, our group previously reported a fluorescent peptide-peptoid hybrid probe to monitor real-time proteasome activity in live cells.⁹ A similar approach was employed to develop an immunoproteasome selective activity-based probe, TBZ-1 (Figure 1B).¹⁰ Although these probes and others have been useful in better understanding the sCP and iCP in cells, one common problem remains in that these probes do not necessarily mimic the proteolysis of endogenous protein substrates, and many lack specificity for the iCP. The Overkleft group has reported several tools to overcome these limitations by developing BODIPY functionalized activity-based probes for monitoring proteasome activity.¹¹ Although their probes show excellent activity, most are not commercially available, are synthetically challenging, or do not allow for easy exchange for different fluorophores. We sought to expand upon and improve the selectivity profile of chemical probes to selectively monitor the iCP in live cells.

Figure 1.

A) Structure of commercially available Suc-LLVY-AMC probe. B) Structure of iCP selective peptide/peptoid hybrid probe, TBZ-1

Over the last decade, several covalent inhibitors have been reported that target the iCP selectively. ONX-0914 and KZR-616 are the gold standards for selectivity and inhibiting activity for the β5i subunit. Inspired by the recent advances in the development of covalent inhibitors targeting the iCP,^12,13 we decided to develop a convenient streamlined synthesis of an alkyne-containing derivative of ONX-0914 (Figure 2). Similar inhibitor-based alkyne-containing probes have recently been reported by our lab for the sCP.^14,15

Figure 2.

A) Structures of iCP inhibitors ONX-0914 and KZR-616. B) Design of new ONX-0914 based probe. Epoxyketone synthesis can be achieved in 3 steps, peptide backbone containing alkyne (1) is easily prepared using standard solid phase synthesis, and azide containing fluorophores can be coupled in solution.

Results and Discussion

Having a peptide backbone and epoxyketone moiety at the C-terminus of ONX-0914, we envisioned an approach where the peptide backbone could be synthesized using standard solid phase and the covalent warhead could be attached in solution (Figure 2). Fmoc-Tyr(OMe)-OH was loaded onto chlorotrityl resin to give us a handle for our SPPS. A PEG linker was added after the remaining amino acids, as this linker length can easily be changed if needed. Lastly, to append a fluorescent dye to the N-terminus, we incorporated an alkyne that would allow for linkage to several commercially available azide-containing fluorescent dyes through a copper(I)-catalyzed azide-alkyne cycloaddition (CuAAC) reaction. The intermediate was then cleaved from the resin using TFA and precipitated out of the solution with cold ether. The alkyne-containing peptide intermediate 1 was subsequently used without further purification.

Next, we set out to synthesize the epoxyketone warhead (Figure 3). Although several papers describe the synthesis of enantio- and diastereomerically pure epoxyketone, we had high variability in yield and ease of the reaction. We found that the synthesis reported by McMinn and coworkers¹² provided the desired diastereomer in fair yield (30%) over three steps (Figure 3A). Briefly, commercially available Fmoc-Phe-OH (2) was converted to a Weinreb amide (3) using EDC as the coupling reagent. The Weinreb amide 3 was treated with isopropenyl magnesium bromide to produce the α,β-unsaturated ketone 4 in 73% yield. The alkene 4 was subjected to classical epoxidation using hydrogen peroxide to produce the epoxyketone substrate as a mixture of diastereomers 5 and 6, which were separated using silica gel chromatography. Fortunately, the desired diastereomer 5 was the major product (dr 1.3:1). The Boc group was removed using 50% TFA in DCM to afford pure epoxyketone 5a, which was used without further purification. To verify the configuration of the epoxyketone, we conducted an alternate synthetic route as well (Figure 3B). According to reported methods,¹⁶ 4 was subjected to a Luche reduction affording the alcohol 7 in high diastereoselectivity. The alcohol 7 was then converted to the epoxide 8 using VO(acac)₂ and TBHP. Alcohol 8 was oxidized to the epoxyketone 5 using Dess-Martin periodinane. The NMR spectra of the desired product 5 matched with the reported spectra (Supplemental Material), confirming we had the correct stereochemistry needed for our probe.

Figure 3.

A) Synthesis of epoxyketone 5a. B) Diastereoselective synthesis of epoxyketone warhead 5 to confirm the stereochemistry.

While designing the probes, we considered several fluorescent tags available commercially. Additionally, since we designed our probes for in cellulo experiments, we chose BODIPY FL-fluorophore due to its stability, large extinction coefficient, and cell permeability. We chose to append the epoxyketone moiety 5a as our final synthetic step, to minimize the risk of the epoxide degrading during purification. With the peptide alkyne 1 in hand, we moved to click the peptide fragment with the fluorophore. Briefly, 1 was reacted with the fluorescent BODIPY azide using standard copper-mediated click chemistry to afford the intermediate 8. After purification on reverse-phase HPLC, we attempted the final amide coupling with the epoxyketone 5a to afford the final compound 9 (Scheme 1).

Scheme 1.

Final synthetic steps to generate the probe, 9.

We then set out to demonstrate the efficacy of 9 as a selective immunoproteasome activity-based probe that could be utilized in a variety of assays simply and effectively. A cell viability assay (CellTiter^®Glo) was performed with 9 and ONX-0914 for 24 hrs in Ramos B-cells, Figure S1, which demonstrated our probe was not significantly toxic at the concentrations we wanted to employ in cellular assays. The results show that modification of the ONX-0914 structure lead to a slight decrease in potency (~0.5 μM vs 1.3 μM IC50) which could result from the removal of the morpholine and replacement with the alkyne. With this positive result in hand, we moved forward with determining the efficacy of the probe 9. To begin, it was necessary to establish cell models that had different iCP expression levels. We decided on Ramos B-cells (Burkitt’s Lymphoma) because of its high levels of endogenously expressed β5i, Figure S2 and MRC-5 (lung fibroblast) because of its ability to induce β5i expression upon treatment with IFN-γ, Figure S3.

We wanted to establish the sensitivity of 9 and its ability to interact with the β5i subunit of the iCP in Ramos B-cells. Cells in a 12 well plate were dosed with 9 at concentrations from 25 μM to 0.1 μM for 1 hr. After incubation with the probe, cells were harvested, lysed, and subjected to SDS-PAGE. The gel was subsequently scanned for fluorescence at the wavelength of the BODIPY fluorophore. Excitingly, we were able to detect bands that correspond to the molecular weight of β5i (Figure S4). There is slight nonspecific staining as expected at higher, not relevant concentrations, as ONX-0914 is known to also inhibit other subunits of the iCP and sCP, but to a lesser extent than β5i.To corroborate our evidence that the probe was binding the β5i subunit, the gel was then subjected to western blot analysis using the anti-β5i (LMP7) antibody (Figure 4A, S5). The band at ~22 K Da is the β5i, while the band at ~50 kDa is tubulin. We decided to use 1 μM of our probe as this concentration was the lowest that provided consistent and sufficient signal for the β5i subunit.

Figure 4.

A) SDS-PAGE analysis of Ramos cells treated with 9 at 1 μM scanning gel for Bodipy fluorophore followed by subjecting gel to western blot (WB) analysis with anti- β5i (LMP7) antibody and anti-tubulin antibody. Full gels in triplicate are provided in the supplemental material. B) SDS-PAGE analysis in Ramos cells dosed with DMSO or ONX-0914 for 1 hr, followed by incubation with 9 for 1 hr then gel scanned for BODIPY fluorescence. *Gel has been cropped where black lines are. Full triplicate gels are in SI C) Quantitation of bands fluorescent bands from B. An unpaired t-test was performed to the treated sample against DMSO for statistical significance. p<0.0001: ****; p<0.0002: ***; p<0.0021: **; p<0.0332: *. All experiments done in triplicate (n=3).

Next, we wanted to test 9 and its ability to monitor changes in iCP activity in cells after inhibition of the iCP using a known inhibitor. ONX-0914 was used at two different concentrations to demonstrate a dose dependent reduction in iCP activity. Ramos B-cells were dosed with inhibitor for one hour, then 9 was added at 1 μM for an additional hour. After incubation, cells were harvested, lysed, and subjected to SDS PAGE. The gel was scanned for fluorescence of the BODIPY fluorophore, and the imaging indicated that ONX-0914 worked as expected, blocking the ability of our probe to interact (Figure 4B, 4C, and S6).

To ensure our probe would work in different cell types that have different iCP expression, we utilized MRC-5 fibroblasts. MRC-5 cells express iCP endogenously, but to a lesser extent than Ramos B-cells. To increase the amount of β5i, these cells can be dosed with IFN-γ as demonstrated previously in Figure S3. After activation of iCP expression with IFN-γ, MRC-5 cells were dosed with 9 using the same conditions and controls as the Ramos B-cell experiment. Excitingly these results were very similar to the Ramos B-cell results indicating this probe is sensitive enough to detect the reduction of β5i activity in a cell line with different iCP activity than Ramos (Figure S7). This highlights the ability to use our probe in drug discovery efforts to find new chemical entities that can affect the expression level of the iCP.

We wanted to demonstrate the different applications this probe could be utilized in and show its ability to be used in a high throughput setting. To begin, we wanted to determine if we could see changes in signal based on the amount of probe dosed on a flow cytometer using unfixed cells. Briefly, cells were incubated for one hour with 9 at different concentrations from 5 μM to 0.5 μM. Then cells were pelleted, rinsed with PBS to remove any unreacted 9, and analyzed. Excitingly we could show an increase in the BODIPY fluorophore channel (FITC) intensity corresponding to our probe reacting covalently with the β5i subunit (Figure S8). The pre-dosing with ONX-0914 control was also repeated and demonstrated a decreased intensity in fluorescent signal as compared to the cells treated with DMSO (Figure 5).

Figure 5.

Median FITC values of the different concentrations of ONX-0914 compared to only probe (Left). Quantification of median FITC values (Right). A one-way ANOVA Dunnett’s multiple comparisons test was performed to compare each concentration to DMSO for statistical significance. p<0.0001: ****; p<0.001: ***; p<0.01: **; p<0.05: *

To demonstrate the selectivity of 9 for interacting with β5i and not β5 from the standard proteasome, we subjected our probe to an assay in which we incubated cells with both our probe 9 and a selective β5 probe that has been previously reported by our lab.^14,15 The β5 epoxomicin probe contains a Cy.7 fluorophore which can be used in combination with the 9 to determine proteasome isoform specific activity in cells (Figure S9). Ramos B-cells, with and without treatment of IFN-γ to increase β5i expression, were dosed with both probes. This should demonstrate that the ONX-probe, 9 has an increase in signal due to the increase in β5i expression, whereas the Epoxomicin probe would have no changes as the amount of β5 would remain the same. The results from this show that we can use 9 to monitor changes in β5i levels upon treatment with cytokines like IFN-γ (Figure 6). We can see two distinct bands corresponding to the β5 and β5i, indicating this combination of probes can be used in studies to monitor the ratio of iCP and sCP activity. To further validate this technique, we also utilized this approach to monitor changes in subunit activity after treatment with an inhibitor selective for β5i. KZR-616 is a modified version of ONX-0914 that has minor changes made to improve its ADME properties such as a 10,000-fold increase in solubility. After treating Ramos cells with KZR-616 we would expect to see our ONX-Bodipy probe have reduced signal due to inhibited β5i, while the signal from the Epox-Cy.7 probe should remain the same as it does not interact with the β5i subunit. Upon treating with both probes and monitoring by flow cytometry we are able to see that when the ONX-probe is used there is a significant reduction in signal, while the Epox-probe has no change. This validates the use of these two probes for selective analysis of their corresponding subunits (Figure S10).

Figure 6.

A) SDS-PAGE of Ramos B-cells treated with and without IFN-γ (5 ng/mL 72 hrs) then incubated with epoxomicin probe and 9 for 1 hr. Red corresponds to Cy.7 epoxomicin probe (β5), Green corresponds to BODIPY ONX-0914 probe (β5i).No overlap of probes is seen showing how these probes can be used to selectively monitor modulation of proteasome isoforms. B) Signal from Onx probe 9 and Epox probe to demonstrate ONX probe can measure changes in iCP after addition of IFN-γ, while Epox probe should not as it is selective for standard proteasome. An unpaired t-test was performed to the treated sample against DMSO for statistical significance. p<0.0001: ****; p<0.0002: ***; p<0.0021: **; p<0.0332: *

Conclusion

Although there are several activity-based probes for studying the standard 20S proteasome, there are limited tools available for studying the immunoproteasome in cellular contexts. This is especially true when trying to correlate sCP and iCP activity in the same experiment. Here we described the modular synthesis and application of an ONX-0914 based probe in a variety of activity assays. The generation of this probe was accomplished through solid phase and solution phase chemistry to generate a more time-efficient, and low-complexity synthesis, without sacrificing yield. Furthermore, installing the alkyne handle allows for the generation of different fluorophore tags for more versatility in applications. Rather than clicking a fluorophore one could install other moieties, such as biotin, for pulldown experiments.

In conclusion, we were able to demonstrate this probe’s ability to interact to the iCP in cells using gel-based assays and flow cytometry. Additionally, the probe was able to detect changes in iCP activity after the addition of inhibitors, or inflammatory cytokines, demonstrating its potential use in drug discovery efforts. We also anticipate that the combination of this new iCP probe along with those available for the sCP will help those interested in studying which proteasome isoform is active in their desired cell type.