A Versatile Synthetic Route to Cycloheximide and Analogues: Potent Inhibitors of Translation Elongation

Keywords

Cycloheximide, CHX, translation elongation inhibitors, protein synthesis inhibitors, ribosome E-site, polysome stabilization, structure-activity relationships, SAR, streptovitacin A, irreversible inhibitors, NHS-ester, ribosome profiling, total synthesis, eukaryotic translation, chlorolissoclimide, lactimidomycin, phyllanthoside, RPL36a, tRNA translocation.

Abstract

Cycloheximide (CHX) is an inhibitor of eukaryotic translation elongation that has played an essential role in the study of protein synthesis. Despite its ubiquity, few studies have been directed towards accessing synthetic CHX derivatives, even though such efforts may lead to protein synthesis inhibitors with improved or alternate properties. Here, we describe the total synthesis of CHX and analogues, and establish structure-activity relationships (SAR) responsible for translation inhibition. The SAR studies aided the design of more potent compounds, one of which irreversibly blocks ribosomal elongation, preserves polysome profiles, and may be a broadly useful tool for investigating protein synthesis.

Introduction

Protein synthesis is a highly coordinated process that involves ribosomes, mRNA, aminoacyl-tRNAs, and various cofactors. Due to the essential role of protein synthesis for life, many organisms for self-defense have evolved to produce secondary metabolites targeting nearly every step of translation. The biological activities of these natural products and their derivatives have been exploited to gain critical insights into tRNA decoding, peptide bond formation, resistance mechanisms to antibiotics, and cellular protein synthesis. In particular, cycloheximide (CHX, 1) has been employed routinely for decades to rapidly and reversibly inhibit elongating ribosomes in order to study protein synthesis and measure protein half-lives.

Figure 1 shows natural products that bind the ribosome E-site and their effects on translation. The diagram illustrates elongating and initiating ribosomes on mRNA, with polypeptide chains and drug binding sites. Cycloheximide (CHX, 1) causes elongation inhibition and is reversible with polysome stabilization. Chlorolissoclimide (2) also causes elongation inhibition and is reversible with polysome stabilization. Lactimidomycin (LTM, 3) causes initiation inhibition and is reversible with polysome run-off. Phyllanthoside (PHY, 4) causes elongation inhibition and is irreversible with polysome destabilization.

Recent biochemical and crystallographic studies demonstrated that 1 binds to the ribosome exit-site (E-site), competitively occupying a pocket where the 3-prime CCA sequences of deacylated tRNAs reside. By blocking translocation of tRNAs, 1 slows down all actively translating ribosomes on mRNA, leading to polysome stabilization. The stalled polysomes can then be visualized in vivo or analyzed to probe post-transcriptional processes. Moreover, polysome stabilization is essential to ribosome profiling, which has revealed details regarding upstream open reading frames, co-translational assembly of proteins, and elongation rates of translating ribosomes. One drawback of stabilizing polysomes with 1 is that binding of 1 to ribosomes is reversible and elongation continues to occur during the course of the experiment. Even at high concentrations of 1, it has been estimated that translation elongation still proceeds at 0.1 to 0.3 amino acids per second, resulting in dose- and time-dependent biases. Thus, irreversible stabilization of polysomes is highly desired to enable more robust downstream experiments to study translation.

Since protein synthesis can be inhibited by stabilizing or disrupting polysomes, several other natural products have been evolved to bind the same pocket as 1 with different effects on polysome stability. For example, chlorolissoclimide (2) stabilizes polysomes in a similar manner to 1, lactimidomycin (LTM, 3) only blocks translation at the initiating codon, while phyllanthoside (PHY, 4) induces dissociation of translating ribosomes from mRNA. Even though high-resolution structures of these molecules are now available, how each inhibitor influences the overall stability of polysomes remains unclear and presents a challenge in designing analogues with additional biological function. Motivated by these questions, herein we report a versatile synthetic route to 1 and potent analogues that enabled the development of an irreversible elongation inhibitor.

Figure 2A shows a structural view of the human ribosome E-site bound by compound 1. Critical bases in the 28S ribosomal RNA (shown in gray) and RPL36a (shown in yellow) are displayed. The structure is from PDB: 5lks. Figure 2B shows the retrosynthetic analysis of compound 1, breaking down the synthesis pathway through intermediates 5, 6, and 7.

Synthetic Strategy and Design

Based on the structure of 1 bound to the human ribosome, we hypothesized that installation of electrophilic functionality extending from the cyclohexanone of 1 may allow covalent interception of nearby lysine residues of RPL36a, a ribosomal protein that comprises the E-site, and subsequent irreversible inhibition of translation. Such a strategy would require alteration of cyclohexanone of 1 without disrupting 1’s activity, but the relative contribution of the cyclohexanone’s substituents to 1’s activity remains elusive. The lack of information could potentially be attributed to the strategy employed in 1’s lone synthesis, wherein the cyclohexanone moiety was prepared in the first stage. Since the importance of glutarimide and C8-OH of 1 has already been established, we devised an alternative route to 1 which would allow for stereodivergent assembly of C11- and C13-substituents late stage. We envisioned cyclohexene 5 as a key intermediate for incorporating different functionality and stereochemistry across the alkene. To access 1 from 5, the desired C13-stereocenter could be installed by directed hydrogenation. The C11-methyl stereocenter in 6 could be secured via a diastereoselective crotylation, while the C8- and C9-stereocenters in 7 could be generated by an Evans aldol reaction.

Total Synthesis of Cycloheximide

The synthesis commenced with an Evans aldol reaction with N-acyl oxazolidinone 8 and aldehyde 9, each of which was obtained in three steps from commercial materials. The aldol product was then protected to provide TBS ether 10 in 98 percent yield over two steps. Chemoselective conversion of 10 to a thioester was achieved using a lithium thiolate, and subsequent Fukuyama reduction delivered aldehyde 11 in 55 percent yield over two steps. Diastereoselective crotylation of 11 was achieved using (E)-crotylboronic acid pinacol ester to yield homoallylic alcohol 12 with the correct C10- and C11-stereochemistry. Ring-closing metathesis of 12 to prepare cyclohexene 13 followed by directed hydrogenation yielded cyclohexanol 14 as a single diastereomer. The C10-carbinol was then oxidized to furnish the corresponding cyclohexanone, and global deprotection of this intermediate using ceric ammonium nitrate completed the total synthesis of 1 in 10 percent yield and 12 steps from commercial 2-(2,6-dioxopiperidin-4-yl)acetic acid.

Scheme 1 shows the total synthesis of cycloheximide (1), streptovitacin A (23a) and irreversible inhibitor (28). The synthetic scheme details multiple reaction conditions and intermediate structures, showing the transformation from starting materials through compounds 8, 9, 10, 11, 12, 13, 14, and various other intermediates (21a, 21b, 22a, 22b, 23a, 23b, 24, 25a, 25b, 26a, 26b, 27, and 28) with specific reagents and conditions for each step.

Structure-Activity Relationship Studies

Following the established route, we synthesized several analogues of 1 with altered C11- and C13-substituents, and evaluated them in a cell-based assay using O-propargyl puromycin (OPP) incorporation. Overall, the C11- and C13-methyl substituents were vital to 1’s activity. Inversion of the C11-stereocenter (15), addition of a second methyl group (16), and removal of the methyl group (17) all abolished inhibitory activity. Inversion of the C13-stereocenter (18) and elongation of C16-methyl to n-butyl (19) also resulted in highly diminished activity. The activity could be recovered when the n-butyl chain was positioned pseudoequatorially (20) instead. These data indicate that CHX analogues likely need to interact with G4370, G4371, and RPL36a Phe56 for successful inhibition. It is noteworthy that other E-site binders (2-4) maintain similar contacts with those residues. For instance, the corresponding methyl groups in LTM (3) are similarly positioned. Moreover, C2-Cl of chlorolissoclimide (2) and C11-methyl of PHY (4) are also oriented in a similar fashion to the C11-methyl group of 1.

Figure 3A shows structures of CHX derivatives synthesized in this study. For each compound, IC50 values for translation inhibition are shown. CHX (1) has IC50 of 2.64 plus or minus 0.54 micromolar. Compound 15 has IC50 greater than 100 micromolar. Compound 16 has IC50 greater than 100 micromolar. Compound 17 has IC50 greater than 100 micromolar. Compound 18 has IC50 greater than 100 micromolar. Compound 19 has IC50 greater than 100 micromolar. Compound 20 has IC50 of 6.98 plus or minus 0.92 micromolar. Streptovitacin A (23a) has IC50 of 0.97 plus or minus 0.26 micromolar. Compound 26a has IC50 of 0.25 plus or minus 0.01 micromolar. Compound 26b has IC50 greater than 100 micromolar. Compound 27 has IC50 of 60.0 plus or minus 18.3 micromolar. Compound 28 has IC50 of 4.16 plus or minus 0.84 micromolar. Error represents standard deviation for n equals 3.

Figure 3B shows dose-response curves demonstrating relative protein synthesis levels (percent maximum on y-axis) after treatment with natural 1, synthetic 1, 26a, or 28 versus vehicle control (0.1 percent DMSO volume per volume). Error bars represent standard error for n equals 3.

Development of More Potent Analogues

Based on these observations, we considered if incorporation of an additional substituent at the C13-position of 1 could lead to a more potent analogue with a functional group handle. Toward this goal, we first targeted streptovitacin A (23a). Mukaiyama hydration of 13 afforded the C13-diastereomeric tertiary alcohols 21a and 21b in 48 percent and 52 percent yield, respectively. Subsequent oxidation to the cyclohexanone and global deprotection then yielded streptovitacin A (23a, 11 percent yield over two steps) and C13-epi-streptovitacin A (23b, 18 percent yield over two steps). In the OPP assay, 23a exhibited robust translation inhibition, thus suggesting the feasibility of derivatizing the C13-position to generate potent analogues.

We then further leveraged the synthetic versatility of 13 to install a quaternary center through a radical-mediated olefin cross-coupling reaction. Hydrogen atom transfer to 13 and subsequent reaction with benzyl acrylate installed the side-chain in 2:1 diastereoselectivity. The C13-diastereomers were separated after Dess-Martin oxidation to give cyclohexanones 25a and 25b in 21 percent and 46 percent yield, respectively, over two steps. Global deprotection then afforded benzyl esters 26a and 26b.

In the OPP assay, 26a was approximately an order of magnitude more potent than 1. The cytotoxicity of 26a was similar to 1, with slightly enhanced selectivity toward certain cancer cell lines. Based on the available structural data, we speculate that the increased potency of 26a may be driven by stabilizing interactions involving the benzyl ester with the E-site in a manner similar to the macrolactone of LTM (3). Given the increased potency of 26a, installation of electrophilic functionality was explored to potentially transform 26a into an irreversible inhibitor. Thus, benzyl ester 26a was converted to N-hydroxysuccinimidyl (NHS)-ester 28 in two steps. Comparable activities were obtained for 28 and 1 in the OPP assay.

Characterization of Irreversible Inhibitor

To determine if 28 irreversibly inhibits protein synthesis, OPP-translation assays were conducted after inhibitor washout. Consistent with the literature, reversible inhibition was observed with 1 and irreversible with 4. Treatment with 28 resulted in irreversible inhibition, which was not due to hydrolysis of 28 or non-specific effects of an NHS-ester. By contrast, inhibition induced by benzyl ester 26a could be rescued by inhibitor washout, albeit to a reduced extent; this partial effect may reflect the increased potency of 26a relative to 1. Additionally, competition experiments between 28 and 1 strongly suggest that 1, in a dose-dependent manner, prevents 28 from binding the same site as 1. Moreover, one minute of incubation time was sufficient to induce irreversible inhibition by 28 at 100 micromolar. Altogether, our data support the notion that NHS-ester 28 effectively irreversibly inhibits protein synthesis and targets the same ribosome E-site as 1.

Figure 4A shows dose-response curves demonstrating relative protein synthesis levels for CHX and 28 treatments after being retained or washed out from the media for 30 minutes. Error bars represent standard error for n equals 3.

Figure 4B shows relative protein synthesis levels (percent maximum on y-axis) as measured by OPP incorporation after compounds (100 micromolar) are retained or washed out for 30 minutes. Error bars represent standard error for n equals 3.

Figure 4C shows relative protein synthesis levels after co-treatment with CHX and 28 at various doses. Cells were preincubated with CHX for 5 minutes before 28 was added. Error bars represent standard error for n equals 3.

Figure 4D shows polysome profiles obtained from 293T cells treated with vehicle (0.1 percent DMSO volume per volume), CHX (100 micromolar), or 28 (100 micromolar) for 90 minutes. The absorbance at 254 nanometers is plotted against distance in millimeters, showing polysome peaks.

Mechanistically, translation inhibition could be driven by either stabilization or dissociation of polysomes (e.g. PHY, 4). To distinguish between the two possibilities, polysome profiles were obtained after treatment of 293T cells with vehicle, 1, or 28. Similar levels of polysomes were observed in samples treated with 1 and 28, demonstrating 28’s ability to globally inhibit translation elongation unlike LTM (3) or PHY (4). It is noteworthy that a single treatment with 28 at 100 micromolar prior to lysis was sufficient to preserve the integrity of polysome. Currently, typical polysome profiling experiments require a high concentration of 1 in all buffers after cell lysis due to the reversibility of 1.

To understand how irreversible inhibition by 28 might be mediated, we conducted immunoprecipitation (IP) mass spectrometry (MS) experiments on RPL36a. Mass shifts corresponding to acylation of Lys22 by 28 were observed in both trypsin and chymotrypsin digests; no other adducts of 28 were detected.

Figure 4E shows HPLC-MS/MS analyses of the lysine 22-modified peptide resulting from trypsin digestion. The peptide sequence shown is NH-HQPHKVTQYK with both b ions and y ions indicated. Observed mass-to-charge (m/z) values are displayed: 266.26 (+1), 310.14 (+1), 482.52 (+2), 532.28 (+2), 539.30 (+1), 620.69 (+2), 638.55 (+1), 728.77 (+2), 1164.85 (+1), 1292.91 (+1), and 1336.98 (+1).

Figure 4F shows a structural view of the human ribosome. E-site tRNA (shown in yellow), P-site tRNA (shown in blue), mRNA (shown in orange), Lys22 (shown in red) of RPL36a (shown in green), and CHX (shown in pink) are displayed. The structure is from PDB: 5lks aligned to 5lzt.

Figure 4G shows a zoomed-in view of the E-site, highlighting the proximity of Lys22 of RPL36a to the binding pocket occupied by CHX.

Since Lys22 is located outside of the putative binding pocket of 28, it remains to be established how the modification of Lys22 is mechanistically relevant to 28’s binding to the E-site and its effect on protein synthesis.

Conclusion

In conclusion, NHS-ester 28 acts as an effectively irreversible inhibitor of protein synthesis by maintaining polysome stability. We anticipate that 28 may serve as an important tool compound to study translation, especially in applications of ribosome profiling where Cycloheximide (1) is suspected of displaying bias due to its reversibility.

The development of irreversible inhibitors like compound 28 is critical for stabilizing the ribosome-mRNA complex during sequencing procedures. This stabilization ensures a more accurate “snapshot” of the translatome. Such advancements in molecular tools are frequently highlighted by the National Human Genome Research Institute (NIH), as they refine our ability to map gene expression with high precision.

Future studies using these CHX analogues to investigate structure-function relationships in the ribosome E-site and interrogate protein synthesis will be reported in due course.