Bioconjugation Strategies for Revealing the Roles of Lipids in Living Cells

CONSPECTUS:

The structural boundaries of living cells are composed of numerous membrane-forming lipids. Lipids not only are crucial for the cellular compartmentalization but also are involved in cell signaling as well as energy storage. Abnormal lipid levels have been linked to severe human diseases such as cancer, multiple sclerosis, neurodegenerative diseases, as well as lysosomal storage disorders. Given their biological significance, there is immense interest in studying lipids and their effect on cells. However, limiting factors include the low solubility of lipids, their structural complexity, and the challenge of using genetic techniques to directly manipulate lipid structure. Current methods to study lipids rely mostly on lipidomics, which analyzes the composition of lipid extracts using mass spectrometry. Although, these efforts have successfully catalogued and profiled a great number of lipids in cells, many aspects about their exact functional role and subcellular distribution remain enigmatic.

In this Account, we outline how our laboratory developed and applied different bioconjugation strategies to study the role of lipids and lipid modifications in cells. Inspired by our ongoing work on developing lipid bioconjugation strategies to generate artificial cell membranes, we developed a ceramide synthesis method in live cells using a salicylaldehyde ester that readily reacts with sphingosine in form of a traceless ceramide ligation. Our study not only confirmed existing knowledge about the association of ceramides with cell death, but also gave interesting new findings about the structure—function relationship of ceramides in apoptosis. Our initial efforts led us to investigate probes that detect endogenous sphingolipids using live cell imaging. We describe the development of a fluorogenic probe that reacts chemoselectively with sphingosine in living cells, enabling the detection of elevated endogenous levels of this biomarker in human disease. Building on our interest in the fluorescence labeling of lipids, we have also explored the use ofbioorthogonal reactions to label chemically synthesized lipid probes. We discuss the development of photocaged dihydrotetrazine lipids, where the initiation of the bioorthogonal reaction can be triggered by visible light, allowing for live cell modification of membranes with spatiotemporal control.

Finally, proteins are often post-translationally modified by lipids, which have important effects on protein subcellular localization and function. Controlling lipid modifications with small molecule probes could help reveal the function of lipid post-translational modifications and could potentially inspire novel therapeutic strategies. We describe how our previous studies on synthetic membrane formation inspired us to develop an amphiphilic cysteine derivative that depalmitoylates membrane-bound S-acylated proteins in live cells. Ultimately, we applied this amphiphile mediated depalmitoylation (AMD) in studies investigating the palmitoylation of cancer relevant palmitoylated proteins in healthy and diseased cells.

Graphical Abstract

INTRODUCTION

Cell membranes are composed of numerous lipids that differ in overall shape, structure, and cell function.⁵ While lipids are broadly defined by their low solubility in water, their chemical structures are highly diverse and can be complex. Lipids are the main structural elements of cell membranes, and are critical for energy storage and cellular signaling.^6,7 Our group is interested in understanding the importance and function of lipid compartments by carrying out the bottom-up synthesis of artificial cells. Doing so provides models for investigating the role of lipids in the origin of life and understanding the requirement of enzymes in the early evolution of cell membranes.⁸ To carry out these studies, our group has pursued chemical as well as chemo-enzymatic synthesis strategies toward generating membrane-forming lipids.^8-12

An outcome of our work on artificial cells has been the development of technologies to manipulate lipid membranes. While we understand much about the rules governing protein and DNA function in the cell, we know much less about the roles of lipids. Lipids are intractable due to their poor solubility and lack of template encoded biosynthesis.¹³ Current tools for studying lipids are predominantly focused on measuring and cataloging the numerous species present in cells.¹⁴ By adopting the techniques resulting from research on biomimetic membranes, we have an opportunity to move beyond cataloging and toward the manipulation and understanding of biological lipids.

Lipids are known to play an essential role in health and irregular lipid levels are often associated with human disease.^15,16 During our investigations on artificial lipid membranes, we gained a deeper understanding about the reactivity of various single-chain lipid precursors. We became motivated to apply this knowledge toward studying lipid metabolism in human cells, with a particular emphasis on lipids that are involved in disease. Understanding the exact role, function, and reactivity of specific lipids would offer insights to help both the prevention and treatment of human disorders where dysregulation of lipids is observed. In this Account, we give an overview of how we progressed from studying lipids for artificial membrane synthesis, to adapting lipid bioconjugation reactions for the development of tools to study lipids in living cells.

CHEMOSELECTIVE SYNTHESIS OF SPHINGOLIPIDS IN LIVING CELLS

Sphingolipids are defined by their sphingosine backbones and play key roles in both membrane structure and cell signaling pathways.^17,18 Various sphingolipid species have been linked to cell proliferation, apoptosis, migration, inflammation, and intracellular trafficking.^19-23 Furthermore, altered sphingolipid levels are associated with several diseases, such as diabetes, cancer, Alzheimer’s, and lysosomal storage diseases.^24-27 Given their biological and clinical importance, there is considerable interest in understanding the behavior and structure–function relationships of these lipids. However, indirect approaches to control lipid activity through genetic manipulations are rarely specific and often have pleiotropic effects.²⁸ Additionally, poor aqueous solubility makes delivery of exogenous sphingolipids into cells difficult. For instance, ceramides are two-tailed sphingolipids that are unable to pass through cell membranes due to their high lipophilicity. They are formed by ceramide synthases that N-acylate long-chain sphingoid bases with various fatty acids.^28,29 Although known to be involved in signaling and programmed cell death, the exact roles and interactions of the various known ceramide species remain elusive. Previous studies reported the apoptotic effect of permeable unnatural short-chain ceramides on cells.^30,31 However, short-chain ceramide analogues differ significantly in their physical membrane properties as well as biological activity compared to natural long-chain ceramides.^32,33 Therefore, these short-chain analogues are limited in their use in structure–function analyses relevant to natural ceramides and do not enable direct studies on how the structure of the N-acyl group of ceramides affects apoptosis.

Inspired by ceramide synthases, we decided to develop a strategy that overcomes the inherent biological limitations of ceramide delivery by chemically ligating single-chain lipid precursors in vivo to yield natural ceramides in a traceless manner. Using cell-permeable single-chain precursors would allow the delivery of such precursors into cells, where they would subsequently react and form the bioactive ceramide species. Since ceramides contain an amide moiety, we adapted a peptide ligation strategy to create lipids selectively. Previous work had demonstrated that N-terminal serine peptides can be acylated by salicylaldehyde esters to form peptide bonds.³⁴ We hypothesized that we could adapt this peptide ligation tool for the generation of ceramides by N-acylation of sphingosine, since sphingosine contains the same amino alcohol moiety as serine necessary for the serine ligation chemistry. However, we had two major concerns if this chemistry could be performed in living cells. First, the previous work was performed in vitro under nonphysiological conditions, and we anticipated that the salicylaldehyde might hydrolyze under aqueous reaction conditions. Second, the previous work used trifluoroacetic acid in the final step of the ligation reaction and it was not clear if the ligation would work without the acidic workup.³⁴

We prepared three palmitic acid salicylaldehyde esters (1–3, Figure 1A) and examined their ability to acylate sphingosine (4). Using unmodified fatty acid salicylaldehyde esters that were analogous to the previously used salicylaldehydes allowed the formation of ceramide at neutral pH but also significantly hydrolyzed. However, adding an electron-donating group at the aromatic ring of the fatty acid salicylaldehyde significantly improved stability while still enabling ceramide synthesis over several hours. This enabled traceless ceramide ligation (TCL) to take place chemoselectively under biophysiological conditions yielding a natural ceramide (7) and the released salicylaldehyde 8. After demonstrating that ceramides could be synthesized selectively in mitochondria-associated membranes using isotopically labeled precursors, we sought to test the effect of ceramide formation on apoptosis. We observed that the formation of ceramide 7 in HeLa cells was accompanied by decreased cell viability (Figure 1B and C). Quantification using fluorescence detection of caspase 3/7 activity (Figure 1D) confirmed the apoptotic effect of ceramide 7 accumulation in cells. Interestingly, this effect correlated with the ability to translocate the apoptosis regulator protein Bax to mitochondria (Figure 1E).

Figure 1.

Investigating the function of ceramides in living cells by traceless ceramide ligation (TCL).1 (A) Proposed mechanism of the traceless and chemoselective reaction of fatty acid salicylaldehyde esters (1–3) with sphingosine (4) forming a natural ceramide 7 under aqueous conditions. (B and C) Simultaneous treatment of cells with both reactive single-chain precursors 3 and sphingosine (sph, 4) induces cell apoptosis as shown in (B) WST-1 proliferation assay and (C) total live cell count. Treatment with one precursor alone does not decrease cell viability. (D) Treatment of HeLa cells with 100 μM 3 and 40 μM 4 results in the activation of caspases indicative of apoptosis. (E) Production of C16:0 ceramide (7) during TCL triggers translocation of eGFP-tagged Bax from the cytosol to mitochondria. (F) Photocontrolled initiation of TCL by photouncaging using 405 nm laser light. (G and H) N-acyl chain saturation-dependent effect on cell viability and Bax localization. Treatment of cells with 100 μM 3 (G) and sphingosine results in significantly more cell death than sphingosine alone while cells treated with 100 μM unsaturated salicylaldehyde 10 (H) and sphingosine or sphingosine alone show no significant difference in cell viability.

We also explored a method to trigger TCL with spatial and temporal control by functionalizing the amino group of sphingosine (4) with a photocleavable coumarin analogue (9, Figure 1F). Photouncaging with laser scanning microscopy released sphingosine which subsequently reacted with 3 to form ceramide and trigger apoptosis in a cell-specific manner. Ultimately, we think that the photoactivation of ceramide formation will facilitate studies on the functional role of sphingolipids in defined cells or cell regions.

Our initial experiments enabled us to directly detect the apoptotic effect of saturated long-chain ceramides within cells using in situ lipid synthesis, confirming previous conclusions that were made indirectly by using short-chain ceramides.³⁰ However, the effect of N-acyl chain length or saturation on ceramide function had not been directly investigated in previous studies. Human cells produce a number of different ceramide species that differ in N-acyl chain length.³⁵ By synthesizing different acylating salicylaldehydes we generated ceramides that differed in both chain length and saturation. We confirmed formation of all ceramides in cells and demonstrated that the quantity of all resulting ceramides remained mostly consistent. By performing TCL in living cells, we discovered that the chain length of the synthesized ceramides did not significantly affect apoptosis. However, we found that the saturation of the N-acyl chain has an important impact on triggering cell death. While TCL of saturated ceramide 7 almost completely reduced the cell viability of cultured HeLa cells after 16 h (Figure 1G), the apoptotic effect of unsaturated ceramide 11 was minor (Figure 1H). Our finding that ceramides with unsaturated acyl chains were unable to trigger apoptosis in cells suggests a role for membrane fluidity and ceramide-enriched membrane domains in the mechanism of ceramide-driven cell death.³⁶

DEVELOPMENT OF A FLUOROGENIC PROBE FOR THE DIRECT DETECTION OF SPHINGOSINE IN CELLS

Lipids play a key role for the function of healthy cells and an abnormal sphingolipid homeostasis can be caused by numerous pathophysiological conditions.³⁷ The single chain lipid sphingosine forms the backbone of all sphingolipids and has itself an important biological role in cell signaling. Indeed, disruption of sphingosine metabolism can cause severe disorders.³⁸ For instance, the lysosomal storage disorder Niemann-Pick disease type C1 (NPC1) results in the accumulation of sphingosine in cells due to a mutation in the NPC1 gene resulting in poor trafficking of sphingolipids.³⁹ As a result, a straightforward and rapid method for detection and quantification of sphingosine might have potential for clinical use in early disease diagnosis. Current sphingosine detection tools mostly rely on analysis using various time-consuming mass spectrometry methods, but the introduction of the necessary expensive instrumentation in general clinical settings is challenging.^40,41

During the development of TCL, we synthesized lipophilic salicylaldehyde acylating agents that reacted specifically with the aminoalcohol moiety of sphingosine. We reasoned that redesign of the salicylaldehyde acylating agents may enable the generation of a fluorogenic probe that could detect sphingosine in living cells.² To convert the salicylaldehyde ester into a fluorogenic probe, we equipped the ester moiety with a hydrophobic fluorescent dye, as well as a fluorescence quencher via the amine substituent at the aromatic ring (12, Figure 2A). We chose Bodipy Fl as fluorophore since this lipophilic green-fluorescent dye facilitates permeability through cell membranes without decreasing the water solubility of the probe. Furthermore, black hole quencher-1 (BHQ-1) was used as a quencher. The final reengineered salicylaldehyde ester fluorogenic probe 12 showed only minor background fluorescence (96% quenched) and reacted with sphingosine in aqueous conditions, which we confirmed via mass spectrometry analysis. We also demonstrated that the reaction was chemoselective in the presence of other biomolecules including other phospholipids, sphingosine-1-phosphate (S1P), and even 10-fold molar excess of serine. During the ligation reaction with sphingosine, the fluorescent dye is transferred onto the amine of sphingosine and the subsequent release from the quenching BHQ-1 salicylaldehyde results in a fluorescence increase of resulting Bodipy Fl-sphingosine derivative 13. To determine if our probe was compatible with live-cell studies, HeLa cells were incubated with exogenously added sphingosine and probe 12. A dose-dependent increase in fluorescence intensity was observed compared to control experiments which included cells treated with an unreactive salicylaldehyde ester derivative (14, Figure 2B) as well as untreated cells (Figure 2C and D). In addition, our control experiments using probe 14 demonstrated that the increase in fluorescence was not due to enzymatic cleavage of the fluorophore from the salicylaldehyde ester in cells.

Figure 2.

Fluorogenic probe for the detection of sphingosine in living cells.2 (A) Reaction of Bodipy Fl- and BHQ-1-functionalized fluorogenic salicylaldehyde 12 with sphingosine yielding fluorescent sphingosine derivative 13. (B) Unreactive salicylaldehyde ester 14 was used as a negative control probe. (C) Fluorescence microscopy images of HeLa cells treated first with either probe 12 or control probe 14, followed by different concentrations of exogenous sphingosine (0–40 μM) for 20 h. (D) Quantified fluorescence response of 12 within populations of cells after treatment with Sph (0–40 μM). (E) Fluorescence microscopy images of healthy and Niemann–Pick disease type C1 (NPC1) patient-derived fibroblasts treated with probe 12 (7.5 μM) for 24 h. (F) Quantified fluorescence response of 12 within populations of cells.

We next explored if we could detect elevated endogenous sphingosine levels in diseased cell lines. NPC1 is a rare disease but the number of patients suffering from NPC may be underestimated or misdiagnosed due to challenges in detecting the disease.⁴² Current diagnostic methods such as filipin staining of cholesterol in fibroblasts and mass spectrometry detection of oxysterols are rather unspecific since these lipids may be increased due to other lysosomal storage disorders.⁴³ Thus, we were curious if our fluorogenic probe 12 identifies cells with a buildup of sphingosine and potentially offer a complementary diagnostic approach. We obtained fibroblasts from healthy patients and patients suffering from NPC1. After treating both cell lines with 12, we detected a statistically significant increase in fluorescence intensity of the NPC1 cell lines compared to the healthy control cell line (Figure 2E and F). We confirmed that the increased fluorescence intensity is likely due to increased accumulation of sphingosine by comparing our imaging results to lipidomic data gathered in collaboration with the Metallo laboratory. Further improvement of these probes may be achieved by enhancing the acylation reaction rate as well as improving the quenching of the fluorescent dye to reduce background signals.

DEVELOPMENT OF PHOTOCAGED LIPID PROBES FOR SPATIOTEMPORAL CONTROL OF MEMBRANE LABELING

The dynamics, trafficking, and composition of lipids in cell membranes is constantly changing in a complex choreography between cell membranes, proteins, and cellular signaling molecules. In this context, fluorescently labeled lipid probes can reveal the subcellular localization and biological function of lipids.^44,45 Unlike for proteins, tools do not yet exist for generating specific fluorescent lipids by genetically modifying proteins involved in lipid synthesis.^46,47 Instead, numerous direct and indirect methods have been developed to tag lipids with a fluorophore. These include tagging lipids using fluorescent fusions of lipid binding protein domains and functionalizing lipids with synthetic fluorophores using organic chemistry. Although lipids can be altered chemically at either the tail or head moieties, these modifications can result in a radical change of the lipids’ overall shape and function, particularly when the lipid is ligated to a bulky and charged fluorescent dye. As an alternative, the use of lipid probes that can be modified in situ by bioorthogonal reactions has emerged as a powerful strategy.⁴⁸ By using smaller bioorthogonal handles, the overall geometry of the phospholipid can be mostly unperturbed. Furthermore, highly fluorogenic bioorthogonal probes can be used for live-cell imaging.⁴⁹ These methods require lipid probes containing bioorthogonal handles that are first incubated with living cells and incorporated into membranes. In the second step, the probes are reacted with a functionalized fluorescent dye via a bioorthogonal reaction. In that context, copper-free azide—alkyne cycloaddition as well as inverse electron-demand Diels—Alder (IEDDA) tetrazine ligations have seen frequent use in live-cell imaging.^50-53 Subsequently, there have been numerous bioconjugation applications as well as new synthetic procedures for accessing symmetrical and unsymmetrical tetrazines as well as novel reactive dienophiles such as cyclopropenes and trans-cyclooctene (TCO).^54-58 Despite the widespread adoption of tetrazine ligations in bioconjugation chemistry, a noted drawback is the challenge in gaining spatial and temporal control over the initiation of the IEDDA reaction. Furthermore, reactive tetrazines are often unstable in water or in the presence of strong nucleophiles, and degradation by hydrolysis constitutes a major issue for live-cell imaging.^59,60 One strategy to achieve stimuli responsive tetrazine ligations is to control the redox state of the tetrazine. Tetrazines can be reduced to relatively unreactive dihydrotetrazines. Our laboratory previously developed a stimuli-responsive tetrazine ligation by selective electrochemical oxidation of dihydrotetrazine-functionalized microelectrodes.⁶¹ Although our group succeeded in electrochemically inducing the reaction between tetrazines and TCO-functionalized enzymes with spatiotemporal control, this approach was limited to electrode surfaces and was not applicable to bioorthogonal reactions in cells. As an alternative, visible light-triggered tetrazine formation could be an ideal method for biological applications, due to the minimally invasive nature of light. In a seminal study, the Fox laboratory reported that light can be used as a stimulus for dihydrotetrazine oxidation under the presence of a photosensitizer such as methylene blue.⁶² The reported dihydrotetrazines were stable toward oxidation in air and the resulting tetrazines were also very resilient toward hydrolysis after the photocatalytic oxidation. This method yielded a number of polymeric fibers functionalized with proteins and fluorophores and was applied to photocatalytic activation of tetrazines in mice using a silicon-rhodamine fluorophore as photocatalyst.⁶³ Most recently, Fox and co-workers reported spatiotemporal control over tetrazine labeling in live cells by photocatalytic activation of dihydrotetrazines using fluorescein fluorophores as photocatalysts and ambient oxygen.⁶⁰

One potential drawback of prior approaches is the necessity of a photooxidant. To simplify light-stimulated tetrazine ligation, our group recently developed photocaged dihydrotetrazines for the spatiotemporal control of bioorthogonal reactions with TCO-functionalized probes.³ These dihydrotetrazine derivatives were readily activated using visible light as a stimulus. Interestingly, we determined that functionalization of dihydrotetrazines with photocleavable protecting groups drastically improved their resilience toward hydrolysis and side reactions with nucleophiles. In our initial experiments, we used photocleavable 1-(2-nitrophenyl)ethyl) carbamate as a protecting group to form photocaged dihydrotetrazine 15 (Figure 3A) and observed almost complete decaging and conversion to tetrazine 16 after irradiation for 2 min using visible light (405 nm). We explored the substrate scope of the photocaged dihydrotetrazines by variation of the photocleavable protecting group forming caged dihydrotetrazines that are susceptible toward different wavelengths (17-18, Figure 3B) and also synthesized light-activatable photocaged dihydrotetrazines more electron deficient than our initial caged dihydrotetrazine 15.

Figure 3.

Development of photocaged tetrazine probes for live-cell labeling lipids with spatiotemporal control.3 (A) Light-activated formation of tetrazine 16 from photocaged dihydrotetrazine 15 by irradiation with LED light (405 nm). HPLC-Spectra before (blue) and after (red) LED irradiation for 2 min. (B) Substrate scope of photocaged dihydrotetrazines by functionalization with various light sensitive cages. (C) Illustration of live-cell lipid labeling. (D) Structures of TCO-functionalized fluorescent labeling agents TCO-AF488 (19) and TCO-AF568 (20). (E) Reaction between photocaged dihydrotetrazine-diacylphospholipid probe 21 and TCO-functionalized fluorescent dye. (F) Fluorescence live-cell labeling demonstrating single-cell photoactivation of tetrazine ligation on the cell membrane of a selected HeLa cell using TCO-AF488 (top) and TCO-AF568 (bottom). (G) Activation of four groups of cells at different locations inside a 0.75 mm by 0.75 mm square area using TCO-AF568 for the tetrazine ligation.

Our major goal was to develop new methods to photo trigger the labeling of lipid species in living cells (Figure 3C) using TCO-functionalized fluorescent imaging agents (Figure 3D) such as TCO-AF488 (19) and TCO-AF568 (20). Therefore, we functionalized the headgroup of 1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamine (DPPE) with a photocleavable dihydrotetrazine (Figure 3E). Incubation of this phospholipid probe (22) with cells in aqueous media did not result in degradation, demonstrating the hydrolytic stability of the probe. To test the spatiotemporal precision of our methodology, we irradiated a single cell among a cell population with a 405 nm laser to specifically activate caged tetrazines. We then subsequently reacted the activated lipid probes with TCO-functionalized fluorescent dyes (Figure 3F). Notably, only the irradiated cell exhibited fluorescence labeling. Additionally, we were able to activate multiple cell populations (Figure 3G) over a wide area and demonstrated that the technique was reproducible and robust. Therefore, our methodology offers a precise method for lipid tagging with single-cell resolution. In the future, we hope that spatiotemporal control will allow for imaging of lipid movement between cellular compartments. For instance, a lipid probe can be activated at a specific subcellular location by light and allowed to traffic within a cell before reaction with fluorescent probes followed by imaging. Nevertheless, our model probe 22 contains a rather bulky headgroup that leads to a different overall geometry and polarity compared to DPPE. Future work could circumvent this issue by minimizing the size of the bioorthogonal handle or functionalizing one lipid tail instead of the headgroup.^56,64

In addition to lipid labeling, our group explored several alternative applications for photocaged tetrazines. For instance, the protected dihydrotetrazine was successfully incorporated into peptides using solid phase peptide synthesis (SPPS) with dihydrotetrazine-functionalized Fmoc-protected amino acids as building blocks. We found that in contrary to tetrazines, photocaged dihydrotetrazines are highly stable when submitted to SPPS conditions and can be carried through several peptide coupling cycles. Subsequent photouncaging and ligation with TCO-functionalized fluorophores validated the potential application of photocaged tetrazines in labeling synthetic peptides. Furthermore, we demonstrated light-triggered drug delivery by combining photocaged tetrazines and dienophile-caged doxorubicin conjugates.

DEVELOPING SMALL MOLECULES THAT CAN DEPALMITOYLATE PROTEINS IN LIVING CELLS

While the most prominent function of lipids may be in cell membranes, lipids play many other key roles in cells. For instance, a large percentage of proteins are post-translationally modified by lipids. These modifications are critical for protein function, often altering subcellular localization, folding, and interaction with other protein partners.⁶⁵ An important lipid post-translational modification is S-palmitoylation.⁶⁶ During enzymatic S-palmitoylation, a protein is equipped with a lipid tail by modification of the thiol of a cysteine residue with palmitic acid, resulting in a fatty thioester linkage. S-palmitoylation is reversible, and several thioesterases are known to catalyze the removal of the lipid modification.⁶⁷ Despite the importance of S-palmitoylation, there are limited chemical tools for manipulating its function in cells, and genetic methods are challenging to implement due to the wide range of acyltransferases that exist.

Our group became interested in developing chemical tools for studying the role of S-palmitoylation in cell biology and disease by leveraging our previous experience with lipid bioconjugation reactions. In the past decade, our group developed various chemical methods for the synthesis of vesicle-forming artificial glycerophospholipids in water, such as lipid synthesis using copper-catalyzed azide–alkyne cycloaddition (CuACC) and most recently, oxime ligation.^9,68 We also adopted a well-known tool in peptide synthesis, the native chemical ligation (NCL), to synthesize vesicle-forming phospholipids.^9,10 NCL was developed in the early 1990s by Kent and co-workers for the synthesis of large and synthetically challenging peptides using unprotected peptide fragments.^69-72 The reaction consists of a terminal thioester that readily undergoes transthioesterification with a terminal cysteine followed by an S,N-acyl shift generating the desired peptide. Our group took advantage of the reactivity between terminal cysteines and thioesters and repurposed the ligation for the synthesis of phospholipids. To do so, we functionalized the sn-1 position of a lysophosphatidylcholine (23, Figure 4A) with cysteine and reacted 23 with a fatty thioester 24 to form a diacyl phospholipid analogue (25).¹⁰ Interestingly, we observed that NCL was extremely rapid, did not require an activating catalyst, and the phospholipids formed vesicles in situ.

Figure 4.

Development of a method for S-depalmitoylation of proteins in living cells.4 (A) Phospholipid synthesis by native chemical ligation (NCL) between a cysteine-functionalized lysolipid (23) and a fatty thioester (24) as basis for amphiphile-mediated depalmitoylation (AMD).10 (B) Schematic representation of AMD between MESNA (26) and amphiphilic cysteine derivative 27 yielding NCL product 28 (C) HPLC/ELSD traces of the reaction between 27 (5 mM) and MESNA (26, 5 mM) to form the N-acylated product 28 demonstrate complete conversion. (D) Structure of hydroxy control probe 29. (E) Illustration of the depalmitoylation of S-palmitoylated proteins (SPP) in live-cell membranes using amphiphile 27 (F) Fluorescence microscopy images of HeLa cells expressing eGFP-HRas or control protein eGFP-KRas4B before and after treatment with 27, hydroxy control compound 29, or control compound TCEP.

Given the rapid reaction we observed between cysteine modified lysolipids and acylthioesters, we hypothesized that amphiphilic aminothiols might react with S-palmitoylated proteins by acyl transfer of the thioester lipid modification. Such a scheme would enable direct manipulation of protein lipid post-translational modifications. In our initial experiments, amphiphilic cysteine derivative 27 (Figure 4B) readily reacted with S-palmitoyl sodium 2-mercaptoethanesulfonate (MESNA, 26), a small molecule surrogate for an S-palmitoylated protein, and yielded the NCL product 28 with complete conversion of both MESNA and amphiphile 28 after 90 min (Figure 4C). Moreover, we used an unreactive hydroxy amphiphile (29, Figure 4D) in control experiments, and reaction of 29 with MESNA did not result in depalmitoylation. We then showed that in live cells the thioester moiety of specific S-palmitoylated proteins (SPP, Figure 4E) reacts rapidly with 27 following the NCL mechanism. We termed this reaction amphiphile-mediated depalmitoylation (AMD).⁴ Initially, we investigated the effect of AMD on the protein HRas. HRas is a member of the Ras superfamily, which are responsible for various biological processes such as cell proliferation, differentiation, and apoptosis.⁷³ Furthermore, mutations in Ras are often associated with cancer and other human diseases^74,75 and therefore represent a potential target for cancer therapy. HRas, NRas, and KRas4A are all S-palmitoylated proteins that require the lipid modification for their proper functioning in cells. We demonstrated that administering amphiphile 27 in live cells leads to depalmitoylation of HRas using a resin assisted capture assay. To monitor the role of palmitoylation on the subcellular localization of HRas, HeLa cells were transfected to express enhanced green fluorescent protein (eGFP)-functionalized HRas. Fluorescence microscopy was used to visualize the relocalization of eGFP-HRas from the plasma membrane to the interior of the cell upon depalmitoylation by 27 (Figure 4F).⁷⁶ The experiments with 27 were compared to experiments administering hydroxy amphiphile 29 and we did not observe relocalization nor increased depalmitoylation of HRas. Furthermore, we administered 27 to cells expressing eGFP-KRas4B, a RAS isoform lacking a thioester lipid anchor, and observed no localization change either.

S-Palmitoylation is crucial for various biological processes in cells and a disrupted S-palmitoylation/depalmitoylation cycle can lead to human disease.^77,78 For instance, the degenerative disorder infantile neuronal ceroid lipofuscinosis (INCL) is due to the inactivation of the depalmitoylating enzyme palmitoyl-protein thioesterase (PPT1) leading to accumulated levels of palmitoylated proteins in cells and ultimately cell death.⁷⁹ Therefore, thioester-reactive nontoxic molecules that S-depalmitoylate the accumulated palmitoylated proteins under biophysiological conditions could provide a novel therapeutic strategy to treat such fatal defects that currently lack treatments. Interestingly, when we treated lymphoblasts of INCL patients with amphiphile 27, we observed a decreased palmitoylation level of GAP43 proteins in the cells. Therefore, using depalmitoylating molecules such as 27 may provide a strategy to create small-molecules mimics of thioesterases like PPT1 and alleviate diseases where hyper palmitoylation is a phenotype.

To understand the selectivity of AMD based depalmitoylation, we surveyed how 27 affected the palmitoylation status of a small panel of 8 proteins that are known to be palmitoylated. Despite the simple structure of 27, we noticed significant differences in deplamitoylation depending on the protein. The highest level of depalmitoylation observed for HRas and very little depalmitoylation for proteins such as flotillin-2 and calnexin. This data suggests that additional selectivity might be gained by modulating the structure of the depalmitoylating aminothiols. The mechanism for the observed selectivity is not clear and might be due to the binding of 27 near the site of palmitoylation.

In a follow-up report, we showed that 27 can also depalmitoylate NRas.⁸⁰ Mutations of the NRAS protein are reported to be responsible for promoting tumor growth in several types of cancer including melanoma.⁸¹ We showed that AMD effectively suppresses the NRas signaling activity by depalmitoylation which triggered apoptosis in WM3000 human cells. The growth of this cell line is known to be driven by activating Q61 mutations in NRAS.

CONCLUSIONS

In our initial investigations on artificial membranes, we developed strategies to synthesize lipids in aqueous reaction conditions which inspired us to develop tools to study the function of lipids as well as lipid–protein posttranslational modifications in cells. For instance, by synthesizing ceramides in living cells from cell permeable precursors, we showed that the apoptotic effect of ceramides is limited to saturated ceramides while unsaturated ceramides do not seem to influence the initiation of cell death. In parallel, we developed various approaches to label lipids for live-cell imaging such as by using fluorogenic small molecule probes that react with sphingosine selectively and can detect elevated endogenous levels of sphingosine in diseased cells. In addition, we recently discovered that photocaged tetrazines enable live cell labeling and imaging of lipids with single cell resolution.

Looking to the future, we hope to continue to further expand the development of tools to study lipids and hope these tools will help reveal new knowledge about the cellular function of lipids.

Our lab is currently exploring new chemical methods to assemble lipids more rapidly in living cells. We are also expanding the repertoire of lipid species we can manipulate to reveal their functional roles in cells.

Biographies

Caroline H. Knittel received her Ph. D. in Chemistry from the TU Berlin in 2021 under the guidance of Prof. Roderich Süssmuth. Then, she joined the group of Prof. Neal K. Devaraj as a postdoctoral research fellow at the University of California in San Diego. Her research interests include the synthesis and bioorthogonal tagging of lipids for live-cell imaging.

Neal K. Devaraj is a Professor and Murray Goodman Endowed Chair in the Department of Chemistry and Biochemistry at the University of California, San Diego. A major research thrust of his lab is understanding how nonliving matter, such as simple organic molecules, can assemble to form life. Along these lines, he has developed approaches for the in situ synthesis of synthetic cell membranes by using selective reactions to stitch together lipid fragments. His lab has also made contributions towards applying bioconjugation reactions to understand the role of specific lipid species in living cells.