ABSTRACT
Sphingolipid dysregulation is involved in a range of rare and fatal diseases as well as common pathologies including cancer, infectious diseases or neurodegeneration. Gaining insights into how sphingolipids are involved in these diseases would contribute much to our understanding of human physiology, as well as the pathology mechanisms. However, scientific progress is hampered by a lack of suitable tools that can be used in intact systems. To overcome this, efforts have turned to engineering modified lipids with small clickable tags and to harnessing the power of click chemistry to localize and follow these minimally modified lipid probes in cells. We hope to inspire the readers of this Review to consider applying existing click chemistry tools for their own aspects of sphingolipid research. To this end, we focus here on different biological applications of clickable lipids, mainly to follow metabolic conversions, their visualization by confocal or superresolution microscopy or the identification of their protein interaction partners. Finally, we describe recent approaches employing organelle-targeted and clickable lipid probes to accurately follow intracellular sphingolipid transport with organellar precision.
Introduction
Sphingolipids are a class of bioactive lipids defined by a long chain amino-alcohol (sphingoid) backbone. They make up ∼10 mol% of eukaryotic lipids (van Meer et al., 2008). Complex sphingolipids, such as sphingomyelin (SM) and glycosphingolipids (GSLs), are important structural components of membranes, particularly of the plasma membrane (PM), where they regulate cell–cell interactions (Eich et al., 2016), affect transmembrane receptors (Coskun et al., 2011; Fantini and Barrantes, 2009) and play important roles in cellular signaling (Haberkant et al., 2008; Hamel et al., 2010; Russo et al., 2016). Smaller sphingolipids, such as ceramide, sphingosine and their phosphorylated derivatives, are known to act as first and second messengers in a variety of signaling pathways, regulating cell growth, differentiation, senescence and apoptosis (Futerman and Hannun, 2004; Quinville et al., 2021). As such, imbalance in sphingolipid levels invariably results in physiological and pathological consequences. In fact, aside from a class of rare genetic disorders called sphingolipidoses (Arenz, 2017; Platt et al., 2018), abnormal sphingolipid levels accompany many common diseases, such as diabetes, cancer and neurodegenerative disorders (Ogretmen, 2018; Pan et al., 2023; Song et al., 2022). Subcellular metabolism (see Fig. 1) and organellar distribution of sphingolipids is tightly regulated by an array of metabolic enzymes, as well as a network of sensors and transport pathways along which lipids are moved. These pathways involve vesicular trafficking and non-vesicular lipid transport, mediated by specialized lipid transfer proteins capable of distributing sphingolipids between organelles (Chiapparino et al., 2016; Lev, 2010).
Sphingolipids undergo constant metabolic turnover inside the cell. Left, de novo synthesis of sphingolipids starts in the cytosolic leaflet of the ER, where serine and palmitoyl-CoA are condensed to form 3-ketosphinganine, which is subsequently reduced to sphinganine and, furthermore, ceramide (Cer). More complex sphingolipids, such as sphingomyelin (SM) and glycosphingolipids (GSL), are produced in the Golgi. Sphingolipids are greatly enriched in the PM, where SM forms up to 20% of lipid content. Turnover of such sphingolipid-rich membranes and breakdown of sphingolipids mostly occurs in the endolysosomal compartments. The resulting sphingosine is then exported and either reintegrated into biosynthetic pathways or phosphorylated to sphingosine-1-phosphate (S1P). Right, structure of the most common sphingolipids. R in hexosylceramide (glucosylceramide) represents additional carbohydrate units (such as glucose, galactose, N-acetylglucosamine, N-acetylgalactosamine and others) conjugated to the hexosylceramide structure to create complex glycosphingolipids.
Sphingolipids undergo constant metabolic turnover inside the cell. Left, de novo synthesis of sphingolipids starts in the cytosolic leaflet of the ER, where serine and palmitoyl-CoA are condensed to form 3-ketosphinganine, which is subsequently reduced to sphinganine and, furthermore, ceramide (Cer). More complex sphingolipids, such as sphingomyelin (SM) and glycosphingolipids (GSL), are produced in the Golgi. Sphingolipids are greatly enriched in the PM, where SM forms up to 20% of lipid content. Turnover of such sphingolipid-rich membranes and breakdown of sphingolipids mostly occurs in the endolysosomal compartments. The resulting sphingosine is then exported and either reintegrated into biosynthetic pathways or phosphorylated to sphingosine-1-phosphate (S1P). Right, structure of the most common sphingolipids. R in hexosylceramide (glucosylceramide) represents additional carbohydrate units (such as glucose, galactose, N-acetylglucosamine, N-acetylgalactosamine and others) conjugated to the hexosylceramide structure to create complex glycosphingolipids.
The study of sphingolipids in biological samples largely depends on reliable separation and detection methods. The most common techniques used for this purposes include thin layer chromatography (TLC), high-performance liquid chromatography (HPLC) and mass spectrometry (MS) (Luberto et al., 2019; Wigger et al., 2019; Zhang et al., 2023). With the constant development of these approaches, it has become possible to isolate, detect and quantify a large number of sphingolipid species in complex samples. However, the small size of sphingolipids and their compartmentalization in nano- and micrometer-scale domains in membranes, as well as their rapid metabolic interconversions, means that studying sphingolipids is still very challenging (Hannun and Obeid, 2018; Kraft, 2016). Here, the use of exogenous sphingolipid probes can provide higher specificity and better resolution owing to the reduction in noise from the cellular lipidome.
Earlier approaches, such as using radiolabeled sphingolipids, gave valuable insights into the metabolic characteristics of the sphingolipid pathways, such as sphingoid base recycling (Chigorno et al., 2005) and protein-mediated lipid transport steps (Hanada et al., 2003). However, although radiolabeled probes are excellent mimics of their endogenous counterparts, their applications mostly remain limited to metabolic analyses. For visualization of the subcellular localization, fluorescently labeled sphingolipids have long been a tool of choice for studies of metabolism and subcellular behavior (Bhabak et al., 2012; Kim et al., 2013; Lipsky and Pagano, 1983), but these probes suffer from a range of artifacts derived from the bulky addition of the fluorophore, mainly incorrect incorporation into biological membranes and subcellular mistargeting. In this Review, we focus on synthetic sphingolipid analogs that feature only small modifications; these are detected after the event in question by addition of reporter molecules through biorthogonal chemistry and are thus applicable in the context of intact cells and organisms. For other methods to study sphingolipids in intact systems outside the scope of this Review, see Box 1, and for a guide on how to choose a suitable sphingolipid probe, see Box 2. As highlighted here, the advantage of such minimally modified probes are easy uptake and incorporation into cellular membranes, close mimicry of the localization of their natural counterparts (such as correct organelle targeting) and recognition by the endogenous cellular machinery.
Activity-based probes and fluorescently labeled sphingolipids that employ Förster resonance energy transfer (FRET) represent an elegant way to measure and visualize enzymatic activity in situ (for an excellent review, see Mohamed et al., 2018). Such probes have been employed in the study of sphingolipidoses and other diseases where sphingolipid imbalance is involved. More recently, a fluorescent turn-on probe was developed that is able to visualize endogenous sphingosine, thus circumventing the disadvantages associated with fluorescently labeled lipids (Rudd et al., 2020). With this probe, a sphingosine accumulation in cells from an individual with Niemann–Pick disease type C was demonstrated.
For rapid elevation of sphingolipids within intact cells, caged compounds were developed. The biological activity of caged sphingolipid is blocked and only restored by ‘uncaging’ using a flash of UV light (Höglinger et al., 2015, 2014). Uncaging of sphingosine has been shown to release Ca2+ from acidic compartments, suggesting that sphingosine is an effector of Ca2+ signaling (Höglinger et al., 2015), whereas the use of caged ceramide 1-phosphate (C1P) helped elucidate that cell proliferation initiated by C1P signaling in macrophages relies on activation of the PI3K/Akt and the MEK–ERK1/2 pathways (Gangoiti et al., 2008; Gomez-Muñoz et al., 2016). Combining the use of caged sphingosine with a chemoselective reaction partner helped achieve in situ synthesis of ceramides controlled by light (Rudd and Devaraj, 2018). This approach showed that the saturation of N-acyl chain in ceramides plays a critical role in triggering apoptosis.
Photoswitchable compounds, i.e. azobenzene-modified probes, are extensively used in studies of membrane organization on supported lipid bilayers, as they can shuttle between liquid-ordered and liquid-disordered membrane regions upon irradiation with UV and blue light (Hartrampf et al., 2023). In cells, simple photoswitchable sphingolipids, such as photosphingosine and photosphingosine 1-phosphate (PhotoS1P), were used to gain optical control of S1P signaling in cell and animal models (Morstein et al., 2019). PhotoS1P reversibly induced S1P3 receptor-mediated nociceptive response in cultured neurons and modulated S1P evoked pain hypersensitivity in mice (Morstein et al., 2019).
Researchers faced with the task of investigating sphingolipid-related biological questions can sometimes be overwhelmed with the choice of probes or methods. The effects of additional functional groups or tags on the behavior of the probe are not immediately obvious. Selecting the most optimal synthetic analog naturally depends on the research question and sometimes on available instrumentation. The predominance of fluorophore-labeled sphingolipids in the past can be attributed to their relative ease of application. However, this comes at the cost of accurately representing biological events in the context of a whole cell. The addition of the fluorophore (a large, planar, rigid molecule) to the lipid (a flexible, smaller molecule) invariably changes the way the probe incorporates into biological membranes and how they are trafficked through the cell. Therefore, fluorophore-labeled sphingolipids are unsuitable for investigating cell biological questions. However, they are still popular substrates for enzymatic assays, in particular when investigating headgroup-modifying enzymes. When applied in vitro, or in lysates, such fluorophore-labeled sphingolipids represent a convenient way of assessing enzyme activity. Alternatively, enzyme activity can be measured more elaborately investigated in the context of whole cells through activity-based probes or fluorescent turn-on or energy transfer probes (see Box 1).
Nowadays, minimally modified sphingolipid probes with small clickable tags have become the tools of choice when precise localization or accurate transport kinetics are important. As close mimics of their natural counterparts, they take part in endogenous metabolic networks and are thus rapidly converted into a multitude of downstream metabolites. This is not always desired and sometimes necessitates the use of specialized cell lines, such as sphingosine-1-phosphate lyase (SGPL1) knockout cells to circumvent loss of the tagged part of the molecule to unrelated metabolic networks. Alternatively, when the actions of a single lipid species are of interest, a caged version of the probe (see Box 1) can help to circumvent rapid metabolic conversion during labeling. When deciding between alkyne- or azide-modified probes, it is important to note that alkyne-modified lipids resemble closest their endogenous counterpart, whereas azide-modified lipids show differences in transport kinetics and are therefore less suited for very precise investigations. In exchange, azide modifications enable live-cell click reactions (Fink and Seibel, 2018), which is not yet possible with alkyne lipids.
Click chemistry and its use in sphingolipid research
Bioorthogonal chemistry uses chemical reactions in biological systems in such a way that reactants do not cross-react with other molecules naturally present inside cells, but instead only selectively and specifically react with their exogenous reaction partners. The most commonly used type of biorthogonal chemistry is click chemistry, the development and application of which was honored with the Nobel prize in 2022. The term click chemistry was coined by Sharpless and coworkers around two decades ago (Kolb et al., 2001) for fast and selective reactions which occur under physiological conditions. Click reactions have to proceed at low reagent concentrations and must be specific while in presence of a variety of competing functional groups and high ion concentrations, as typical for biological samples. Most often, carbon–heteroatom reactions are now used to label and track biological molecules (Baskin et al., 2007). In lipid biology, most reactions with lipid probes rely on azide–alkyne cycloadditions (Fig. 2). Such a cycloaddition was first described in 1960s (Huisgen, 1963), but gained relevance only upon the discovery that a metal catalyst, such as Cu(I), accelerated the rate of cycloaddition 106-fold (Rostovtsev et al., 2002; Tornøe and Meldal, 2001). Different types of click reactions exist today: the copper-catalyzed azide-alkyne cycloaddition (CuAAC) can be used for different bioconjugation applications in fixed or lysed samples (Meldal and Tornøe, 2008) and is the reaction of choice for most applications due to its efficiency and high reaction kinetics. To circumvent the intrinsic toxicity of copper, a copper-free variant of click chemistry called strain-promoted azide-alkyne cycloaddition (SPAAC) has been developed to proceed rapidly and selectively in living systems (Baskin et al., 2007). SPAAC is most commonly used for visualization of lipid probes inside cells (Walter et al., 2017). Inverse electron demand Diels–Alder reaction (IEDDA) has a very fast kinetics and finds application in generating fluorophore labeled sphingolipids in living cells (‘tetrazine click’) (Erdmann et al., 2014). In the following sections, we will focus on how these reactions have been applied to advance our understanding of sphingolipid biology, particularly with regards to lipid metabolism and subcellular localization.
Types of click chemistry reactions used in sphingolipid research. Copper-catalyzed click chemistry (CuAAC) uses Cu(I) ions to catalyze the cycloaddition between a terminal alkyne and an azide. It is a widely used type of click chemistry and finds applications in in vitro, as well as in fixed samples. Most sphingolipid probes use this approach because of its efficiency and high reaction kinetics. Copper-free click chemistry (SPAAC) takes advantage of strained alkynes, which, compared to terminal alkynes, have decreased activation energy and therefore do not require a catalyst. SPAAC circumvents using of the cytotoxic copper and is therefore suitable for click reactions in living samples. Inverse electron demand Diels–Alder (IEDDA) reactions have very fast kinetics and finds application in generating fluorophore-labeled sphingolipid compounds in living cells.
Types of click chemistry reactions used in sphingolipid research. Copper-catalyzed click chemistry (CuAAC) uses Cu(I) ions to catalyze the cycloaddition between a terminal alkyne and an azide. It is a widely used type of click chemistry and finds applications in in vitro, as well as in fixed samples. Most sphingolipid probes use this approach because of its efficiency and high reaction kinetics. Copper-free click chemistry (SPAAC) takes advantage of strained alkynes, which, compared to terminal alkynes, have decreased activation energy and therefore do not require a catalyst. SPAAC circumvents using of the cytotoxic copper and is therefore suitable for click reactions in living samples. Inverse electron demand Diels–Alder (IEDDA) reactions have very fast kinetics and finds application in generating fluorophore-labeled sphingolipid compounds in living cells.
Metabolic studies
Alkyne lipids represent the most minimally modified lipids probes (Table 1). They were initially established as convenient substrates for in vitro assays. Here, alkyne-modified sphingolipid derivatives are subjected to enzymatic assays, separated by TLC and detected by click reaction with fluorogenic coumarin-azide (Beatty et al., 2006; Gaebler et al., 2013; Sivakumar et al., 2004). In another experimental setup, alkyne-modified SM was synthesized (Sandbhor et al., 2009), which retained its ability to be catabolized by sphingomyelinase, making it a good probe for sphingomyelinase activity. By showing that alkyne lipids are reliably modified by sphingolipid-processing enzymes, alkyne lipids were established as substrates to be applied also in cellular assays. For instance, they helped elucidate how viruses hijack endogenous lipid processing machinery to support their own replication (Benedyk et al., 2022). In that study, alkyne-modified sphingosine was fed to cellular models of herpesvirus HSV-1 infection, where its conversion into SM was enhanced. More detailed investigations specified that this process occurs by direct activation of ceramide transfer protein (CERT) by a viral protein (Benedyk et al., 2022).
Clickable lipids can also be used to evaluate cellular enzymatic activity in fixed cells. Recently, alkyne-containing sphingosine was used as a reporter in a ‘fix and click’ assay in two different primary cell types from colonic epithelial tissue – a proliferative and a differentiated cell type (Gallion et al., 2022). Cells treated with sphingosine-alkyne were fixed, which halted enzymatic activity. Sphingolipid metabolites were labeled by click chemistry with fluorophore and subjected to capillary electrophoresis with fluorescence detection. This assay allowed the characterization of the activity of sphingolipid-modifying enzymes, such as sphingosine kinase (SphK) and ceramide synthase (CerS), in a single-cell setup (Gallion et al., 2022). In the future, similar assays could be applied for analysis of primary patient samples, for example to assess intratumor heterogeneity in order to design a suitable therapeutic approach.
Sphingolipids are components of most foods and are also produced by gut microbes. Dietary biorthogonal sphinganine was utilized in studies of diet–microbiome interactions (Lee et al., 2021). Oral administration of sphinganine alkyne to mice followed by isolation of gut microbes and click chemistry labeling revealed that dietary sphinganine is nearly exclusively assimilated by Bacteroides, a gut commensal bacterial genus known to produce sphingolipids. Moreover, metabolic tracing has shown that Bacteroides can metabolize dietary sphinganine into dihydroceramides that bear fatty acyl side chains with 15–22 carbons (C15–C22). This discovery helped elucidate how sphingolipid-producing bacteria play a major role in processing dietary sphinganine (Lee et al., 2021).
Alkyne-functionalized analogs also found application in studies on deoxysphingolipids. Deoxysphingolipids are non-canonical, clinically relevant sphingolipids produced when serine palmitoyltransferase (SPT) uses L-alanine instead of L-serine for condensation reaction with palmitoyl-CoA (Penno et al., 2010). Buildup of deoxysphingolipids in cells leads to pathologies, such as slow dying of sensory and autonomic neurons in individuals affected by heredetary sensory neuropathy type 1. An alkyne deoxysphinganine analog was developed to study its metabolic processing (Alecu et al., 2017). Interestingly, this analog locates very rapidly, before any of its downstream metabolites are detected, to mitochondria. Deoxysphinganine is metabolized only into selected species, namely deoxyceramide and deoxydihydroceramide. Accumulation of these metabolites in mitochondria leads to mitochondrial fragmentation and dysfunction. Use of alkyne lipid analogs thus provided an explanation for the toxicity of deoxysphingolipids and the characteristic neuropathic phenotype (Alecu et al., 2017). Treating mouse fibroblasts with the same analog led to vesicle accumulation in the autophagosomal apparatus and to aggregation of intracellular lipids (Lauterbach et al., 2021). In macrophages, N-acylated metabolites, such as deoxyceramide and deoxydihydroceramide, also led to generation of lipid aggregates and crystals, which facilitated the activation of the NLRP3 inflammasome. In summary, experiments with alkyne-functionalized deoxysphingolipids have established a link between deoxysphingolipid pathology in mitochondria and the activation of the innate immune system (Lauterbach et al., 2021).
Besides alkynes, clickable sphingolipid derivatives equipped with an azide moiety are widely utilized (Table 1). For example, ω-azidosphinganine has been used in metabolic assays involving tissue extracts, such as rat liver microsomes, as well as in assays with isolated enzymes (Fink et al., 2021). Studies regarding the behavior of the clicked analogs revealed that a BODIPY-labeled ω-azidosphinganine can still be recognized by CerS to generate labeled dihydroceramide and, in the same assay, recognized by dihydroceramide desaturase to form ceramide. Therefore, a fluorophore-conjugated lipid analog can be converted by more than one consecutive enzymatic conversion in a cell-free system, indicating that clickable sphingolipids can be in principle applied to study metabolic conversions in in vitro setups (Fink et al., 2021). Nevertheless, some enzymes such as sphingosine kinase 1 (SphK1) can process unclicked ω-azidosphinganine but not its BODIPY-clicked version. In metabolic assays with clickable lipids, it must therefore be considered that certain enzymes might display lower tolerance towards the clicked substrate, or they may be inhibited by components of the click-reaction mixture (Fink et al., 2021).
Azidosphingolipids are well suited for MS studies with multiplexed samples. Here, the same sphingolipid analog can be click-labeled with different tags, which allows analysis of sphingolipidomes from different cell populations (Garrido et al., 2015). Differentially tagged lipid extracts can be pooled together to reduce experimental and processing errors. In turn, a simultaneous analysis can distinguish sphingolipidomes of different cells from each other by characteristic masses conferred by their respective click tags. Such differential tagging can be also performed in living cells by utilizing a copper-free click reaction (Garrido et al., 2015). This was demonstrated in a proof-of-principle experiment comparing cells treated with CerS inhibitors. As expected, the treated cells displayed a substantial reduction in incorporating ω-azidosphingosine into ω-azidoceramide and ω-azidosphingomyelin (Garrido et al., 2015).
A particular class of sphingolipid analogs for biophysical, metabolic and signaling studies are photoswitchable ceramide probes, called clickable and azobenzene containing ceramides (caCers) (Kol et al., 2019). These probes take advantage of an azobenzene-containing N-acyl chain. Azobenzene can reversibly switch conformation from cis to trans, depending on irradiation with UV-A or blue light, respectively. The conformational change modifies the curvature of N-acyl chain, thus largely affecting the biophysical properties of caCers and their behaviour in membrane bilayer (Frank et al., 2016). The cis-trans photoswitching also impacts the enzymatic processing of these probes. caCers incubated with yeast lysate overexpressing sphingomyelin synthase 2 were converted into SM more efficiently upon their isomerization to cis with UV-A light. Similarly, the processing of caCer into glucosylceramide by glucosylceramide synthase was faster while in the cis form, suggesting that a general biophysical property of caCers underlies this phenomenon (Kol et al., 2019). To improve cell permeability and advance studies on cellular physiology, short caCers were developed (scaCer). scaCers can easily penetrate through the PM and incorporate into cell membranes. They are recognized by endogenous enzymes and can trigger signaling cascades leading to apoptosis, a feature that was not observed with their long chain counterparts (Morstein et al., 2021).
Of note, to facilitate metabolic studies of sphingolipids, a cell line deficient in S1P lyase (SGPL1) was generated to prevent metabolism of clickable sphingolipids into phospholipids by degradation of S1P (Gerl et al., 2016). Sphingosine-1-phosphate lyase deficient (SGPL1−/−) cells were used to examine the metabolic fate of photoactivatable and clickable sphingosine (pacSph), an alkyne sphingosine with an additional modification (see chapter 5). Indeed, in SGPL1−/− cells, only sphingolipid derivatives of pacSph, but no modified phosphatidylcholine species were observed. The SGPL1−/− cell line is currently used as a reliable tool for studying sphingolipid - protein interactions using clickable lipid analogs (Altuzar et al., 2023; Gerl et al., 2016).
Overall, clickable sphingolipids represent a versatile tool for a wide range of experimental setups, from MS detection of sphingolipid metabolites in cell and tissue extracts to detail oriented studies focused on assessing enzymatic activity in vitro.
Localization
Compared to fluorescent lipids, the advantages of clickable lipids in localization studies are clearly demonstrated: their small, noninterfering tag (usually positioned at the terminal position of the sphingoid base or fatty acid chain) does not interfere with the endogenous localization of the lipid, yet it can be detected post fixation via click reaction with a variety of fluorophores that provide the desired fluorescence properties (Höglinger, 2019; Sternstein et al., 2023).
In in vitro approaches, azide-modified ceramide analogs containing the functional group in the hydrocarbon chains or in the polar head were reconstituted into phospholipid bilayers in giant unilamellar vesicles. Click reaction in situ with a fluorogenic dye helped visualize the ceramide-fluorophore conjugates emitting at ∼440 nm, thus allowing imaging with two photon microscopy at 760 nm (Garrido et al., 2012). However, investigations of lipid analogs in intact cells were still hampered by low efficiency of the click reaction. To circumvent this, the introduction of a copper-chelating moiety such as a picolyl greatly enhanced the reaction rate with its alkyne partner (Uttamapinant et al., 2012), increasing detection sensitivity. In addition, new protocols, in particular investigations into suitable fixation methods resulted in good compatibility with the use of fluorescent proteins or antibody staining (Altuzar et al., 2023; Gaebler et al., 2016).
With these technological advances, the application of clickable sphingolipids began to diversify. One major field of research was studies of host-pathogen interactions. It had previously been demonstrated that sphingosine and other sphingoid bases delivered in cell culture medium have a bactericidal effect on pathogenic bacteria (Drake et al., 2008; Fischer et al., 2012). However, it had not been known whether sphingosine could interact with bacteria directly inside the host cells. To answer this question, azido-sphingosine analogs ω-N3-sphingosine and 1-N3-sphingosine were used to demonstrate that host cell-derived sphingosine incorporated into bacterial membrane in infected cells and interfered with intracellular survival of bacteria, presumably by membrane damage and subsequent lysis of bacterial cells (Solger et al., 2020).
Further research into the bactericidal effects of small sphingolipids also took advantage of correlated light and electron microscopy (CLEM). Treating samples of Gram-negative bacteria with azido-sphingosine and azido-ceramide followed by click reaction and imaging by light and electron microscopy provided a detailed view of how sphingolipids incorporate into the outer bacterial membrane, where these lipids appeared to form elongated structures, which likely impair bacterial viability (Peters et al., 2021).
Azido-sphingosine was also used in studies on viral infection (Lang et al., 2020). Here, macrophages were loaded with azidosphingosine and infected by herpes virus HSV-1. Fluorophore click and co-staining with a marker of intraluminal vesicles showed that azido-sphingosine accumulates in intraluminal vesicles. This supported the authors' hypothesis that HSV-1 infection is prevented by macrophages, which trap HSV-1 viral particles by fusing them with sphingosine-rich vesicles inside the multivesicular bodies and targeting them for lysosomal degradation (Lang et al., 2020).
Even more detailed insights into the importance of sphingolipids in host-pathogen interactions came from the application of clickable sphingolipids in expansion microscopy (ExM). ExM enables super-resolution microscopy using standard confocal fluorescence microscopes by employing hydrogel swelling. Hydrogel facilitates sample expansion by a factor of four to ten, allowing up to ∼20 nm lateral resolution by a confocal laser scanning microscope (Wassie et al., 2019). However, ExM necessitates primary amino groups for reaction with glutaraldehyde or other reagents during fixation and linking into gel. To this end, Götz et al. developed a double-functionalized ceramide analog, α-NH2-ω-N3-C6-ceramide, applicable for sphingolipid ExM (Götz et al., 2020). This synthetic lipid contains an azide group for click labeling and a primary amine for chemical fixation and linkage into sodium acrylate and acrylamide-based gels. Due to its short chain, α-NH2-ω-N3-C6-ceramide efficiently partitions into mammalian membranes. It can be used in conjunction with fixation and immunostaining of cellular proteins, as well as with membrane markers used in super-resolution microscopy such as mCling (Revelo et al., 2014). However, the addition of the amino group confers a positive charge to a naturally neutral molecule, which likely influences its subcellular localization. With α-NH2-ω-N3-C6-ceramide, Götz and colleagues visualized ceramide uptake by several bacterial pathogens in infected HeLa cells. Within few minutes of incubation, this ceramide derivative strongly accumulates in bacterial membranes, and this accumulation is prevented by inhibiting CERT. By combining ExM with another technique, structured illumination microscopy (SIM), it is possible to resolve spatially separated inner and outer bacterial membrane in infected cells. In this way, it was uncovered that α-NH2-ω-N3-C6-ceramide decorated both membranes, which hints on the involvement of active sphingolipid transport during the elusive process of bacterial membrane biogenesis from host-derived lipids (Götz et al., 2020). α-NH2-ω-N3-C6-ceramide was also utilized to study the inhibition of viral replication by short chain ceramides (Brenner et al., 2023). Here, cells were treated with α-NH2-ω-N3-C6-ceramide and infected with SARS-CoV-2. Like its natural sibling, α-NH2-ω-N3-C6-ceramide also inhibited viral replication. ExM with fluorophore-clicked samples showed that the probe is localized to lysosomes. As coronaviruses utilize lysosomes for egress from cells (Ghosh et al., 2020), these results suggest that short ceramides block SARS-Cov-2 replication in this organelle (Brenner et al., 2023).
An alternative to azide-alkyne click reactions for super-resolution microscopy of sphingolipids is the use of “tetrazine click” chemistry. In this approach, a trans-cyclooctene containing ceramide (Cer-TCO) was fed to cells together with a tetrazine-conjugated near-infrared dye silicon-rhodamine (SiR). In cells, these biorthogonal reagents undergo a tetrazine-click reaction and generate SiR-Cer (Erdmann et al., 2014). SiR-Cer is nontoxic and it enables prolonged imaging of Golgi structures in 3D confocal microscopy, as well as in stimulated emission depletion (STED) microscopy.
Clickable sphingolipid probes can also support investigations of T-cell activation (Collenburg et al., 2016). In activated T cells, ceramide is redistributed into clusters on the plasma membrane, and it is largely excluded from the center of the immunological synapse (Mueller et al., 2014). The stimulation-mediated ceramide redistribution in living T cells is very fast. To resolve this process, several azido-functionalized short and long chain ceramides (C6 and C16, respectively) were synthesized (Collenburg et al., 2016). Among them, C6 azidoceramide efficiently incorporated into clusters in membrane of immune cells and, upon T cell activation, partitioned into larger aggregates, likely along with endogenous ceramides generated by neutral sphingomyelinase (Collenburg et al., 2016). C6 azidoceramide helped elucidate that two distinct subpopulations of T cells, regulatory and conventional T cells, vary in their response to ceramide levels in membranes (Wiese et al., 2021). Ceramide concentrations were in turn proposed to work as a rheostat maintaining the balance between regulatory and conventional T cells in humans (Wiese et al., 2021).
When utilizing several different clickable derivatives, it must be kept in mind that the derivatives may differ in accessibility of the clickable group for a chemical reaction, depending on its position on the fatty acyl chain (Walter et al., 2017). The acyl chain length and the position of azide moiety were shown to impact the membrane properties of individual azido-ceramide analogs (Walter et al., 2017). Additionally, the click chemistry workflow, i.e. the click labeling with fluorophore before or after feeding the analog to living cells (“preclicking” and “postclicking”, respectively), also affects membrane incorporation. Long and short chain ceramides preclicked with a fluorophore were more efficiently incorporated into plasma membranes of cultured Jurkat cells, and they frequently show tendency to accumulate in intracellular vesicular structures (Walter et al., 2017). The same study showed that postclicked long chain ceramides, such as α-N3-C16-ceramide and ω-N3-C16-ceramide, have lower click labeling efficiency, as the azide group is shielded in the plasma membrane and inaccessible for click reaction with the fluorophore. In contrast, short chain ceramide analogs α-N3-C6-ceramide and ω-N3-C6-ceramide are easily postclicked, as their azide-functionalized side chain sticks out of the membrane (Walter et al., 2017).
In the above examples, the fidelity of the clickable probes to report accurate localizations corresponding to their endogenous counterparts most critically depended on the fixation procedure used. Except the ceramide derivates used for ExM, lipid probes lack reactive functionalities towards aldehyde fixatives. Therefore, lipid mobility cannot be completely abolished by fixation. Of note, one study found that cell permeabilization with saponin prior to click reaction did not cause mislocalization or signal reduction (Gaebler et al., 2016); however, this was performed with oleate-alkyne probes that are highly present in cells and incorporated into a large variety of different lipid species. In our hands, we found improved signal-to-noise by including a photocrosslinking step before fixation (Haberkant et al., 2016). This resulted in the formation of protein-lipid complexes, which stabilize sphingolipids in place and provide increased resistance towards harsher solvent-based fixation and washing steps. Such a photocrosslinkable and clickable (pac) sphingosine analog (pacSph) was used in metabolic studies (confirming the sensitivity of metabolic interconversions towards inhibition of the endogenous metabolizing enzymes) alongside localization studies using fluorescent microscopy (Haberkant et al., 2016). The use of the same probe in both methods allowed to correlate the predominant metabolite (small sphingolipids at short incubation periods compared to higher sphingolipids at longer labeling times) with the observed localization, corroborating the subcellular localizations of the respective metabolic turnover steps (Haberkant et al., 2016). In further development of pacSph, a photocleavable protection group was added (Höglinger et al., 2017). So-called trifuntional sphingosine could be delivered to cells but became active only upon illumination with long-wavelength UV-light. The subcellular localization of the released pacSph was followed with high temporal resolution. Employing trifunctional sphingosine in confocal microscopy as well as CLEM, it was revealed that sphingosine is accumulated in lysosomes in cellular models, as well as in fibroblasts from individuals affected by Niemann–Pick disease type C1 (NPC1) hinting at a potential role of NPC1 during sphingosine export (Höglinger et al., 2017). The same assay was employed in a study investigating the inhibitory effect of mycobacterial mycolic acids on the cellular NPC1 protein (Fineran et al., 2016). Aside from the benefits of having a photocrosslinkable group in sphingolipid probes in localization studies, another advantage is that the photocrosslinking step can be utilized in workflows to analyze protein-lipid interactions within the context of intact cells, as described in more detail in the following chapter.
Lipid-protein interactions
Photocrosslinking of lipids to interacting proteins in intact cells is one of the few methods to investigate lipid interactions with membrane proteins. The principle of photocrosslinking consists in irradiation with UV light, which generates covalent links to molecules in the immediate vicinity. Photocrosslinking ensures that the captured lipid-protein interaction persists through lysis until detection. Initial efforts gave rise to radiolabeled photoactivatable sphingosine (PhotoSph) (Haberkant et al., 2008). PhotoSph contains a crosslinkable diazirine group and a 3H isotope in the sphingoid backbone. Upon cell labeling, a wide range of PhotoSph-derived radioactive photolabile sphingolipid species are generated in situ (Haberkant et al., 2008). PhotoSph was most prominetly used to investigate the specific interaction of p24 protein with SM (Contreras et al., 2012). Given that such radioactively labeled photocrosslinkable lipids restricted the investigations to only one protein of interest at a time, improving the probe involved replacing the radioactive label with a clickable group, which gave rise to pacSph (Haberkant et al., 2016). PacSph allowed purification of crosslinked proteins (for instance by biotin-avidin pulldown) followed by proteomic analysis, thus being able to identify an ensemble of novel sphingolipid-interactors in a single experimental setup. As with all proteomic workflows, a careful use of controls, such as different crosslinkable lipids, as well as non-crosslinked samples, is required to avoid false positives. As pacSph was rapidly metabolized by SphK into S1P and fed into phospholipid synthesis, the use of SGPL1−/− cells was necessary to restrict downstream metabolism to sphingolipids. In this setting, 38 proteins were identified to interact with sphingolipids, among them already known interactors such as p24 (Haberkant et al., 2016). Later assays using pacSph showed that sphingolipids generated by ceramide synthase 6 were involved in insulin resistance in obesity (Hammerschmidt et al., 2019).
Photoactivatable and clickable ceramide analogs, pacCers, differ in their application compared to pacSph, given the poor solubility of long chain ceramides in aqueous solutions and therefore their poor uptake (Sot et al., 2005). To circumvent this issue, short chain ceramides such as pac-C7-ceramide can be fed to cells and undergo metabolic conversion into long chain ceramides using cellular machinery (Deng et al., 2021). Pac-C7-ceramide was applied in a proteome-wide screen, which identified translocating chain-associated membrane protein 2 (TRAM2) as a sphingolipid sensor involved in translocation of transmembrane proteins (Deng et al., 2021). Long-chain pacCer was used to identify ceramide-binding proteins in cytosolic fractions of mammalian cells (Bockelmann et al., 2018). Closer investigations revealed StAR-related lipid transfer domain protein 7 (STARD7), a protein with primary functions in importing PC into mitochondria, as a ceramide interactor. Furthermore, it was suggested that STARD7-mediated PC transport might be negatively regulated by ceramides (Bockelmann et al., 2018). PacCer has also been utilized in a screen to find ceramide interactors in isolated mitochondria; this screen identified voltage dependent anion channel 2 (VDAC2) as ceramide effector responsible for ceramide induced cell death (Dadsena et al., 2019). Utilizing an in silico approach in combination with mutagenesis and probing the generated mutant in photocrosslinking experiments showed that interaction of VDAC2 with ceramide in cellular membranes depends on a single residue (Dadsena et al., 2019). pacCer is thus established as a suitable tool for identifying and validating ceramide-binding proteins in cytosolic fractions or isolated organelles.
Given that higher sphingolipids are difficult to deliver to live cells, while clickable sphingosine is easy to deliver, but rapidly metabolized into a host of downstream products (including phospholipids), it remained cumbersome to investigate the interaction of proteins with a specific lipid. In addition, many sphingolipids function as second messengers, and as such they exert their function on very short time scales and in a highly localized manner (Gomez-Muñoz et al., 2016; Höglinger et al., 2015; Kolesnick, 1991; Obeid et al., 1993). To solve these issues, we started developing probes equipped with a photocleavable protecting group called a ‘cage’, usually located at a position that is crucial for the function of a particular lipid (Höglinger et al., 2017, 2014). The cage group blocks biological activity, as well as entrance into metabolic pathways, and allows the labeling of cells without activating signaling pathways and causing cell-wide lipid changes. The active probe is released (uncaged) by a flash of light. In this way, it is possible to acutely elevate sphingolipid levels and follow the subcellular distribution, metabolic fate and protein interactors by virtue of the photocrosslinking and click groups.
Such a caged, photoactivatable and clickable (‘trifunctional’) sphingosine has been employed to identify 66 proteins uniquely interacting with sphingosine in living cells (Höglinger et al., 2017). This approach yielded unprecedented temporal control in releasing sphingosine and catching its interactors. However, the properties of the cage group mistarget the probe to all cellular membranes, even those in which endogenous sphingosine would not be found. To gain better spatial precision over the uncaging and crosslinking reactions, multifunctional lipids equipped with organelle targeting functionalities have more recently been synthesized (Jiménez-López and Nadler, 2023).
Multifunctional organelle-targeted lipid probes
Organelle-targeted caged compounds (Fig. 3) are pre-localized to their respective organelle prior to photorelease. For simple synthesis of organelle targeted compounds, a ‘click-cage’ has been designed, which has allowed facile modification of the cage by organelle-targeting groups (Wagner et al., 2018). In a proof-of-concept study, a variety of organelle-targeting groups were chosen to target arachidonic acid and sphingosine to ER, mitochondria, lysosomes and PM (Wagner et al., 2018). When the sphingosine probe was uncaged in the ER, mitochondria and lysosomes, it elicited a transient Ca2+ response as seen in previous reports (Höglinger et al., 2015). In contrast, the uncaged probe in the PM did not trigger such a Ca2+ response (Wagner et al., 2018).
Multifunctional lipid probes and their application in lipid imaging and interactome studies. Different chemical groups enable initial targeting of the caged probe into different organelles, such as the PM, ER, lysosomes and mitochondria. Upon uncaging, the active probe is released inside target organelles and begins to participate in subcellular trafficking, metabolism and signaling machineries. There are several options for use of multifunctional lipid probes in different experimental setups: (1) for metabolic analysis, sphingolipid probes can be subjected to click reaction with coumarin for rapid identification of major metabolites by TLC. If more detailed analysis is required, lipid mass spectrometry detects metabolites by characteristic masses conferred by the alkyne tag. (2) For visualization, a photocrosslinking step is performed at a timepoint of choice after uncaging. Photocrosslinking leads to formation of stable protein–lipid complexes, thus immobilizing sphingolipid probe inside the cell. After cell fixation, click chemistry with fluorophore is used for visualization by fluorescence microscopy. (3) For identification of protein interactors, photocrosslinking step is performed to create protein–lipid complexes. Click reaction with biotin enables the isolation of these complexes by avidin pulldown. Subsequently, protein mass spectrometry can be used to detect protein interactors. Alternatively, if known interaction candidates exist, these can be tested by subjecting isolated protein–lipid complexes to western blotting with candidate-specific antibodies. B, biotin.
Multifunctional lipid probes and their application in lipid imaging and interactome studies. Different chemical groups enable initial targeting of the caged probe into different organelles, such as the PM, ER, lysosomes and mitochondria. Upon uncaging, the active probe is released inside target organelles and begins to participate in subcellular trafficking, metabolism and signaling machineries. There are several options for use of multifunctional lipid probes in different experimental setups: (1) for metabolic analysis, sphingolipid probes can be subjected to click reaction with coumarin for rapid identification of major metabolites by TLC. If more detailed analysis is required, lipid mass spectrometry detects metabolites by characteristic masses conferred by the alkyne tag. (2) For visualization, a photocrosslinking step is performed at a timepoint of choice after uncaging. Photocrosslinking leads to formation of stable protein–lipid complexes, thus immobilizing sphingolipid probe inside the cell. After cell fixation, click chemistry with fluorophore is used for visualization by fluorescence microscopy. (3) For identification of protein interactors, photocrosslinking step is performed to create protein–lipid complexes. Click reaction with biotin enables the isolation of these complexes by avidin pulldown. Subsequently, protein mass spectrometry can be used to detect protein interactors. Alternatively, if known interaction candidates exist, these can be tested by subjecting isolated protein–lipid complexes to western blotting with candidate-specific antibodies. B, biotin.
The power of organelle targeting in disentangling metabolic fates of sphingosine has been demonstrated by comparing mitochondria- and lysosome-targeted isotope-labeled sphingosine. (Feng et al., 2018). Mitochondrial sphingosine was predominantly phosphorylated into sphingosine 1-phosphate (S1P), whereas lysosomal sphingosine was directed to the production of C16 and C24 ceramides in the ER (Feng et al., 2019). Interestingly, mitochondria-released sphingosine generated proportionally more C16 ceramide, whereas lysosomal sphingosine generated more C24 ceramide (Feng et al., 2019). The distinct metabolic outcome indicates that the subcellular sources of sphingoid building blocks greatly affect the metabolic profile of ceramide formation, in spite of the fact that all CerSs localize in the ER. Subcellular compartmentalization thus plays a crucial role in sphingolipid metabolism.
Building on the established organelle-targeting chemistry, we recently designed a lysosomal-targeted, photoactivatable and clickable sphingosine analog called lyso-pacSph (Altuzar et al., 2023). By forcing the probe to accumulate in lysosomes before release by UV irradiation, we were able to investigate sphingosine recycling and transport from the lysosome into the ER in normal and pathological conditions. Making use of its photocrosslinking group, we could identify the known transmembrane sterol transporters NPC1 and LIMP2 as sphingosine interactors (Altuzar et al., 2023). The same probe was then used to follow the metabolic fate and subcellular localization of lysosomal sphingosine in NPC1- and LIMP2-deficient cells to show that NPC1 and, to a lesser extent, LIMP2 are involved in sphingosine recycling and efflux from the lysosome. Besides providing insight into the pathology of NPC disease, the use of lyso-pacSph also demonstrated that lysosomal sphingosine utilizes the same export machinery as lysosomal cholesterol (Altuzar et al., 2023).
As a next step, lyso-targeted sphingosine was then used to elucidate the mechanisms and proteins involved in the export of lysosomal sphingosine into the ER for sphingolipid biosynthetic pathways (Hempelmann et al., 2023 preprint). The results of metabolic assays suggested that STARD3, another known sterol transporter located in the endo/lysosomal membranes, is involved in trafficking of lysosomal sphingosine for re-use in the biosynthetic pathway at the ER (Hempelmann et al., 2023 preprint). This study highlights the power of employing organelle-targeted, caged and clickable sphingolipid derivates in following the subcellular transport routes of sphingolipids and represents another piece in the ever-growing body of evidence for a co-regulation of sphingolipid and cholesterol transport and metabolism.
Conclusions
Research on sphingolipid biology is rapidly advancing and requires constant innovations in tools to answer the complicated questions associated with its functions. Click chemistry and synthetic sphingolipid analogs provide a plethora of strategies for detection, visualization and relative quantification in a variety of setups, from in vitro through cellular experiments up to in vivo experiments in whole animals. Even though a multitude of clickable sphingosine and ceramide analogs has been developed by the chemistry community (as illustrated in Table 1), to date only a few probes are readily available for biologists. Expanding the list of commercially available clickable sphingolipids would therefore be highly desirable in the long term.
Currently, most clickable sphingolipids have been validated for standard assays in cultured cells. However, more advanced studies will still necessitate further development of more-complex chemical tools that are equipped with a combination of functional groups. For example, supplementing clickable probes with organelle-targeted cages will be of tremendous importance to follow sphingolipid transport and acute signaling in a time-resolved manner on a single-organelle level. The recent progress in designing of novel cage and photosensitive protective groups, multicolor uncaging (Amatrudo et al., 2014; Farley et al., 2021; Olson et al., 2013), as well as PM leaflet-specific uncaging (Schuhmacher et al., 2020), could greatly advance our understanding of the sphingolipid asymmetry and complex distribution in cellular membranes, as well as elucidating the functional outcomes of these phenomena. The presence of photocrosslinking groups will not only enable the detection of which protein–lipid interactions occur, but also help mapping interaction surfaces by methods such as crosslinking mass spectrometry (O'Reilly and Rappsilber, 2018). Complementation with other chemical moieties and tags that facilitate immobilization and detection of sphingolipid probes in cellular or tissue samples will expand the scope of utilization of these tools in super-resolution microscopy. Here, developing quantitative readouts of click labeling would be of special advantage. Obtaining detailed information about quantities of probes incorporated into membranes, quantities of lipids that form functional domains, which affect protein effectors or trigger signaling pathways would answer many open questions in the field of sphingolipid biology. Similarly, quantitative standards for MS with click-labeled sphingolipids would greatly improve metabolic studies. The recent development of clickable reporters for MS studies of alkyne lipids has already resulted in highly sensitive tracing procedures, enabling parallel processing and simultaneous analysis of several samples (Thiele et al., 2019). However, this technology has not yet been applied to sphingolipid research. A combination of such clickable MS reporters with alkyne-sphingolipids will certainly lead to faster and more precise sphingolipidomic analyses of many samples undergoing different experimental procedures, such as treatment with drugs.
Clickable analogs have a great potential to dissect biological processes regulated by sphingolipids with unprecedented detail and precision while minimizing artefacts. We envision that they will see a much wider application in the future and become an essential component in a new era of sphingolipid research.
Footnotes
Funding
Our work in this area was supported by the Deutsche Forschungsgemeinschaft (project number 278001972; TRR 186, project A19 to D.H., as well as JA 3315/1-1 to D.J.)
References
Competing interests
The authors declare no competing or financial interests.