Harnessing Clickable Acylated Glycerol Probes as Chemical Tools for Tracking Glycerolipid Metabolism

Abstract

We report the use of clickable monoacylglycerol (MAG) analogs as probes for the general labeling of glycerolipids during lipid metabolism. Incorporation of azide tags onto the glycerol region was pursued to develop probes that would generally label glycerolipids, in which the click tag would not be removed through processes including acyl chain and headgroup remodeling. Analysis of clickable MAG probes containing acyl chains of different length resulted in widely variable cell imaging and cytotoxicity profiles. Based on these results, we focused on a probe bearing a short acyl chain (C₄-MAG-N₃) that was found to infiltrate natural lipid biosynthetic pathways to produce click-tagged versions of both neutral and phospholipid products. Alternatively, strategic blocking of the glycerol sn-3 position in probe C₄-MEG-N₃ served to deactivate phospholipid tagging and focus labeling on neutral lipids. This work shows that lipid metabolic labeling profiles can be tuned based on probe structures and provides valuable tools for evaluating alterations to general lipid metabolism in cells.

Graphical Abstract

Introduction

The lipidome or the collection of lipids that exists in a cell is extremely complex, in that lipids are products of complicated and interconnected metabolic pathways resulting from constant interconversion between different lipid structures bearing myriad headgroup and acyl chain compositions. Within the intricacies of lipid metabolism, the composition and subcellular localization patterns of lipids often dictate their function, with the capacity to trigger physiological responses. In fact, the abundance of certain lipids is so tightly regulated because they act as signaling molecules for key biological processes.^[1] Such is the case for many glycerolipids that act as mediators of molecular signaling events that enable proper cellular function in eukaryotes.^[2] For a few examples, membrane glycerophospholipids regulate cellular functions by binding and activating various proteins,^[3] triacylgycerols (TAGs) play a major role in lipid droplet formation,^[4] and diacylglycerols (DAGs) play a prominent role in cell signaling by transmitting downstream messages following primary signal recognition events.^[5] As a consequence, dysregulation of such lipids is recognized as a hallmark of many diseases including cancer.^[6] Due to the complexities of lipid metabolic networks, efficient tools and methods that facilitate the tracking of lipid metabolism are critical.

Several approaches for tracking lipid flux have been developed, each of which is characterized by their own scope of applications and limitations. Classically, while strategies for producing radiolabeled lipid derivatives have been invaluable for analyzing lipid metabolism through scintillation counting^[7] as well as stable isotope labeling and mass spectrometry,^[8] these approaches are not as effective for determining lipid subcellular localization, which is ideally accomplished through imaging. Fluorescent lipid analogs have aided in localization studies but commonly disrupt lipid function, localization, and membrane properties due to the introduction of bulky reporter tags that render them structurally distinct from the lipid they are designed to mimic.^[9] Another exciting avenue for studying lipid metabolism and dynamics uses multifunctional lipid probes that typically contain both a photocrosslinking moiety and a click chemistry conjugation handle within the hydrophobic tail.^[10] Such probes have enabled simultaneous lipid analysis and characterization of lipid-protein interactions,^[11] while some lipid surrogates provide information for intracellular lipid localization and lipid dynamics.^[12] Metabolic labeling comprises another exciting modern approach in which labeled lipid products can be traced in cells by utilizing conservative modifications to lipid precursors containing diminutive click tags that are less likely to disturb lipid biosynthesis, structure, and function.^[13]

The metabolic labeling approach takes advantage of exogenously applied precursors that bear a biorthogonal “click” tag such as an azide or alkyne to hijack lipid biosynthesis (Figure 1). The resulting tagged lipid products can be conveniently derivatized via click chemistry to append various reporter tags of use for detecting and tracking lipid biosynthesis. A common lipid metabolic labeling approach has focused on labeling specific phospholipid structures using clickable probe analogs of substrates that map onto the headgroups of target glycerophospholipids^[14] including phosphatidylcholine (PC),^[15]phosphatidic acid (PA),^[16] phosphatidylserine (PS),^[17] phosphatidylinositol (PI)^[18] and glycophosphatidylinositol (GPI) anchors.^[19] This bioorthogonal strategy also extends to labeling ceramide or other sphingolipids.^[20] However, a complementary strategy would enable the labeling of broader families of glycerolipids to obtain a snapshot of changes in lipid metabolism in response to different cellular processes. Such an approach would be comparable to biorthogonal non-canonical amino acid tagging (BONCAT), which has been beneficial for analyzing changes in global protein synthesis.^[21] One method that enables broad lipid labeling utilizes click-tagged fatty acid (FA) precursors to infiltrate the hydrophobic regions of different lipid structures, an approach that has commonly been applied to scrutinize FA metabolism, β oxidation,^[22] and total lipid profiling.^[23] With regard to glycerolipid labeling, a drawback of this approach is that the tag can be removed through acyl chain remodeling and FAs are also diverted into structures other than phospholipids, such as through posttranslational protein lipidation.^[24],[25]

Figure 1.

Lipid metabolic labeling strategy. Click-tagged precursors are substrates for metabolic incorporation and production of labeled lipids. These labeled products can be detected and analyzed via fluorescence microscopy, TLC, and MS, after derivatization with click-partner fluorophores or a mass tag.

In order to obtain a broad snapshot of glycerolipid metabolism, the glycerol scaffold of lipids provides an attractive motif for metabolic labeling since this region is conserved within glycerolipids. Therefore, probes can be developed to target glycerolipid biosynthesis in a manner by which the clickable tag will not be excised if an acyl chain or lipid headgroup is modified through metabolism. Herein we introduce azide-tagged glycerol and monoacylglycerol (MAG) probes as a distinct avenue for lipid tracing. While MAGs are short-lived but common intermediates for glycolipid and glycerophospholipid biosynthesis,^[26] clickable MAG probes have not, to our knowledge, been leveraged as an exogenous biorthogonal chemical reporter for the synergistic analysis and visualization of global glycerolipid flux. Moreover, implementation of MAG probes may also allow access to free glycerol analogs, which has previously been utilized for metabolic engineering studies.^[27] Due to these attributes, we set out to explore the efficacy of clickable glycerol and MAG probes for metabolic labeling of different glycerolipid families.

Results and Discussion

To explore glycerolipid labeling, we developed a series of azide-tagged probes of glycerol as well as MAGs bearing different lengths of acyl chains attached to oxygen. The azide tag was specifically chosen to enable post-derivatization via the strain-promoted azide-alkyne cycloaddition (SPAAC) through reaction with cyclooctyne-containing reagents. The core design features a glycerol derivative, Gly-N₃ (Figure 2A), in which an azidomethylene moiety is appended at the sn-1 position in place of a hydrogen atom on the glycerol backbone. This approach draws from previous phospholipid probes we developed that retained protein binding properties despite the addition of the azidomethylene tag.^[28],[29] Along with Gly-N₃, we also synthesized a panel of monoacylglycerol derivatives containing a single FA tail appended to the sn-3 position (C_n-MAG-N₃, where n denotes the total number of carbons in the acyl chain). This includes analogs with varying acyl chain lengths ranging from short (C₄-MAG-N₃), to medium (C₁₀-MAG-N₃, C₁₂-MAG-N₃, C₁₄-MAG-N₃), and long (and unsaturated, C_18:1-MAG-N₃) (Figure 2A). These MAG probes were explored as an avenue for improving cell-permeability, after which the acyl chain could be hydrolyzed by intracellular lipases/esterases to manipulate acyl chains_.^[30] Finally, we additionally developed C₄-MEG-N₃ as a control probe for comparison in which the sn-3 alkyl chain is attached via an ether linkage to block hydrolytic cleavage by hydrolytic enzymes. The synthetic routes for C_n-MAG-N₃ derivatives and short chain probes C₄-MAG-N₃ and C₄-MEG-N₃ are shown in Scheme S1 and Scheme S2, respectively.

Figure 2.

A) Structures of azide-tagged glycerol probe Gly-N₃, MAG probes of type C_n-MAG-N₃, and ether-linked probe C₄-MEG-N₃. B) Probes were assessed for cytotoxicity by measuring cell or optical density (OD) over time and then subjected to fixation and then SPAAC to image labeled molecules through fluorescence microscopy.

To study metabolic labeling using Gly-N₃ and MAG-N₃ probes, we employed the yeast Saccharomyces cerevisiae, which has proven to be a valuable model system for eukaryotic lipid metabolism.^[31] Moreover, glycerol kinase from S. cerevisiae has previously been shown to tolerate unnatural modification of its glycerol substrate.^[32] We began evaluation of cellular labeling activities by conducting fluorescence microscopy experiments following treatment of cells with each probe (1 mM) followed by SPAAC with the fluorescent reagent dibenzocyclooctyne-cyanine 3 (DBCO-Cy3). After incubation of these glycerol and MAG derivatives with cells, the media was removed and washed. Next, formaldehyde (3.7%) solution was added to the washed cells for fixation. Azide-bearing lipids were then labeled through conjugation to DBCO-Cy3 via SPAAC (Figure 2B), washed to remove unreacted dye, and visualized using confocal fluorescence microscopy. In these experiments, initial glycerol analog Gly-N₃ did not show evidence of incorporation into cells since signal from cells treated with 1mM Gly-N₃ was not more intense when compared to negative control cells not treated with Gly-N₃ (Figure S1). These results suggested that Gly-N₃ probe was either not effective at infiltrating metabolic pathways to label glycerolipids or that cell entry of this compound was unsuccessful.

On the contrary, cellular fluorescence images resulting from C_n-MAG-N₃ probe treatment showed widely variable signal strength and localization depending upon the length of the acyl chain. In particular, C₄-MAG-N₃ primarily resulted in fluorescence that was localized at the periphery of cells (Figure 3A). This was noted as a positive since our goal was to label glycerolipids, many of which are localized at cellular plasma membranes. It is important to note that the localization of the observed signal may be impacted by the accessibility of the dye reagent (DBCO-Cy3), which could enhance labeling of the plasma membrane as the first point of contact. It is also possible that lipids could remain mobile after fixing cells, which could also impact localization. Nevertheless, the results show that C₄-MAG-N₃ is effective at labeling cells. On the contrary, probes with longer acyl chains yielded variable fluorescence patterns. For MAG probes bearing intermediate length chains (C₁₀-MAG-N₃, C₁₂-MAG-N₃), we instead observed diffuse cytosolic staining (Figure 3C, D). We speculated that this signal profile may result from cell death, since it is characteristic for the plasma membranes of dead yeast cells to rupture, which may lead to accumulation of excess dye.^[33] We investigated this possibility further via cytotoxicity experiments that will be described in the next section. Interestingly, further increasing of chain length in the probe C₁₄-MAG-N₃ (Figure 3F) culminated in a return to fluorescence localized at the cell periphery, although the resulting signal was rather weak compared to that resulting from C₄-MAG-N₃ treatment. Probe C_18:1-MAG-N₃ instead resulted in fluorescent clumps within cells that we will explore further in future experiments (Figure 3G).

Figure 3.

Fluorescence imaging of cell labelling using clickable C_n-MAG-N₃ probes and C₄-MEG-N₃. S. cerevisiae cells were treated with 1.0 mM of either C₄-MAG-N₃ (A), C₄-MEG-N₃ (B), C₁₀-MAG-N₃ (C), C₁₂-MAG-N₃ (D), C₁₄-MAG-N₃ (F), C_18:1-MAG-N₃ (G), or ethanol (H) from the initial OD₆₀₀ of 0.2 and incubated for 12 hrs. After removal of media and washing steps, cells were then fixed and click-labeled with 1 μM DBCO-Cy3, washed to remove unreacted dye, and subjected to confocal microscopy. In each panel, the left image shows fluorescence while the right side provides a bright field image. Images are representative of n = 180 fluorescing cells from 6 biological replicates. Scale bars indicate 5 μm. (E) Quantification of fluorescence intensity (Mean fluorescent intensity, MFI) for C₄-MAG-N₃ and C₄-MEG-N₃ by flow cytometry with clickable dye DBCO-AF647. Error bars represent standard errors from 6 biological replicates. Significance was determined by Brown-Forsythe and Welch’s ANOVA test (*** p = 0.0001; **** = p <0.0001).

Based on the promising labeling properties we observed for C₄-MAG-N₃, we next compared this to imaging resulting from ether-linked derivative C₄-MEG-N₃. We expected that the ether-linked chain at the sn-3 position in the latter probe would disable cleavage by esterase enzymes, which we hypothesized would preclude metabolic labeling of phospholipid biosynthesis due to the unavailability of the sn-3 hydroxyl group for phosphorylation and headgroup introduction. We found that confocal images resulting from C₄-MEG-N₃ exhibited signal localization that was similar to C₄-MAG-N₃ but appeared to be significantly less intense (Figure 3B). This difference was quantified by flow cytometry, which confirmed diminished fluorescence using C₄-MEG-N₃ probe compared to C₄-MAG-N₃ (Figure 3E). It should be noted that due to instrumental wavelength filters, these experiments were performed via SPAAC with a different clickable reagent (DBCO-AF647). The results from these experiments provided initial evidence in line with our hypothesis by suggesting that the probe C₄-MEG-N₃ is less effective at metabolic labeling compared to C₄-MAG-N₃ probe, which will be discussed further through subsequent TLC experiments. We also quantified fluorescence signal for long chain C_n-MAG-N₃ probes (Figure S2).

Since we had speculated that certain probes were affecting cell viability, we next moved to study cytotoxicity effects of probes that produced fluorescent signal. To do so, S. cerevisiae cells were grown in galactose media in the presence of 1 mM of each probe and cell density (OD₆₀₀) was measured at 0, 12, 14, and 24 hour time points. We observed that MAG probes with medium chain fatty acid tails (C₁₀-MAG-N₃, C₁₂-MAG-N₃), which had resulted in intensive cytoplasmic fluorescence signal and exhibited significantly higher fluorescence compared to control (Figure S2), indeed abrogated cell growth since OD₆₀₀ maintained at lag phase around 0.2 or 0.4 (Figure S3A). This phenomenon is in line with prior literature reports that S. cerevisiae cells are known to undergo a pro-apoptotic mechanism upon exogenous treatment with MAGs with medium acyl chain length.^[34],[35] Otherwise, the C₄-MAG-N₃, C₁₄-MAG-N₃, C_18:1-MAG-N₃, and C₄-MEG-N₃ analogs enabled log-phase cell growth that was in range of no probe control cells (NPC). Moreover, the calculated doubling times (Dts) from the growth curves (14 hrs) for C₄-MAG-N₃ and C₄-MEG-N₃ were not significantly different from that of untreated cells (Figure S3B).

We next set out to confirm glycerolipid labeling and investigate the lipid classes that are labeled by the C₄-containing probes, which we selected for further analysis based on their favorable imaging properties. We initially performed thin-layer chromatography (TLC) analysis by extracting lipids from cells incubated with either C₄-MAG-N₃ or C₄-MEG-N₃, grown for 12 or 24 hrs, and subjecting lipid extracts to fluorescence labeling via copper-catalyzed azide-alkyne cycloaddition (CuAAC) with ethynylnaphthalimide “naphth”, a nonpolar fluorogenic reagent that minimally impacts R_f of lipids, followed by analyzing via TLC separation (Figure 4). Because these probes are expected to label both non-polar (i.e. TAG/DAG) and polar (i.e. glycerophospholipid) products (see summary of potential products in Figure 6), we conducted TLC analysis to separate and visualize labeled products using a two-step elution method through a combination of polar and then non-polar eluants. As shown in Figure 5A, C₄-MAG-N₃ treatment led to the appearance of fluorescent bands seen in fluorescent images that appeared close to commercial standards for both non-polar lipids (MAG, R_f = 0.79 and DAG, R_f = 0.84–0.92, non-polar) as well as two phospholipid classes (PC, R_f = 0.34 and PS, R_f = 0.28). New spots correlating with PC and PS were only observed when cells were harvested at log phase, which indicates that labeled lipids undergo rapid turnover. The R_f values for commercial standards of the appropriate lipids are listed on the left side of Figure 5, with arrows pointing toward their position. The R_f values for new spots are shown on the right side of each TLC image. These new spots appear slightly lower than commercial standards, which can be explained since these probes are modified via CuAAC with naphth dye. Treatment with C₄-MEG-N₃ led to similar high R_f spots that were comparable to commercial standard for DAG (R_f = 0.84–0.92, lanes 6–7, Figure 5B).

Figure 4.

General experimental outline for fluorescence-based TLC and MS analysis via CuAAC. S. cerevisiae cells were incubated with 1.0 mM C₄-MAG-N₃ or C₄-MEG-N₃ for 12 hrs, harvested, washed, and subjected to a lipid extraction and CuAAC protocol with either naphth dye for TLC separation and visualization or with an alkyne quaternary ammonium reagent, AlkBuQA, for LCMS analysis.

Figure 6.

A simplified representation of lipid biosynthetic transformations in S. cerevisiae indicating potential glycerolipid products from C₄-MAG-N₃ or C₄-MEG-N₃ probe. Synthetically added tails are in red, biosynthetically added FA tail of varying lengths (m) are in black, and the methylene azide tag is in blue. A) Acyltransferases can convert C₄-MEG-N₃ into “DAG” and “TAG”. Note that the C₄-ether is counted as one tail (ie: C₄-ether + fatty acyl tail = “DAG”). B) Labeled nonpolar lipids can be produced from acyltransferase activity from C₄-MAG-N₃. C) Alternatively, labeled neutral lipids can be accessed from the glycerol-3-phosphate (G-3-P) pathway. From C₄-MAG-N₃, MAG lipase/esterase activity can result in glycerol-N₃ (Gly-N₃) and free fatty acid (FFA). Then, through the cytidine diphosphate diacylglycerol (CDP-DAG) pathway, labeled phospholipids can be produced. Note that the pathways illustrated do not encompass all lipid transformations.^{[26, 30a]}

Figure 5.

TLC image showing C₄-MAG-N₃ labels neutral lipids and phospholipids while C₄-MEG-N₃ labels only neutral lipids. Lipids were extracted and subjected to CuAAC reaction with ethynylnaphthalimide “naphth”. Clicked extracts were loaded onto a TLC plate for lipid separation by elution with chloroform/methanol/water/acetic acid and then cyclohexane/ethyl acetate and then analyzed by fluorescence imaging. A) Lane 1 contained no probe control lipid extracts. Lane 2 was loaded with a stock of C₄-MAG-N₃ that was clicked with naphth. Lanes 3 and 4 were naphth-clicked lipid extracts from yeast incubated with C₄-MAG-N₃. Lane 3 contained lipids isolated from 24-hr incubation while lane 4 contained lipids obtained from 12-hr incubation. The R_f values and positions of lipid standards for DAG (16:0/16:0), MAG (18:1), and different phospholipids are indicated with arrows on the left side of the figure (viewed after primulin staining). R_f values for new spots are shown on the right side of each image. B) TLC plate representing click-derived fluorescence from lipid extracts when S. cerevisiae cells were incubated with 1.0 mM C₄-MEG-N_3. Lane 5 was loaded with a stock of C₄-MEG-N₃ that was clicked with naphth. Lanes 6 and 7 contained lipids from 24-hr and 12-hr incubation, respectively.

In contrast, there were no clear bands corresponding to polar lipids (R_f < 0.60) resulting from cells grown with C₄-MEG-N₃. These results support our hypothesis that C₄-MAG-N₃ would be effective for labeling phospholipids while this activity would be deactivated for C₄-MEG-N₃ by irreversibly blocking the sn-3 position where the phospholipid headgroup would reside.

While probe C₄-MAG-N₃ produced TLC bands that correlate with phospholipids, it is important to note that these spots were faint, particularly in comparison to bands for labeled products that overlap with neutral lipids. This suggests that C₄-MAG-N₃ is less effective as a substrate for enzymes that produce phospholipids, perhaps because the added azidomethylene tag is problematic for these particular enzymes. On the contrary, probe C₄-MEG-N₃ can be seen as advantageous since our data support that this probe bypasses the labeling of phospholipids altogether but retains robust labeling of neutral lipids. Therefore, this strategy is expected to be beneficial for focusing metabolic labeling within the family of neutral lipids including DAG and TAG products.

To further characterize the identities of potential labeled products (see Figure 6), we additionally performed MS analysis after whole-cell lipid extraction. For this purpose, azide-tagged lipids were post-derivatized through CuAAC reaction with N-ethyl-N,N-dimethylhex-5-yn-1-aminium (AlkBuQA, Figure 4) to enhance LCMS detection, as described in previous reports,^[16b],[36] followed by LCMS analysis. Cells treated with C₄-MAG-N₃ yielded mass peaks corresponding to a range of labeled lipids including both phospholipids (Table S1; PA, PS, PE, PC; representative mass spectra shown in Figure S4) and neutral lipids (Table S2; MAG, DAG, and TAG; representative mass spectra shown in Figure S5). In our analysis, while we detected products resulting from the click reaction with AlkBuQA, masses corresponding to unclicked lipids (mN₃) were still detected from the same samples. Labeled lipids containing the azido-methylene tag were detected in both positive and negative ESI conditions. On the contrary, C₄-MEG-N₃ treatment was found to result in mass peaks correlating with labeled neutral lipids (Table S3; DAG and TAG; representative mass spectra shown in Figure S6), but not phospholipids. In all cases, mass peaks were not observed for untreated/NPC samples. These results further support our expectation that C₄-MAG-N₃ labels a broad range of neutral lipids and phospholipids while blocking the sn-3 position of C₄-MEG-N₃ would deactivate phospholipid labeling.

Conclusion

In this work, a series of azide-tagged MAG probes were prepared and assessed for their lipid metabolic labeling properties. While Gly-N₃ did not exhibit increased fluorescence compared to control, we found that each C_n-MAG-N₃ probe resulted in a distinct fluorescence pattern and variable cytotoxicity. The short chain C₄-MAG-N₃ probe in particular was found to successfully infiltrate lipid metabolism and result in fluorescence at the plasma membrane. Labeled lipid classes were identified using TLC separation and further characterized down to lipid species resolution using LCMS. In line with our expectations, probe C₄-MAG-N₃, resulted in the labeling of both neutral lipids and phospholipids (albeit the latter less effectively), while our data show that C₄-MEG-N₃ exclusively labeled neutral lipids, which we attribute to blocking of the sn-3 position. Overall, this work showcases the potential of broadly labeling glycerolipids in cells using probe C₄-MAG-N₃, while probes of type C₄-MEG-N₃ could be particularly advantageous for focusing the metabolic labeling of neutral lipids. Additionally, we have shown that the portfolio of labeled lipids can be controlled through subtle modifications to the structures of probes that act as mimics of biosynthetic precursors. We expect these probes to be valuable chemical tools for studying lipid remodeling and discerning how lipid metabolism shifts in cells in response to different stimuli and biological processes.

Experimental Section

Materials

All lipid standards, L-α-phosphatidylcholine (mixed isomers from chicken egg), L-α-phosphatidic acid (sodium salt from chicken eggs), and 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine, were purchased from Avanti Polar Lipids, Inc. NMR spectra were obtained using Varian Mercury 300, 500, and 600 MHz spectrometers. Mass spectra were obtained with a JEOL DART-AccuTOF mass spectrometer or a Waters Synapt G2-Si electrospray ionization mass spectrometer with a quadrupole-time-of-flight mass analyzer (Milford, MA). Dibenzocyclooctyne-cyanine (DBCO-Cy3) was obtained from Sigma Aldrich. AFDye^™ 647 DBCO (DBCO-AF647) was purchased from Click Chemistry Tools. 4-Ethynyl-N-ethyl-1,8-napththalimide was synthesized as previously reported.^[37] Thin-layer chromatography (TLC) glass plates, pre-coated 0.25 mm silica gel without fluorescent indicator (20×20cm) were purchased from Sorbent Technologies (Norcross, GA).

Methods

Yeast Strains and Culture Conditions:

S. cerevisiae TRY 181 (wild-type, uraΔhisΔ) cells were streaked on YPD (1% yeast extract, 2% peptone, 2% dextrose, 2% agar) plates and grown at 30 °C. A loop of cells was inoculated into fresh 2% synthetic medium as the preculture. 2% minimal media (YNB) was prepared by combining the following: 6.7g of yeast nitrogen base (w/o AAs), 20 mg uracil, 20 mg L-histidine, and 20 g galactose dissolved in 1 L Milli-Q water and filtered via a pre-sterilized Millipore vacuum filtration system. All cells were grown in liquid medium at 30 °C in a 225 rpm shaker. After ~18 h, cells from the seed culture were diluted back to an initial OD₆₀₀ of 0.2 in fresh media and supplemented with either Gly-N₃, C_n-MAG-N₃, C₄-MEG-N₃ or ethanol in an equivalent concentration (5 mL total volume with growth media) until harvest. For growth curve studies, an aliquot was taken at designated timepoints for OD measurement (Figure S3). For in vivo metabolic labeling, cells were harvested at 12 hours (12 and 24 hours for TLC) by centrifugation (3000 × g for 5 min at 4–0°C) to remove growth media. After removal of growth media, the cells were washed three times with Milli-Q water (pelleted at 3000 × g, 5 min for each wash). MAG-N₃ probe stocks were kept in ethanol (100 mM or 50 mM) so MAG-N₃ addition into 5 mL cultures did not exceed 10% by volume to prevent leaky membranes and to ensure no detrimental effects on the growth of S. cerevisiae.^[38]

SPAAC and Fluorescence Microscopy:

Washed cells were fixed for 15 min (3.7% PFA in PBS). Fixed cells were washed with 1x PBS and resuspended in 5 mL 1x PBS. 1 mL cell suspension (diluted to an OD₆₀₀ of 0.60/mL) was transferred to a fresh tube and incubated with DBCO-Cy3 clickable dye (1 μM final concentration in Milli-Q) for 1 hr at rt. The cells were finally washed with Milli-Q, vortexed, and spun down (5000 ×g) a total of three times. Finally, resuspension in 1mL 1x PBS (pH = 7.4) gave clicked yeast samples. Cell culture samples (3 μL) were mounted onto microscope slides and covered with a coverslip, immobilized with acrylic nail polish and visualized under a confocal microscope (Leica SP8 White Light Laser Confocal microscope) using a 63x oil objective. The samples were imaged by differential interference contrast (DIC) to first locate the cells and a zoom factor of 6 or 8 was applied to properly view the yeast cells. Then, samples were excited at 554 nm and fluorescence images were collected between 559–620nm with a HyD detector. The laser strength, gain, and offset settings were kept constant between samples. All images acquired on S. cerevisiae cells were taken as a Z-stack of 7–14 z-slices with 1.0μm increments, after which lightning deconvolution was applied through Leica Application Suite (LAS) V4.4 software. Images were processed the same way using FIJI (v.2.1.0/1.53c).

Quantification of Fluorescence Intensity by Flow Cytometry:

Samples for flow cytometry were first fixed (3.7% PFA in PBS), incubated with AFDye^™ 647 DBCO (2 μM in from 0.11 mM stock in Milli-Q), and washed in similar fashion as samples for fluorescence microscopy. After washing and centrifugation, cells were diluted to an OD₆₀₀ of 0.06 or 1.1 × 10⁶ cells/mL, centrifuged (200 × g for 3 min), and the resulting pellets were resuspended in 1 mL of 1X PBS buffer and kept on ice for flow cytometry with FACSCalibur LSR II flow cytometer (Becton Dickinson). After exclusion of debris (FSC-A vs SSC-A plot) and gating for single cells (FSC-W vs FSC-A), AFDye^™ 647 DBCO fluorescence intensity was recorded in the APC channel. Flow cytometry data were obtained for 50,000–60,000 gated events per sample. For unstained controls, probe-supplemented yeast samples were only fixed and not incubated with click-dye and were prepared each run for a representative negative population. Biological replicates were analyzed on the same day and analysis was performed using FlowJo software (v.10.11 FlowJo LLC, OR, USA). For flow cytometry experiments, samples were also viewed under the confocal microscope, excited at 633 nm, and fluorescence was collected between 651–671nm with a HyD detector. Statistical analysis was performed on Graphpad Prism 9.1.

Lipid Extraction for Mass Spectrometry and Fluorescence-based Thin Layer Chromatography:

The lipid extraction method followed a prior procedure with minor modification.^[39] Log phase samples (15 mL volume) were placed on ice and cold trichloroacetic acid (TCA) solution was added to a final 5% concentration for an hour incubation. Growth media and TCA suspension were removed and then cells were centrifuged and washed 3 times with MQ water. Cells were either used directly for lipid extraction or stored at −20 °C. Cells warmed to rt were resuspended in 1 mL lipid extraction mix of (15/15/5/1/0.018) 95% EtOH/H₂O/diethyl-ether/pyridine/4.2 N NH₄OH aqueous solution. Samples were vortexed for 1 minute and then incubated for 15 minutes at 55°C. This extraction was repeated once more for a total of two times. Collected extracts were dried with a steady N₂ stream. Lipid extracts were resuspended in 16 μL CHCl₃. For the preparation of the 8.60 mM Cu(I) click solution, 3.80 mg Cu(I)(MeCN)₄BF₄ was dissolved in 605 μL acetonitrile and 800 μL of ethanol. 200 μL of this copper solution was combined with 10 μL of N-ethyl-N,N-dimethylhex-5-yn-1-aminium or AlkBuQA (100 mM in MeOH) or 10 μL of 4-ethynyl-N-ethyl-1,8-naphthalimide dye “naphth” (22.5 μM in CHCl₃/MeOH) as the click mix and vortexed for 30 s. The click mix (30 μL) was added to lipid extracts (suspended in 16 μL CHCl₃), as well as to aliquots of unmetabolized C_n-MAG-N₃ probe stocks (0.5 mM). The clicked samples were flushed with N₂, vortexed, spun down, and then incubated on a 42°C water bath for 16 h. After CuAAC incubation, the clicked samples were vortexed and spun down (4,600 × g). For mass spectrometry, samples were diluted with 30 μL of CHCl₃/MeOH/H₂O (73:23:3) prior to LCMS. For TLC, clicked samples were loaded onto a thin-layer chromatography (TLC) plate in 5 μL increments (40 μL total lipid extracts), with drying in between each addition. Lipid samples were separated by TLC using glass plates of pre-coated 0.25 mm silica gel without fluorescent indicator (20 × 20 cm). The spots were dried with a heat gun prior to the addition of phospholipid standards: 2–4 μL of PS, PC, PA, DOPE, MAG and DAG standards (3–5 mM in CHCl₃/MeOH). The plate was once again dried with a slow stream of nitrogen and then developed in a solvent mixture of chloroform/methanol/water/acetic acid (65:25:4:1, v/v/v/v) until an Rf or 0.6. The plate was air-dried for an hour and then developed with 1:1 cyclohexane/ethyl acetate.^[40] Next, the TLC plate was imaged using a Geldoc with a SYBR Green emission filter (grayscale images not shown) processed using ImageJ (LUT inverted). The “click” spots (Figure 5) were marked with a lead pencil to differentiate later from the spots due to primulin staining. The plate was dipped in primulin stain (5 mg primulin per 100 mL 9:1 acetone/water) and then imaged again (primulin-stained image not shown).

Ultra-High Performance Liquid Chromatography High Resolution Mass Spectrometry (UHPLC-HRMS)^[41] and Analysis:

An UltiMate 3000 ultra-high performance liquid chromatography system (UHPLC, Dionex, Sunnyvale, CA) was used to inject 10 μL of sample onto a CORTECS C18 column (90 Å, 2.7 μm, 2.1 mm × 150 mm; Waters) controlled at 40◦C. Mobile phase A was 60:40 acetonitrile/water with 10 mM ammonium formate as a buffer and 0.1% formic acid while mobile phase B consisted of 90:10 2-propanol/acetonitrile with 10 mM ammonium formate as a buffer and 0.1% formic acid. The gradient used follows: t = 0 min: 32% solvent B flow rate of 0.30 ml/min, t = 1.5 min: 32% solvent B flow rate 0.4 ml/min, t = 2.5 min: 45% solvent B flow rate 0.4 ml/min, t = 5 min: 52% solvent B flow rate 0.3 ml/min, t = 8.0 min 58% solvent B flow rate 0.3 ml/min, t = 11.0 min 68% solvent B flow rate 0.4 ml/min, t = 14.0 min 75% mobile phase B flow rate of 0.4 ml/min, t = 18.0 min 80% solvent B flow rate of 0.4 mL/min, t = 21 98% solvent B flow rate 0.45 ml/min, t = 25.0 min 32% solvent B flow rate 0.3 ml/min until equilibration at 30 min. Eluent was introduced to the mass spectrometer via an electrospray ionization (ESI) source, with the following parameters: sheath gas 30 (arbitrary units), aux gas 8 (arbitrary units), sweep gas 3 (arbitrary units), spray voltage 3 kV, capillary temperature 300°C. Mass analysis was performed using an Exactive Plus (Thermo Scientific, Waltham, MA) mass spectrometer operated in dual polarity mode. Masses were detected in full scan mode within a scan range of 100–1,500 m/z, operated at a resolution of 140,000, with an automatic gain control target of (AGC) of 3 × 106 ions, and a maximum IT time of 100 ms. Full scan data was complemented with all ion fragmentation data at a resolution of 35,000 utilizing 35 eV collisional energy. Raw files were converted to mzML files using MS convert and full scan data was evaluated using El MAVEN software. For peak identification, retention times of labeled lipids were compared with respect to natural lipid of the same lipid species. Labeled lipids (from a generated compound list) were identified (C12 isotope as parent peak) based on their accurate mass (±5 ppm).