Cobalt carbonyl complexes as probes for alkyne-tagged lipids

Abstract

Monitoring lipid distribution and metabolism in cells and biological fluids poses many challenges because of the many molecular species and metabolic pathways that exist. This study describes the synthesis and study of molecules that contain an alkyne functional group as surrogates for natural lipids in cultured cells. Thus, hexadec-15-ynoic and hexadec-7-ynoic acids were readily incorporated into RAW 264.7 cells, principally as phosphocholine esters; the alkyne was used as a “tag” that could be transformed to a stable dicobalt-hexacarbonyl complex; and the complex could then be detected by HPLC/MS or HPLC/UV_349nm. The 349 nm absorbance of the cobalt complexes was used to provide qualitative and quantitative information about the distribution and cellular concentrations of the alkyne lipids. The alkyne group could also be used as an affinity tag for the lipids by a catch-and-release strategy on phosphine-coated silica beads. Lipid extracts were enriched in the tagged lipids in this way, making the approach of potential utility to study lipid transformations in cell culture. Both terminal alkynes and internal alkynes were used in this affinity “pull-down” strategy. This method facilitates measuring lipid species that might otherwise fall below limits of detection.

The exploitation of reactions that are orthogonal to common functional groups encountered in biology has led to the development of probes for molecules that prove useful in a variety of biomedical applications. The azide group is particularly useful as a tag for a molecule of interest since it does not react with most functional groups common to cells and biological fluids while its reactions with phosphines (1–3) and alkynes (4, 5), Huisgen dipolar cycloaddition (6), permit probing for a specific molecule in a complex chemical environment (7–9). The alkyne functional group itself is not common in mammalian biology, and we have employed it as an affinity-tagged surrogate for the small diffusible electrophile 4-hydroxy-2-nonenal (4-HNE). A triple bond differs from a saturated molecule by only four hydrogen atoms, and its introduction as a tag into a probe molecule could, under the right circumstances, represent a minimal structural perturbation to a natural or biologically active compound. Furthermore, reaction of the alkyne-tagged molecule with an azide reagent could then be used to develop an alkyne affinity isolation strategy based on an inverse click approach; i.e., the alkyne serves as the tag, and the azide as the pull-down reagent.

4-HNE, a remnant of (ω-6) polyunsaturated fatty ester peroxidation, reacts with proteins and nucleic acids to give covalent adducts, and the pathology associated with modification of these important biomolecules has been of widespread interest (10). An alkynyl-substituted analog has equivalent cellular toxicity as 4-HNE itself and this analog, alkynyl-4-HNE (a-4-HNE), modifies proteins by reaction with nucleophilic centers. The modified proteins can be isolated from complex biological mixtures by ligation with an azido biotin pull-down reagent, enriching those proteins and peptides that have been modified (11, 12).

The success of the alkynyl-tag approach with 4-HNE led us to the design and study of alkynyl-tagged lipids such as those shown in Fig. 1 (13). Several natural lipid analogs modified with an alkynyl functional group were synthesized with the intention that these compounds could be used as lipid surrogates in cells, fluids, and tissues. The utility of cobalt octacarbonyl alkyne complexation (14–17) has also been demonstrated as a means for immobilization (18–20) and enrichment of compounds bearing the alkyne tag, making the alkyne functional group a potential “tag” for tracking lipids in biology as shown in Scheme 1 (13). We report here that the alkyne cobalt complexes themselves can be used as reporter tags for lipid classes and molecular species, making quantitative analysis of individual molecules or lipid classes straightforward. We also demonstrate that the strategy is not limited to a terminal alkyne, as is the case for “click” cycloaddition, expanding the variety of structures that can be used as surrogate lipids.

Fig. 1.

Alkynyl surrogates of natural lipids.

SCHEME 1.

Alkyne affinity strategy for terminally or internally tagged lipids.

MATERIALS AND METHODS

Materials

Commercial anhydrous CHCl₃ was used as received. Purification by column chromatography was carried out on silica gel, and TLC plates were visualized with phosphomolybdic acid. 2-Diphenylphosphinoethyl functionalized silica gel was purchased from Aldrich chemicals. Co₂(CO)₈ was obtained from Strem Chemicals.

Synthetic procedures

¹H, ¹³C, and ³¹P NMR spectra were collected on a 300 MHz NMR. The syntheses of di-15a-PPC and its corresponding lyso-15a-PC have been previously described (13). The detailed synthetic procedures for the preparation of the lipids shown in Fig. 1 are presented in the supplementary data.

Alkynyl lipid cellular incorporation and enrichment

Incorporation into RAW cells.

RAW 264.7 cells were obtained from the American Type Culture Collection and maintained in Dulbecco's Modified Eagle Medium (DMEM) supplemented with high glucose and pyridoxine hydrochloride (Gibco, Grand Island, NY) containing 10% heat-inactivated fetal bovine serum (Hyclone, Logan, UT) and 1% penicillin streptomycin (Gibco, Grand Island, NY). Cells were plated onto 60 mm tissue culture dishes (Corning Inc., Corning, NY) at 2 × 10⁶ cells/ dish and incubated in a humidified chamber at 37°C/5% CO₂ for either 3 h (in the case of fatty acid labeling) or 24 h (in the case of glycerophospholipid labeling) to allow cells to adhere to the plates. Plating media and nonadherent cells were removed and replaced with freshly sonicated labeling media [50 µg/ml 7a-16:0 or 15a-16:0, or 2 mg/ml di-15a-PPC in DMEM containing 10% heat-inactivated fetal bovine serum, 1% penicillin streptomycin, and 0.25 mg/ml fatty acid- and endotoxin-free BSA (Sigma-Aldrich, St. Louis, MO)]. Cells were incubated with 3 ml labeling media per dish for either 20 h (in the case of fatty acid labeling) or 5 h (in the case of glycerophospholipid labeling) in a humidified chamber at 37°C/5% CO₂. After labeling, each plate of cells was washed twice with 5 ml DPBS (Mediatech, Inc., Manassas, VA), then cells were scraped into 600 µl ice-cold 0.1N HCl:MeOH (1:1) and transferred to microcentrifuge tubes (Laboratory Product Sales, Rochester, NY). Cellular lipids were extracted by the addition of 300 µl ice-cold CHCl₃ to the scraped cells, vortexed for 5 min at 4°C, and then centrifuged at 4°C for 5 min at 18,000 g. The extracted lipids were stored at −20°C until analysis.

Alkyne cobalt octacarbonyl complexation and enrichment.

Stock solutions of the alkynyl fatty acids or phospholipids were made up in methanol (1 mM). Samples (20 µl of each stock solution) were prepared in microcentrifuge tubes equipped with a stir bar and diluted to 200 µl with methanol. An aliquot (50 µl) of the initial sample was reserved for analysis. To the remaining sample was added Co₂(CO)₈ (5 mg) followed by stirring at room temperature for 30 min. The solution should be dark orange-brown at this stage of the reaction. 2-Diphenylphosphinoethyl-functionalized silica gel (50 mg) was added to the solution and heated at 55°C with vigorous stirring. After 2 h, the slurry was transferred to a spin-tube, a microcentrifuge tube fitted with a removable 0.45 µm nylon filter. The reaction tube was washed with methanol (200 µl) and added to the spin-tube. Then the sample was centrifuged at 13,200 g for 5 min. The silica gel was washed with additional methanol for phospholipids or 1% HOAc/methanol for fatty acids (200 µl × 2), vortexed, and centrifuged. This filtrate was reserved as the wash sample. The filter containing the silica gel was transferred to a new microcentrifuge tube. Methanol (400 µl) and Fe(NO₃)₃ (40 mg) was added and stirred at room temperature for 30 min. The sample was centrifuged at 13,200 g for 5 min, then the silica was washed with additional methanol or acidic methanol (200 µl) as described above. This sample is referred to as the release. See Scheme 1 for the process.

Quantitation based on cobalt complex 349 nm UV.

The CHCl₃ layer from a Bligh and Dyer cellular extraction was removed and transferred to a new microcentrifuge tube. The sample was concentrated on a SpeedVac and resuspended in methanol (200 µl). To the sample was added 7a-palmitic acid (100 nmol) as an internal standard and Co₂(CO)₈ (5 mg). After the reaction mixture had stirred at room temperature for 30 min, it was analyzed by HPLC with UV detection at 349 nm. The area under curve for the di-15a-PPC-cobalt complex was corrected by a factor of two, taking into account it has two alkynes. For mass spectral quantitation, a sample was equally divided and half subjected to UV analysis of the cobalt complex as described. The other half was spiked with 33:0 plasmanyl PC (33:0 PCe) (20 nmol) as an internal standard and analyzed by LC/MS (see the supplementary data). A response factor was determined using authentic standards of di-15a-PPC and 33:0 PCe. Phospholipids were detected as their M+H adduct. Dimers of the phospholipids were also detected and were included in the integration. For both quantitation methods, the reaction mixture was analyzed on an Ascentis C8 column using a mobile phase consisting of A (5 mM NH₄OAc in H₂O) and B (5 mM NH₄OAc in methanol). The phospholipids were eluted with a gradient of 80 to 100% B over 10 min, held for 15 min, and back to 80% B over 5 min. Mass spectral analysis was carried out on an ion trap in positive ionization mode with the following ESI source parameters: capillary temperature 200°C, capillary voltage 17 V, spray voltage 4.5 kV, tube lens offset 15 V, and sheath gas (nitrogen) 20.

Alkynyl lipid enrichment cells.

The CHCl₃ layer from the Bligh and Dyer extraction was removed and transferred to a new microcentrifuge tube. Two samples were combined, concentrated, and resuspended in methanol (200 µl). The enrichment of the di-15a-PPC was carried out as described above. Prior to mass spectral analysis, the excess Fe(NO₃)₃ was removed from the sample. The release sample was concentrated, then resuspended in 400 µl CHCl₃, 200 µl methanol, and 200 µl 0.1 N HCl. The sample was vortexed for 1 min and centrifuged at 13,200 g for 1 min to separate the layers. The CHCl₃ layer containing the phospholipids was removed and concentrated. The sample was resuspended in 200 µl methanol and analyzed by LC/MS using the conditions described above. Detailed procedures for separation and quantitation of lipid classes along with examples of alkynyl lipid enrichment from cell culture are presented in the supplementary data.

RESULTS AND DISCUSSION

Cobalt complex formation and characterization

Conditions suitable for use in enriching alkyne-tagged lipids from cells were sought so that organelle distribution and metabolism studies could be carried out with tagged precursors. In general, cobalt complexes were formed by addition of Co₂(CO)₈ to a solution of the alkyne in methanol. The cobalt reagent itself is unstable and readily oxidizes upon exposure to air, evident by a color change from orange to purple. In contrast, the cobalt-alkyne complexes are stable as long as they are not exposed to oxidative conditions. Complexes form rapidly at room temperature, generally within 10 min, producing an orange solution. Although methanol is the preferred solvent for lipid substrates, acetonitrile, methylene chloride, isopropanol, and tetrahydrofuran were used in some cases for cobalt complexation. Minimal amounts of water could be used with methanol as a cosolvent, but when it exceeded 20%, significant precipitation of Co₂(CO)₈ occurred and complexation to the alkyne failed. For assays on alkynyl lipids in cell culture, cells were washed to remove media and scraped into acidic methanol, and the lipids were extracted using a modified Bligh and Dyer protocol (21, 22).

The formation of the cobalt-alkyne complexes could be readily monitored by HPLC/UV. The peak due to the alkyne disappears after the Co₂(CO)₈ treatment and a much less polar compound corresponding to the cobalt-alkyne complex generally appears in the chromatogram. The cobalt-alkyne complexes have a characteristic absorbance at 349 nm with a molar absorption coefficient found to be ∼3,700 M⁻¹ cm⁻¹ for several of the alkyne-cobalt complexes described here. Thus, monitoring the chromatogram at 349 nm immediately identifies the complex and permits a quantitative analysis of that species based upon ϵ. Some of the alkyne complexes could be isolated and further characterized by NMR analysis. This generally showed a downfield shift of acetylenic protons from 2.1 to 6.3 ppm, and a more modest shift of propargylic protons from 2.1 to 2.9 ppm (see supplementary Fig. I).

HPLC/MS proved to be a useful tool for the analysis of complex lipid mixtures containing alkynyl-tagged lipids. The intact alkynyl-cobalt phospholipid complexes produced strong molecular ions on ion-trap mass spectrometers, whereas analysis on quadrupole instruments gave inconsistent results. With trifluoroacetic acid or ammonium acetate additive, a strong [M+H]⁺ peak was observed in the positive mode. Some fragmentation corresponding to sequential CO loss was also detected (Fig. 2). The [M+OAc]⁻ adduct could also be seen in negative mode. For phospholipid alkynes, the cobalt complexes were less readily detected than the parent alkynes, mass spectral analysis being less sensitive for the bis-dicobalt complex compared with the mono-complexed species which was, in turn, observed at lower signal than the parent phospholipid. Increasing the number of alkynes, and hence the number of cobalt species bound to the molecule, results in lower signal intensity.

Fig. 2.

HPLC/MS analysis of a mixture of di-15a-PPC and its mono- and bis-dicobalt hexacarbonyl complexes. The top panel shows the total ion current of the mixture, and the other panels show the mass spectrum (M+H) of each compound. Analysis was carried out with C8 HPLC using a H₂O/MeOH gradient with trifluoroacetic acid as additive.

Cobalt complex ligation to solid support and release

Phosphines are good ligands for alkyne-cobalt complexes, and phosphines on polystyrene (18), Novagel (20), and silica gel were examined as potential solid supports for ligation of the alkynyl lipid cobalt complexes. For small nonpolar alkynes, polystyrene-phosphine beads worked well. The solid-supported phosphine was added to a solution of the cobalt-alkyne complex in methanol and heated at 70°C with gentle stirring under argon for an hour, and under these conditions, immobilization on beads was essentially complete. Polystyrene beads were problematic, however, for studies with polar lipids, and the silica-supported phosphine was ultimately used in most of our studies.

For immobilization of alkyne lipid analogs on the silica-supported phosphine, the reaction could be carried out at 55°C in methanol. This lower temperature eliminated transesterification of the phospholipids, a reaction that was sometimes observed at higher temperatures and evident by the formation of fatty acid methyl esters. In addition, there was significantly less nonspecific binding of the phospholipids to the silica support as compared with polystyrene. Capture of fatty acids and phospholipid alkynes was nearly complete on silica-phosphine beads after an hour at 55°C; terminal alkynes appeared to react somewhat faster than their internal alkyne counterparts. Although the number and position of the alkyne in a substrate has some influence on the rate of immobilization, all phospholipids studied were successfully captured within 2 h at 55°C. Capture could also be achieved at 37°C for 5 h or overnight at room temperature.

After immobilization on a solid support, beads were washed to remove compounds not bearing the alkyne/cobalt functionality, and the parent alkyne was released from the support by mild oxidation of the cobalt complex. Many oxidants were screened to find conditions that resulted in the highest recovery of the alkyne. Trimethylamine oxide, ceric ammonium nitrate, air, and HCl resulted in the release of 30–50% of alkyne from solid support, whereas KSCN, ^tBuOOH, and tetrabutyl ammonium fluoride (TBAF) failed in our hands to give measureable free alkyne. Ferric nitrate produced the best results by far (14), with 70–80% recovery of the alkyne. Upon immobilization of the cobalt-alkyne complex on support, washing with methanol and then release using Fe(NO₃)₃ generally provided the best results. If nonprotic solvents were used in the ferric nitrate oxidation reaction, alkene Z→E isomerization of unsaturated fatty acids and esters attended the release from the solid support, whereas alkene isomerization was not observed in methanol solvent.

Mixtures of alkynyl and natural lipids were made up to test the conditions for separations of the alkynyl-tagged lipids from their natural counterparts. Treatment of the lipid mixtures in methanol with excess Co₂(CO)₈ at room temperature for 30 min was followed by addition of phosphine on a silica support. This mixture was heated at 55°C for 2 h with vigorous stirring and then transferred to an Eppendorf centrifuge tube fitted with a 0.45 µm nylon filter. The solid support was washed and filtered several times to remove the nonbound compounds. Subsequent treatment with Fe(NO₃)₃ at room temperature for 30 min released the alkyne. This enrichment protocol generally works for both fatty acids and phospholipids.

Fig. 3 shows a typical separation of synthetic alkynyl lipids and the corresponding natural compounds. Fig. 3A shows separation of alkynyl analogs of palmitic (a-Palm) and linoleic (a-Lin) acids from the natural acids by cobalt complex formation, “pull down” on the phosphine solid support, wash of natural fatty acids, and release of the alkynyl fatty acids after Fe(NO₃)₃ oxidation. The same sequence shown in Fig. 3B was used for treatment of the phospholipid mixture made up of DPPC and three alkynyl analogs of that phosphocholine, di-15a-PPC, di-7a-PPC and P-15a-PPC. The compound 1-palmitoyl 2-acetyl-sn-glycero-3-phosphocholine [P(OAc)PC] was added as an external standard. DPPC is the dominant phospholipid present in the wash solution, whereas the three alkynyl lipids are all retained on the silica solid support and released upon reaction with ferric nitrate. There is some alkyne in the wash, either from incomplete immobilization onto the solid support or premature oxidation of the cobalt at this step. In addition, there is some cross-contamination evident in the release due to nonspecific binding, but the alkynyl lipids are greatly enriched in the cobalt complex catch-and-release strategy. Similar studies were carried out with fatty acids and esters, and similar results were obtained.

Fig. 3.

Separation of alkynyl analogs from natural lipids. HPLC/MS analysis of (A) a mixture of linoleic, palmitic, a-linoleic, and 15a-palmitic acids and (B) a mixture of di-15a-PPC, di-7a-PPC, P-15a-PPC, DPPC, and P(OAc)PC as external standard. Top panels show the initial mixture, middle panels show the washings after immobilization of the alkyne-cobalt complexes onto the solid support, and the bottom panels show the recovery of the alkynes upon oxidative release from the support.

Quantitation using cobalt complex

The characteristic absorbance of alkyne cobalt complexes at 349 nm has the potential for use in the detection and quantitation of alkyne lipid surrogates. The extinction coefficient for the cobalt complex is essentially the same for terminal and internal alkynes (3,700 M⁻¹ cm⁻¹), and endogenous phospholipids or fatty acid esters do not generally absorb at this wavelength. This permits quantitative analysis of alkynyl lipids extracted from a biological mixture, e.g., by the use of HPLC/UV_349nm when coupled with addition of a known amount of an alkyne standard that forms a cobalt complex. This method greatly simplifies the quantitation in that a single standard is used, and response curves need not be generated for every compound of interest. In addition, it greatly simplifies the detection of alkynyl species in a complex mixture simply by UV analysis at 349 nm.

Model assays.

This approach was tested in model assays of alkynyl phospholipids making use of various added alkynyl standards that form cobalt complexes having convenient retention times relative to the phospholipid complexes on normal and reverse-phase HPLC. The mixtures of alkynyl phospholipids and standards were treated with Co₂(CO)₈, then analyzed by HPLC with UV detection at 349 nm (see supplementary Fig. II). With addition of an appropriate alkyne standard, either normal-phase or reverse-phase HPLC/UV on the alkyne cobalt complexes can be used to report the amounts of alkynyl lipids present in a mixture. Lipids bearing two alkynyl groups analyze appropriately if the molar absorption is assumed to be twice that of a mono-alkyne. From the results in Table 1, one can see that the amount of alkyne based on the cobalt quantitation is in good agreement with the theoretical value.

TABLE 1.

Quantitation of di-15a-PPC, di-7a-PPC, and P-15a-PPC by HPLC/UV detection of their cobalt complexes

Phospholipid	Reverse Phase15a-palmitic std	Reverse PhasePh-butyne std	Normal Phase15a-P(OAc)PC std	Theoretical
di-15a-PPC	23 nmol	17 nmol	19 nmol	20 nmol
di-7a-PPC	20 nmol	16 nmol	19 nmol	20 nmol
P-15a-PPC	24 nmol	19 nmol	22 nmol	20 nmol

Alkynyl lipids in cell cultures.

The utility of the cobalt complex for quantitative analysis is illustrated for the incorporation of an alkynyl lipid into cells. After incubating RAW 264.7 cells with media enriched with di-15a-PPC for 5 h, the cells were washed to remove excess lipid and scraped into acidic methanol, and the phospholipids were extracted using a modified Bligh and Dyer protocol. The lipid extract was treated with Co₂(CO)₈ and 7a-palm was added as an internal standard. The mixture was then analyzed by HPLC with UV detection at 349 nm (see supplementary data for a typical chromatogram). The amount of di-15a-PPC that is associated with the cells could then be determined based on its UV absorbance relative to the internal standard. Four separate samples were analyzed by this method, and they were shown to contain 11.9, 7.5, 10.2, and 7.9 nmol/10⁶ cells of di-15a-PPC. The fourth sample was also analyzed by mass spectrometry using an unnatural 33:0PCe as a standard. By MS analysis, cells were found to have 7.6 nmol/10⁶ cells of di-15a-PPC, a result that is nearly identical to the cobalt quantitative analysis. With these values and the known concentration of lipid used in the incubation, an incorporation of di-15a-PPC in cells of 0.5–0.7% was calculated.

Because the media containing excess di-15a-PPC was washed away, we know that this amount of incorporation does not represent free di-15a-PPC. However, we should point out that it is difficult to determine how much of the di-15a-PPC was incorporated into the cell or is merely associated with the outer membrane. From our previous work, we know that some of the di-15a-PPC is indeed incorporated into the cell. In that work, primed RAW 264.7 cells enriched with di-15a-PPC were treated with phorbol 12-myristate 13-acetate (PMA) in the presence of n-butanol. The di-15a-PPC was converted to the transphosphatidylation reaction product 32:4 alkynyl PtdBuOH via the phospholipase D metabolic pathway, demonstrating that this metabolic pathway is still active with the alkynyl PC as substrate.

Alkynyl fatty acids in cell culture

RAW 264.7 cells were incubated with 15a- or 7a-palmitic acid for 15 h, then the cells were lysed, and the lipids were extracted. In this experiment, the alkynyl fatty acids should be taken up by the cells and incorporated into a variety of phospholipids and other esters. Subsequent observation of a complex mixture of alkynyl phospholipids indicates that these fatty acids have been successfully used in metabolic processes. Several different analyses, described below, were carried out on these lipid extracts to demonstrate the utility of the cobalt approach in identifying these lipids.

Separation of lipid classes.

The lipid extract was treated with Co₂(CO)₈ and the resultant mixture was analyzed by normal-phase HPLC under conditions in which lipid classes (i.e., phosphatidylethanolamine, phosphatidylserine, phosphatidylinositol, and PC) were separated. 15a-P(OAc)PC was added as an internal standard so that the alkynyl phospholipid cobalt complexes could be assayed and lipid class distribution could be determined by the characteristic 349 nm absorption. The chromatographic analysis included the phosphocholines, eluting at ∼36 min (see Fig. 4), the ethanolamines (21 min), along with the serines (28 min), inositols (20 min), and free fatty acids (12 min). Triglycerides, diacylglycerols, sterol esters, and other more esoteric phospholipids, such as cardiolipins, were not analyzed in these studies, but the approach illustrated here could be expanded to include a broader range of lipid classes. As shown in Fig. 4, the alkynyl cobalt-modified phospholipids are distributed almost entirely into the phosphocholine lipid class. Triplicate analysis of lipid extracts from experiments with 7a-palmitic acid enrichment showed that over 96% of the alkynyl phospholipids were present as phosphocholines. For 15a-palmitic acid, some 12–14% of the alkynyl fatty acid was detected in the phosphoethanolamine fraction, the remainder being present as phosphocholines. The total amount of alkynyl phospholipids generated was quantitated based on its UV absorbance relative to the internal standard and is presented in Table 2 (row 1). Incorporation of 15a-palmitic and 7a-palmitic resulted in the formation of 2.4–7.2 and 1.8–4.9 nmol/10⁶ cells of alkynyl phospholipids, respectively. There is not a significant difference in the total amount of phospholipid formed for each of the fatty acids, although there is some variation between cell batches. No free fatty acid was detected (Table 2, row 3), suggesting that there was significant incorporation into the cell in some esterified form. The alkynyl phospholipid cobalt-complex fractions shown in Fig. 4 were collected and subsequently analyzed by mass spectrometry. The major alkynyl phospholipids identified by m/z of the cobalt complex molecular ion were 15a-PPPC and di-15a-PPC for experiments with 15a-palmitic, and the corresponding 7a-PPPC and di-7a-PPC were the major products when 7a-palmitic was the fatty acid added to the cellular media.

Fig. 4.

Alkynyl fatty acid phospholipid class incorporation determined by HPLC/UV of cobalt complexes. Lipid extracts from cells incorporated with (A) 7a-palmitic acid or (B) 15a-palmitic acid. Extracts were spiked with 15a-P(OAc)PC, treated with Co₂(CO)₈, and analyzed by normal phase HPLC (hexanes/iPrOH/NH₄OAc gradient) with UV detection at 349 nm.

TABLE 2.

Quantitative analysis of alkynyl fatty acid incorporation and alkynyl phospholipid formation

Analysis	15a-Palmitic(nmol/106 cells)			7a-Palmitic(nmol/106 cells)
aPCs (NP class sep)a	4.3	2.4	7.2	4.9	1.8	2.5
aPCs (pull-down)b	ND	1.7	4.0	ND	1.6	3.1
aFA free (NP)c	0	0	0	0	0	0
aFA (LiOH)d	ND	ND	14.4	ND	ND	9.0
% inc aFA into PLe	4	2	6	4	2	2
% inc aFA totalf	ND	ND	12	ND	ND	8

Each series of experiments carried out on the same cell batch, but separate samples. Values are reported as nmol/10⁶ cells. ND, not detected.

^aNormal phase class separation; quantitation of cobalt complex carried out in triplicate.

^bMass spectral quantitation after alkyne enrichment.

^ca-Palmitic acid detected in normal phase class separation.

^dTotal amount of a-palmitic acid detected after hydrolysis of phospholipids, quantitation of cobalt complex.

^eincorporation of a-palmitic acid into phospholipids.

^fTotal amount of a-palmitic acid incorporation.

Alkynyl lipid affinity purification.

The lipid class separation and subsequent mass spectral analysis provide a preliminary identification of the alkynyl phospholipids. However, minor alkynyl phospholipids may be difficult to detect unless specifically enriched or captured using the cobalt approach. Analysis of the cobalt pull-downs were carried out on lipid extracts from experiments in which alkynyl fatty acids were incorporated into RAW 264.7 cells. Typical chromatograms for these analyses are presented in Fig. 5. The phospholipids 15a-P(OAc)PC and P(OAc)PC were added to the cell extract to monitor the efficacy of the pull-down and release strategy. This not only gives us an indication of how well the enrichment is working on a sample by sample basis but also allows us to carry out quantitation of individual phospholipid species. As shown in the Fig. 5, all of the P(OAc)PC is seen in the wash fractions whereas most of 15a-P(OAc)PC is observed in the ferric nitrate release fraction, indicating that the capture-and-release technique is working and that we should have only alkynes enriched in the release sample. This greatly simplifies the mass spectral analysis and identification of the alkynyl phospholipids.

Fig. 5.

HPLC/MS analysis of lipid extracts of RAW 264.7 cells incubated with 15a-palmitic acid. Top panel shows total cell extracts, middle panel shows wash fraction after cobalt complex formation, and bottom panel shows ferric nitrate release from phosphine silica pull-down. Analysis was carried out with C8 HPLC using a H₂O/MeOH/NH₄OAc gradient.

Because the release fraction in Fig. 5 is the result of an alkyne capture-and-release strategy, one can reasonably assume that most, if not all, of the constituents observed in the HPLC/MS chromatogram have alkyne functionality. Based on that assumption, a careful analysis suggests that a number of minor phospholipid products are present but that the two major products present in the release fractions for both 7a- and 15a-palmitic acid cell incubations are phosphatidyl cholines containing either one or two alkynyl fatty acid esters at sn1 and sn2. These two lipids account for 59 and 68% of the total phospholipid for 15a- and 7a-palmitate incorporation. This is consistent with the results described above for the mass spectral identification of the cobalt-alkyne complexes.

Other identifiable lipids were found in the release fractions for the 15a- and 7a-palmitate cellular incorporation experiments. In addition to the major phosphocholine class compounds, some plasmalogen and ethanolamine structures are also suggested for some of the minor constituents, all of which had MS characteristics consistent with the presence of one alkynyl palmitoyl group as a glyceryl ester.

The amount of alkynyl phospholipids was determined relative to the standard for two samples and is presented in Table 2 (row 2). The amount of phospholipids formed from the 15a-palmitic acid is just slightly higher than for the 7a-palmitic acid-labeled cells. In addition, the amount of phospholipid formed based on mass spectral quantitation is in good agreement with the cobalt quantitation of the lipid classes described above. Not only does this result validate the cobalt approach but it also indicates that there is good recovery of the phospholipids after enrichment.

Base hydrolysis and fatty acid analysis.

The lipid extract from alkynyl fatty acid incorporation into RAW 264.7 was treated with 3M LiOH to hydrolyze all fatty esters, and fatty acids were then extracted. As the original cell extract contained little, if any, detectable free fatty acids, this extract included all fatty acids that were originally present as esters in the cell. The fatty acid extract was spiked with an alkyne internal standard and treated with Co₂(CO)_8. Reaction of the fatty acid mixture with Co₂(CO)₈ followed by HPLC/UV_349nm with an internal standard permitted estimation of total amount of fatty acid incorporation into the cells during the incubation period. By this analysis, we estimate that the total amount of 15a- and 7a-palmitic acid incorporated was 14.4 and 9.0 nmol/million cells, respectively, during the 15 h incubation of the fatty acids with RAW 264.7 cells (Table 2, row 4). This corresponds to 12 and 8% total incorporation for 15a- and 7a-palmitate, respectively. The slightly higher incorporation of 15a-palmitate is consistent with the slightly higher levels of alkynyl phospholipids present.

HPLC/MS analysis was also carried out on the fatty acid mixture obtained by base hydrolysis of the cellular extract, see Fig. 6. In addition to 15a-palmitic, this analysis shows that 17a-stearic acid, another alkynyl fatty acid, was present in the mixture. Independent synthesis, extracted ion chromatography, and cobalt complex formation of the two-carbon chain extension product of the alkynyl palmitate confirms its presence in the lipid extract. For 15a-palmitic acid cellular incubations, 16% of the fatty acid extract is in the chain elongation form, whereas experiments with 7a-palmitic acid give rise to 4% of the metabolic product in the mixture. The terminal alkyne seems to be incorporated somewhat better in the cell and phospholipids than the internal alkyne, and it appears to undergo metabolic chain extension more readily as well. Nonetheless, both compounds are readily incorporated into the lipids of RAW 264.7 cells, and both undergo at least some of normal metabolism expected of endogenous fatty acids.

Fig. 6.

LC/MS analysis of fatty acids in cell extracts after base hydrolysis. Top panel shows total ion current of fatty acid mixture, middle panel shows extracted ions for [15a]16:0 and [7a]18:0, and bottom panel is a reference spectrum of authentic standards of [15a]16:0, [7a]18:0, and 18:2. Analysis was carried out with C18 HPLC using a H₂O/MeOH/NH₄OAc gradient.

Conclusions

Tracking the cellular incorporation of lipids is essential to understanding lipid metabolic and signaling processes. Most of the studies to date have utilized radioactive-labeled compounds to monitor changes in lipids during cellular processes. Although these procedures are widely used and easily performed, they provide a limited amount of information about the lipid species involved in signaling events. Furthermore, even advanced methods of mass spectrometry have limitations with regard to discriminating very complex mixtures of isobaric species. The main drawback for all of these methods is the inability to discriminate between labeled and naturally occurring lipids in very complex mixtures that contain in excess of 1,000 species of phospholipids.

The use of alkyne tags for substrates has become popular with the interest in click chemistry. Alkyne-modified substrates, in conjunction with click chemistry, offer an approach to capture the substrate through selective modification of the alkyne moiety. While this approach has been successfully used in a wide variety of applications, it has some potential drawbacks as a technique for tracking the cellular incorporation of lipids. The alkyne tag alone is not sufficient for identifying the modified lipid because the alkynyl lipid is isobaric with naturally occurring compounds. Modification of the alkynyl lipid through click chemistry would allow one to identify the alkynyl lipid in a complex mixture. However, quantification of dozens of lipid species simultaneously would not be practical, as standards for these heavily derivatized phospholipids are not commercially available. Finally, conventional click chemistry is only viable for terminal alkynes and results in the permanent modification of the phospholipid of interest.

This work demonstrates that alkynyl probes have the potential to provide a broad-scope method for quantitative analysis of cellular lipid transformations when coupled with an approach involving cobalt complex formation, pull-down on a phosphine solid support, and oxidative release. The utility of these probes is enhanced by powerful chromatographic methods for separation of lipid classes and molecular species, as well as by the fact that the alkyne-cobalt complex has a signature UV absorption at 349 nm. As shown in these studies, detection of the alkyne lipid analogs has been optimized such that they can be used to address questions on lipid metabolic and signaling pathways (23).