|Home | About | Journals | Submit | Contact Us | Français|
When photoirradiated, an unnatural sialic acid analog can covalently capture the complex formed by ganglioside GM1 and cholera toxin subunit B.
Cell surface glycoconjugates mediate essential intra- and inter-cellular interactions. Many of these recognition events are regulated by the presence or absence of sialic acids,1 a family of nine carbon α-keto acids attached to glycoproteins and glycosphingolipids. While sialylated molecules serve important regulatory roles for mammalian cells, they are also exploited by many pathogens.2 For example, influenza viruses utilize hemagglutinin to bind sialylated cell surface receptors, while Trypanosoma cruzi’s trans-sialidase transfers sialic acids from host glycoconjugates to the parasite.2 Sialylated glycosphingolipids, known as gangliosides (Fig. 1a), are of particular interest because they often serve as the principal recognition elements for the invasion of host cells by viruses, bacteria, and their secreted toxins.3
Explicitly identifying ganglioside interaction partners and deciphering the molecular details of these complexes are challenging tasks. Most glycan-mediated complexes have low affinities and their transient natures make traditional purification techniques impractical, as the unstable complexes cannot withstand the washing steps. Furthermore, the amphiphilicity of gangliosides poses additional constraints on purification. Methods to capture ganglioside-protein complexes in their native settings would represent a powerful tool for characterizing these labile complexes. Our strategy is to introduce photoreactive groups into glycan components of cellular glycoconjugates. We accomplish this using metabolic oligosaccharide engineering, a technique that has been highly successful at introducing a wide range of chemical modifications into various monosaccharides, including sialic acid (Fig. 1b).4 Diazirine modification of cellular sialosides is achieved by culturing mammalian cells with a diazirine-containing N-acetylmannosamine analog, Ac4ManNDAz, compound 2 (Fig. 1c), which is metabolized by cells and displayed on the cell surface as the sialic acid analog, SiaDAz, 4 (Fig. 1c). Upon UV irradiation, the diazirine can be activated to a carbene. If positioned appropriately, the highly reactive carbene can form a covalent bond between sialic acid and the sialoside binding partner(s). The covalently-linked complex can then be analyzed and the components identified. We recently demonstrated the incorporation of a diazirine into the glycans attached to cell surface sialic acid binding lectin CD22 and used the diazirine’s crosslinking capacity to capture CD22 oligomers.5
Encouraged by reports of the incorporation of unnatural sialic acid analogs into gangliosides,6 we investigated whether mammalian cells are capable of producing SiaDAz-containing gangliosides. Jurkat cells, which have been used previously in metabolic oligosaccharide engineering,6f,7 were cultured with 2 (Ac4ManNDAz) for 72 hours. As a control, cells were also cultured with the 2-epimer of Ac4ManNDAz, Ac4GlcNDAz – compound 3 (Fig. 1c), which is not expected to be metabolized to SiaDAz.8 Additionally, Jurkat cells were cultured with Ac4ManNAc to determine whether exogenous addition of a normal metabolite in the sialic acid biosynthesis pathway affects ganglioside production. Acetylated sugars9 were added to the media at a final concentration of 100 μM. After harvesting the cultured cells, the gangliosides were extracted using established methods10 (described in supplementary information) and resolved using high performance TLC (HPTLC) plates. To visualize gangliosides, the HPTLC plates were stained with resorcinol, which, upon heating, reacts specifically with sialic acid-containing molecules.11
By resorcinol staining, we observed that untreated Jurkat cells synthesize multiple gangliosides (Fig. 2a). These data were corroborated by mass spectrometry data, which indicated the presence of gangliosides GM3, GM2, GM1, and GD1a, each present in two forms that arise from different fatty acid chain lengths (Fig. 2b, S4 and Table S1). While cells cultured with 1 produced gangliosides whose profile appeared identical to that of untreated cells, cells that were cultured with 2 produced novel ganglioside products (Fig. 2a). We interpret these new bands to correspond to diazirine-modified gangliosides, consistent with the prediction that introduction of the diazirine side chain should increase the hydrophobicity of gangliosides and their mobility in the TLC eluent. In addition to these new molecules, the 2-treated cells continue to generate their natural gangliosides. We confirmed production of diazirine-modified gangliosides by MALDI-TOF-MS (Fig. 2c, S5, and Table S2). Gangliosides from cells cultured with 3 were indistinguishable from those from untreated cells (Fig. 2a), suggesting that 3 is not efficiently metabolized to 4 in Jurkat cells.
To determine if SiaDAz-containing gangliosides can still be recognized as specific receptors, we used a fluorescently-labeled GM1-binder, cholera toxin subunit B-Alexa Fluor 488 conjugate (CTxB-Alexa488), to probe HPTLC plates (Fig. 2d).12 As expected, CTxB-Alexa488 binds GM1 in all of the Jurkat cell extracts. Importantly, CTxB-Alexa488 also recognizes additional species in extracts from cells cultured with 2; we assign these additional bands to diazirine-modified GM1 (GM1-SiaDAz). These results provide evidence that Jurkat cells cultured with 2 produce gangliosides that contain 4 and, further, that CTxB recognition of GM1 is maintained upon installation of the unnatural diazirine side chain on sialic acid.
With the observation that GM1-SiaDAz is recognized by CTxB, we next investigated whether we could covalently capture the ganglioside-protein complex in a cellular setting using photoirradiation. In brief, Jurkat cells were cultured with no additional sugar, 1, 2, or 3 for 70 hours, followed by the addition of CTxB for 45 min to allow cell surface binding. Cells were exposed to 365 nm light for 45 min to initiate photocrosslinking. After cell lysis and SDS gel electrophoresis, proteins were transferred to a polyvinylidene fluoride (PVDF) membrane, which was then probed for the presence of CTxB (Fig. 3a and S6) by standard immunoblotting techniques. When we analyzed lysates from Jurkat cells cultured with 2 and photoirradiated, we observed both unmodified CTxB and an additional band whose mobility was consistent with the addition of a GM1 molecule to CTxB. In contrast, untreated, 1-treated, and 3-treated cells displayed a single band corresponding to unmodified CTxB (Fig. 3a). In complementary experiments where we probed the membrane for GM1, multiple bands were observed only when cells were cultured with 2, consistent with the formation of a covalent GM1-CTxB complex (Fig. S6a). Experiments performed without photoirradiation confirmed that formation of the putative GM1-CTxB complex was UV-dependent (reported in supplemental information, Fig. S6).
To confirm that the novel complex consisted of CTxB crosslinked to a ganglioside, we performed cell culturing and crosslinking in the presence of a glycosphingolipid biosynthesis inhibitor, N-butyldeoxygalactonojirimycin (NB-DGJ).13 When the inhibitor was added to cultured cells, we observed a marked decrease in GM1 expression (Fig. S6d and f) as well as a decrease in CTxB binding (Fig. 3b and S6e). Importantly, the intensity of the band corresponding to the putative GM1-CTxB species decreases dramatically. Taken together, these results demonstrate that Jurkat cells cultured with Ac4ManNDAz produce a diazirine-modified ganglioside that can be efficiently photocrosslinked to CTxB.
In summary, we demonstrate the use of metabolic oligosaccharide engineering to enable a mammalian cell line to produce photoactivatable gangliosides. Metabolically-produced GM1-SiaDAz is recognized by a GM1 binding partner, CTxB, and photoirradiation can be used to efficiently capture this ganglioside-protein complex. These results show promise for using this photoreactive tool to capture and characterize ganglioside-mediated interactions. The ganglioside GM1 is of special interest because it is highly enriched in membrane microdomains commonly known as lipid rafts.14 The ability to metabolically introduce a photocrosslinker into GM1 and to form covalent complexes with neighboring proteins represents a method for gaining information about the composition of lipid rafts with minimal disruption to the native organization of membrane molecules.
M.R.B. thanks the NSF for a graduate research fellowship. M.R.B. and C.M.W. acknowledge the support of Stanford University. We would like to thank Kim Orth (UT Southwestern Medical Center) for sharing Jurkat cells and Dr. Yan Li (UT Southwestern Medical Center Protein Chemistry Technology Center) for assistance with ESI-MS analysis. We would also like to thank Dr. Kevin Luebke, Peter Vu, and Dr. Fan Yang (UT Southwestern Medical Center) for critical comments on the manuscript. Financial support was provided by the University of Texas Southwestern Medical Center, the March of Dimes (5-FY06-913), the Welch Foundation (I-1686), and the National Institutes of Health (GM090271). This research was supported in part by the National Institutes of Health (NIH/NCRR)-funded grant entitled ‘Integrated Technology Resource for Biomedical Glycomics’ (grant no. 1 P41 RR018502-01) to the Complex Carbohydrate Research Center (Athens, GA). This content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of General Medical Sciences or the National Institutes of Health. J.J.K. is an Alfred P. Sloan Research Fellow.