Fluorescent BODIPY-labelled GM1 gangliosides designed for exploring lipid membrane properties and specific membrane-target interactions

Chemistry and Physics of Lipids 159 (2009) 38–44

Contents lists available at ScienceDirect

Chemistry and Physics of Lipids journal homepage: www.elsevier.com/locate/chemphyslip

Fluorescent BODIPY-labelled GM1 gangliosides designed for exploring lipid membrane properties and specific membrane-target interactions Ilya Mikhalyov b , Natalia Gretskaya b , Lennart B.-Å. Johansson a,∗ a b

Department of Chemistry, Biophysical Chemistry, Umeå University, S-901 87 Umeå, Sweden Shemyakin & Ovchinnikov Institute of Bioorganic Chemistry, Russian Academy of Sciences, ul. Miklukho-Maklaya 16/10, Moscow 117871, Russia

a r t i c l e

i n f o

Article history: Received 2 December 2008 Received in revised form 16 January 2009 Accepted 16 January 2009 Available online 24 January 2009 Keywords: BODIPY-labelled gangliosides GM1 ganglioside Förster energy transfer

a b s t r a c t New fluorophore-labelled GM1 gangliosides have been synthesised and spectroscopically characterised. Spectroscopically different BODIPY groups were covalently linked, specifically to either the polar or the hydrophobic part of the ganglioside molecule. The absorption and fluorescence spectroscopic properties are reported for 564/571-BODIPY- and 581/591-BODIPY-labelled GM1 . Each of the different BODIPY groups is highly fluorescent and depolarisation experiments provide molecular information about the spatial distribution in lipid bilayers, as well as order and dynamics. From experiments performed on two spectroscopically different BODIPY:s, specific interactions can be revealed by monitoring the rate/efficiency of donor–acceptor electronic energy transfer. Systems of particular interest for applying these probes are e.g. mixtures of lipids, and peptides/proteins interacting with lipid membranes. © 2009 Elsevier Ireland Ltd. All rights reserved.

1. Introduction The molecular insights obtained from optical spectroscopic methods, such as single-molecule fluorescence spectroscopy (García-Sáeza and Schwille, 2008; Zander et al., 2002) and fluorescence imaging (Bagatolli et al., 2003) are highly dependent on the existence of well-defined and designed fluorophore-labelled biomolecules. The biomolecules could be proteins, peptides (Bogen et al., 1999), nucleotides (Neubauer et al., 2007), and lipids (Grechishnikova et al., 1996). The gangliosides are the most complex class of glycolipids, and they contain one or more sialic acid (N-acylneuraminic acid), rendering them negatively charged (cf. Fig. 1). Gangliosides are most abundant in the plasma membrane of nerve cells in which they constitute 5–10% of the total lipid mass. It has been strongly suggested that the ganglioside GM1 acts as a

Abbreviations: BODIPY-FL-GM1, N-(BODIPY-FL-pentanoyl)-ganglioside; 564/ N-(BODIPY-564/570-pentanoyl)-neuraminosyl-ganglio570-BODIPY-C5-GM1 (p), side GM1 labelled in the polar (p) head group; 564/570-BODIPY-C5-GM1 (np), N(BODIPY-564/570-pentanoyl)-ganglioside GM1 labelled in the non-polar (np) region; 564/570-BODIPY-C5-asialo GM1 , N-(BODIPY-564/570-pentanoyl)-asialo-ganglioside GM1 labelled in the non-polar region; 581/591-BODIPY-C5-GM1 (p), N-(BODIPY581/591-pentanoyl)-neuraminosyl-ganglioside GM1 labelled in the polar (p) head group; 581/591-BODIPY-C11-GM1 , N-(BODIPY-581/591-undecanoic)-ganglioside GM1 labelled in the non-polar (np) region; DOPC, 1,2-dioleoyl-sn-glycero-3phosphocholine; eSM, egg sphingomyelin; GM1 , ganglioside GM1 ; r, steady-state fluorescence anisotropy; r(t), time-resolved fluorescence anisotropy; , the fluorescence lifetime. ∗ Corresponding author. Tel.: +46 90 786 5149; fax: +46 90 786 7779. E-mail address: [email protected] (L.B.-Å. Johansson). 0009-3084/$ – see front matter © 2009 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.chemphyslip.2009.01.005

surface cell receptor in the uptake of bacterial toxin, which causes the debilitating diarrhea of cholera (Zhang et al., 1995; Aman et al., 2001). Numerous studies show that amylo-␤-petides undergo conformational changes and aggregate as a result of interacting with a wide range of neuronal lipid membrane systems. This is one argument for their connection with the Alzheimer’s disease (Bokvist et al., 2004; Yip and McLaurin, 2001). The present work describes newly synthesised fluorophorelabelled GM1 lipids with respect to their optical spectroscopic properties. Moreover their suitability for probing natural gangliosides at a molecular level in various in vitro, as well as in vivo systems is addressed. The potential applicability of the various BODIPYlabelled gangliosides is also discussed. 2. Materials and methods BODIPY-labelled fatty acids and lipids were purchased from Invitrogen Inc., (USA) and Avanti Polar Lipids Inc. (Pelham, AL, USA), respectively. The gangliosides GM1 and GD1a were isolated from bovine brain, as described by Svennerholm (1974) and deacetyl-ganglioside GM1 and de-acetyl, de-acyl-ganglioside GM1 by following the description of Sonnino et al. (1985). The chemical structures of the fluorescent groups and gangliosides used are displayed in Fig. 1. Acetyl anhydride, Triton X-100 were purchased from Aldrich (USA), Silica gel 100 (63–200 ␮m) and TLC plates Silica gel 60 were purchased from Merck (Germany). Exposure to light was avoided in all handling of the dyes. All solvents used were reagent grade and freshly distilled.

I. Mikhalyov et al. / Chemistry and Physics of Lipids 159 (2009) 38–44

39

Fig. 1. Structures of the BODIPY-labelled gangliosides GM1 (García-Sáeza and Schwille, 2008; Zander et al., 2002; Bagatolli et al., 2003; Bogen et al., 1999) and the BODIPYlabelled Asialo-ganglioside GM1 (Neubauer et al., 2007), n = 12, 14.

Preparation of vesicles. The LUV vesicles were prepared by the extrusion technique using an extruder manufactured by Lipex Biomembranes Inc. (Vancouver, Canada). Briefly, the dried lipid film, containing bovine sphingomyelin, cholesterol, ganglioside GM1 and labelled GM1 , was hydrated to the necessary concentration by 20 mM Tris–HCl buffer (pH 7.4), containing 1 mM disodium salt of EDTA. This mixture was freeze-thawed five times and then passed ten times through double polycarbonate filters (Nucleopore) with a pore size of 100 nm. Steady-state and time-resolved fluorescence experiments. The emission and excitation fluorescence spectra were recorded using a Fluorolog® -3 (Jobin Yvon Inc., USA) spectrometer equipped with

Glan-Thompson polarisers. The spectral bandwidth of the excitation and emission monochromators was 2 nm. The fluorescence spectra were corrected. Fluorescence lifetime decays were measured by using the time-correlated single-photon-counting (TCSPC) technique. The instrument was a PRA 3000 (PRA, Canada and the excitation source was NanoLED N-14, 471 nm, (IBH, Scotland) pulsed diode, operated at 800 kHz or PicoQuant laser, 470 nm. The excitation and emission wavelength were selected using a set of filters (Melles Griot, The Netherlands). An interference filter centred at 470 nm (FWHM = 9.3 nm) was used for the excitation, and for the emission interference filters centred at 520 nm (FWHM = 28 nm), 540 nm (FWHM = 28 nm), and long pass filters from 550 to 610 nm.

40

I. Mikhalyov et al. / Chemistry and Physics of Lipids 159 (2009) 38–44

The fluorescence decays were collected over 1024 channels with the resolution of 50 ps/ch, with at least 10,000 photons in the peak maximum for the lifetime experiments, which were performed with the emission polariser set at magic angle (54.7◦ ) relative to the excitation polariser. For the time-dependent anisotropy experiments, fluorescence decays were collected with the difference of 50,000 counts in the peak maxima of the fluorescence decays recorded for the parallel and perpendicular polariser settings. The mass spectra were recorded by using a MALDI-TOF system (Micromass, Manchester, UK). Absorption spectra were recorded on Cary 5000 (Varian, UK). The Förster radii were calculated from the absorption and the corrected fluorescence spectra were carefully determined. Using these data, the Förster radius (R0 ) was calculated from

 R0 =

9000(ln 10)(2/3)˚J 1285 n4 NA

1/6

 ε()f ()4 d

J= band

here f(), , ε(), ˚ and NA denote the corrected normalised fluorescence spectrum, the wavelength (nm), the molar absorptivity of the acceptor, the fluorescence quantum yield of the donor and the Avogadro constant, respectively. Throughout the calculations of R0 , the orientation factor 2  = 2/3 was taken to be a reference state. 3. Synthesis N-(BODIPY-564/570-pentanoyl)-neuraminosyl-ganglioside GM1 labelled in the polar (p) head group (564/570-BODIPY-C5-GM1 (p)) was synthesised as described previously (Marushchak et al., 2007). N-(BODIPY-564/570-pentanoyl)-ganglioside GM1 (564/570BODIPY-C5-non-polar-GM1 ) was synthesised from de-acetyl, de-acyl ganglioside GM1 by N-acylation with mixed anhydride of BODIPY-564/570-pentanoic acid, as described elsewhere (Acquotti et al., 1986). In brief, 4.0 mg (10 ␮mol) of BODIPY-564/570pentanoic acid was dissolved in dried acetonitryl (4 ml), 30 ␮l of triethylamine (solution in acetonitryl, 1:20 (v/v), 1.1 ␮mol) and 35 ␮l of iso-butyl-chlorformiate (solution in acetonitryl, 1:20 (v/v), 1.1 ␮mol) were added. The reaction mixture was stirred during 15 min at room temperature. After evaporation it was dissolved in 3 ml ethylacetate, washed with 2 ml water, then with a 2 ml saturated water–NaCl solution and stored over anhydrous sodium sulphate for 1 h. The solution of mixed anhydride with BODIPY-564/570-pentanoic acid was evaporated and dissolved in 1 ml dried THF. The de-acyl, de-acetyl-ganglioside GM1 (10 mg, 8 ␮mol) was dissolved in 2 ml THF–water, 5:1 (v/v), 1 ␮l (10 ␮mol) triethylamine and the solution of mixed anhydride with BODIPY-564/570-pentanoic acid in 1 ml THF was added. The reaction mixture was stirred overnight at room temperature. BODIPY-564/570-pentanoyl, de-acetylganglioside GM1 was isolated and purified by column chromatography on 1.5 g Silica gel 100 (0.4 cm × 25 cm) with chloroform–methanol–water, 65:25:4 (v/v/v), as the eluent, Rf = 0.17 (chloroform–methanol–15 ␮M aqueous calcium chloride, 60:35:8 (v/v/v)). BODIPY-564/570pentanoyl, de-acetylganglioside-GM1 was dissolved in 2 ml 0.5 M aqueous sodium bicarbonate and then 20 ␮l acetyl anhydride was added. The reaction was kept for 15 min at room temperature and desalted by dialysis against water overnight. The final BODIPY-564/570-pentanoyl GM1 was purified by column chromatography on 1.5 g Silica gel 100 (0.4 cm × 25 cm) with chloroform–methanol–water, 65:25:2 (v/v/v), as the eluent; Rf = 0.26 (chloroform–methanol–15 ␮M aqueous calcium chloride, 60:35:8 (v/v/v)), which is also obtained for the natural GM1 . The

product was obtained as a blue powder in the quantity of 3.2 mg (yield: 19%). MS data: 1678.71 (M1+ +Na), the calculated mass was 1678.76. This corresponds to the molecular specie, which contains the C18:1 sphingosine base; 1706.71 (M2+ +Na). The calculated mass was 1706.76 and it corresponds to the molecular specie that contains the C20:1 sphingosine base. N-(BODIPY-581/591-pentanoyl)-neuraminosyl-ganglioside GM1 (581/591-BODIPY-C5-GM1 ) was synthesised from de-acetyl ganglioside GM1 by N-acylation and mixed with the anhydride of BODIPY-581/591-pentanoic acid, as described for the N-(BODIPY-564/570-pentanoyl)-neuraminosyl-ganglioside GM1 (Marushchak et al., 2007). The mixed anhydride solution of BODIPY-581/591-pentanoic acid was prepared, as described above. The de-acetyl-ganglioside GM1 (5 mg, 3.5 ␮mol) was dissolved in the mixture of THF–water, 7:1 (v/v), 0.5 ␮l triethylamine and mixed anhydride, prepared from 2.1 mg (5 ␮mol) BODIPY-581/591-pentanoic acid in 2 ml of THF was added. The reaction mixture was stirred overnight at room temperature. Following evaporation, BODIPY-581/591-GM1 was purified by column chromatography on 1.5 g of Silica gel 100 (0.4 cm × 25 cm) with chloroform–methanol–water, 65:25:2 (v/v/v), as the eluent; Rf = 0.47 (chloroform–methanol–15 ␮M aqueous calcium chloride, 60:35:8 (v/v/v)). 2.9 mg (41%) of the product was obtained as a dark blue powder. MS data, 1929.03 (M1+ +Na), was calculated to be 1929.03, and it corresponds to the molecular specie that contains the C18:1 sphingosine base; 1957.04 (M2+ +Na), corresponds to a calculated mass of 1957.03 for the molecular specie that contains the C20:1 sphingosine base. N-(BODIPY-581/591-undecanoic)-ganglioside GM1 (581/591BODIPY-C11-GM1 ) was synthesised, as described for the N-(BODIPY-564/570-pentanoyl)-ganglioside GM1 . The mixed anhydride of BODIPY-581/591-undecanoic acid was prepared from 5 mg fatty acid, as described above. The de-acyl, de-acetyl-ganglioside GM1 (10 mg, 8 ␮mol) was dissolved in 2 ml THF–water, 5:1 (v/v), 1 ␮l (10 ␮mol) triethylamine and the solution of mixed anhydride of BODIPY-581/591-undecanoic acid in 1 ml f THF was added. The reaction mixture was stirred overnight at room temperature. BODIPY-581/591-undecanoyl, de-acetylganglioside GM1 was isolated and purified by column chromatography on 1.5 g Silica gel 100 (0.4 cm × 25 cm) with chloroform–methanol–water, 65:25:4 (v/v/v), as the eluent, Rf = 0.17 (chloroform–methanol–15 ␮M aqueous calcium chloride, 60:35:8 (v/v/v)). BODIPY-581/591undecanoyl and de-acetylganglioside-GM1 were dissolved in 2 ml 0.5 M aqueous sodium bicarbonate and 20 ␮l acetyl anhydride was added. The reaction was kept for 15 min at room temperature and desalted by dialysis against water overnight. The final BODIPY-581/591-undecanoyl GM1 was purified by column chromatography on 1.5 g Silica gel 100 (0.4 cm × 25 cm) with chloroform–methanol–water, 65:25:2 (v/v/v), as the eluent; Rf = 0.26 (chloroform:methanol:15 ␮M aqueous calcium chloride, 60:35:8 (v/v/v)), as is also found for the natural GM1 . The product was obtained as a blue powder in the quantity of 3.2 mg (yield: 19%). MS data, 1788.87 (M1+ +Na). A calculated mass of 1788.87 was obtained for the molecular specie that contains the C18:1 sphingosine base; 1816.89 (M2+ +Na). The mass 1816.90 was calculated for the molecular specie that contains the C20:1 sphingosine base. N-(BODIPY-564/570-pentanoyl)-asialo-ganglioside GM1 (564/570-BODIPY-C5-asialo GM1 ) was synthesised from lysoasialo-ganglioside GM1 by N-acylation with mixed anhydride of BODIPY-564/570-pentanoic acid. The lyso-asialo-ganglioside GM1 was synthesised by alkaline hydrolysis, as previously described (Sonnino et al., 1985). The mixed anhydride of BODIPY-564/570pentanoic acid was obtained, as described above. 5 mg (5 ␮mol) of lyso-asialo-ganglioside GM1 , was dissolved in THF:water (5:1, v/v) and then the solution of mixed anhydride, prepared from 2 mg (5 ␮mol) BODIPY-561/570-pentanoic acid in 1 ml THF was

I. Mikhalyov et al. / Chemistry and Physics of Lipids 159 (2009) 38–44

41

added. The mixture was stirred overnight at room temperature. BODIPY-564/570-pentanoyl-asialo-ganglioside GM1 was isolated and purified by column chromatography on 1.5 g Silica gel 100 (0.4 cm × 25 cm) with chloroform–methanol–water (65:25:2, v/v/v), as the eluent, Rf = 0.31 (chloroform–methanol–15 ␮M aqueous calcium chloride (60:35:8, v/v/v)), as is also found for the natural asialo-GM1 . The product was obtained as a blue powder in the quantity of 1.8 mg (1.3 ␮mol, yield: 26%). MS data, 1387.66 (M1+ +Na). A calculated mass of 1387.67 was obtained for the molecular specie, containing the C18:1 sphingosine base; 1415.69 (M2+ +Na), corresponds to the calculated mass of 1415.70 for the molecular specie that contains C20:1 sphingosine base. 4. Results and discussion 4.1. Chemical and spectroscopic properties of the BODIPY-labelled gangliosides The chemical structures of the synthesised and characterised BODIPY-labelled gangliosides are displayed in Fig. 1. The fluorescent group is either specifically linked to the extensive polar headgroup, or to one of the acyl chains of the gangliosides. The different BODIPY groups used are: the spectroscopically fluorescein-like BODIPY (BODIPY-FL), and the derivatives 564/570-BODIPY and 581/591BODIPY. The absorption and fluorescence spectra of BODIPY-FL, 564/570-BODIPY and 581/591-BODIPY are gradually redshifted due to the increasingly extended ␲-electronic systems. The corresponding absorption and fluorescence spectra of these compounds dissolved in methanol are shown in Fig. 2. The peak absorption and fluorescence spectra of 564/570-BODIPY and 581/591-BODIPY solubilised in GM1 -micelles and lipid vesicles are somewhat redshifted (ca. 10 nm) relative to the spectra recorded for the compounds dissolved in methanol. This indicates that the BODIPY groups are exposed to a more polar environment in the micellar and vesicle systems (cf. Fig. 3). The fluorescence spectrum of BODIPY-FL overlaps with the absorption spectra of 564/570-BODIPY and 581/591-BODIPY, and the fluorescence spectrum of 564/570-BODIPY overlaps with the absorption spectrum of 581/591-BODIPY (cf. Fig. 2). Hence, electronic energy transfer may occur from BODIPY-FL to 564/570- and 581/591-BODIPY as well as from 564/570- to 581/591-BODIPY. The Förster radii of donor–acceptor energy transfer in methanol from BODIPY-FL to 564/570-BODIPY, as well as from BODIPY-FL to 581/591-BODIPY were calculated to be 60 and 58 Å, respectively. The Förster radii obtained for the corresponding pairs in GM1 micelles are equal, and have the value of 58 Å. The Förster

Fig. 2. Absorption (—, - - -, – – –) and normalised fluorescence (· · · · ·, - · · - · · -, - · - · -) spectra of BODIPY-FL-C5-GM1 , BODIPY-564/570-C5-GM1 and BODIPY-581/591-C11GM1 , respectively. All compounds were dissolved in methanol.

radii for the 564/570-BODIPY and 581/591-BODIPY are 70 and 68 Å in methanol and GM1 micelles, respectively. In previous studies of BODIPY-FL-GM1 (Bergström et al., 2002; Mikhalyov et al., 2002) fluorescent ground dimers were observed at high mole fractions of the probe. However, studies of 564/570-BODIPY-GM1 and 581/591-BODIPY-GM1 under the same conditions did neither reveal any emitting dimers, nor any distortions of the absorption spectra which are typically observed for strong ground state interaction between monomers. The used fractions of the new ganglioside probes were up to 10 mol% in DOPC vesicles. The obtained fluorescence lifetimes of 564/570-BODIPY-GM1 and 581/591-BODIPY-GM1 , separately solubilised in GM1 micelles at different probe-labelled lipid/lipid ratios, are summarised in Table 1. For the 564/570-BODIPY-GM1 and 581/591-BODIPY-GM1 , the fluorescence decay is very well described by a single exponential function with a lifetime equal to 4.1 and 5.1 ns, respectively. Neither the labelling of the ganglioside in the polar headgroup (p) nor in the non-polar lipid region (np), seems to influence the photophysics relaxation. It is also worth noting that the lifetimes are independent of the mole fractions of labelled-GM1 , which indicates that the influence of ground state dimers is negligible (Bergström et al., 2002). 4.2. Fluorescence depolarisation of BODIPY-labelled gangliosides Various BODIPY-labelled gangliosides have been studied when solubilised in GM1 micelles and lipid bilayers. In general, the depolarisation experiments reveal quite high values of the steady-state

Fig. 3. (A) The corrected normalised fluorescence spectra of 581/591-BODIPY-C5-GM1 (p) solubilised in: vesicles composed of GM1 /bSM/Chol (molar ratio = 3:4:3) and doped with 0.3 mol% of the probe (solid line); GM1 micelles, doped with 1 mol% (dotted line), methanol (dashed line). The peak maxima of the fluorescence in vesicles, micelles and methanol are at 595, 599 and 590 nm, respectively. Also displayed are the steady-state emission anisotropy (r) obtained for the corresponding vesicle (solid line) and micellar (dotted line) system upon excitation at 550 nm. (B) The corrected normalised fluorescence spectra of 581/591-BODIPY-C11-GM1 (np) solubilised in; vesicles composed of GM1 /bSM/Chol (molar ratio = 3:4:3) and doped with 0.3 mol% of the probe (solid line); GM1 micelles, doped with 1 mol% (dotted line), methanol (dashed line). The peak maxima of the fluorescence in vesicles, micelles and methanol are at 596, 599 and 590 nm, respectively. Also displayed are the steady-state emission anisotropy (r) obtained for the corresponding vesicle (solid line) and micellar (dotted line) system upon excitation at 550 nm.

42

I. Mikhalyov et al. / Chemistry and Physics of Lipids 159 (2009) 38–44

Table 1 The fluorescence lifetime () and steady-state anisotropy (r) obtained for 0.033 and 1 mol% of 564/571-BODIPY- and 581/591-BODIPY-labelled gangliosides in GM1 micelles at 294.8 K. The BODIPY groups are either labelled in the polar (p) or nonpolar (np) parts of the gangliosides. The r-value corresponds to an average taken over the wavelength range indicated in the parenthesis. System

r(ex @ 530a /550b nm)

 (ns)

564/571-BODIPY C5-GM1 (p) 1 mol% 564/571-BODIPY C5-GM1 (p) 0.033 mol% 564/571-BODIPY C5-GM1 (np) 1 mol% 564/571-BODIPY C5-Asialo-GM1 (np) 1 mol% 581/591-BODIPY C5-GM1 (p) 1 mol% 581/591-BODIPY C5-GM1 (p) 0.033 mol% 581/591-BODIPY C11-GM1 (np) 1 mol%

0.202 (559–634)c 0.299 (579–653) 0.185 (552–651) 0.177 (554–648) 0.235 (596–675) 0.318 (590–675) 0.207 (590–665)

5.1 5.1 5.1 5.0 4.1 4.1 4.2

a b c

Refers to the excitation of 564/571-BODIPY. Refers to the excitation of 581/591-BODIPY. Range used in calculating the average anisotropy.

excitation and emission anisotropies, as is exemplified in Fig. 3 and Table 1. Lipid vesicles. The rather high anisotropy values are exemplified for 564/570-BODIPY- and 581/591-BODIPY-GM1 , when solubilised in unilamellar vesicles composed of GM1 /sphingomyelin/ cholesterol (in the molar ratios of 3:4:3). This mixture is of particular interest because it mimics the composition of sphingomyelin-cholesterol enriched domains (Brown and London, 2000). The mole fraction of labelled gangliosides in the vesicles was 0.3%. The steady-state anisotropy values determined for 564/570-BODIPY-C5-GM1 (p), 581/591-BODIPY-C5-GM1 (p) and 581/591-BODIPY-C11-GM1 (np) at 293 K were found to be 0.150, 0.224 and 0.175, respectively. For the 564/570-BODIPY- and 581/591-BODIPY-derivatives the excitation wavelengths were 530 and 550 nm, respectively. There is, however, a significant difference between the values obtained for the head- and tail-labelled 581/591-BODIPY-GM1 , which could be expected since the BODIPY groups should be located in environments with different molecular packing. The slightly wavelength independent excitation and fluorescence anisotropies of the 564/570-BODIPY- and 581/591BODIPY gangliosides are compatible with pure electronic S0 ↔ S1 transitions in the spectral regions studied. GM1 micelles. The gangliosides are amphiphiles in which the hydrophobic part consists of a sphingosine and a fatty acid, while the bulky hydrophilic part consists of several sugar units. Thereby the gangliosides spontaneously form micelles in water (Curatolo, 1987; Sonnino et al., 1994). These micelles are relatively large with a radius of 59 Å (Sachl et al., 2008, unpublished data), which implies that the rotational correlation time of the micelles is about 200 ns. Thus the influence of rotational tumbling, of the micelles,

is negligible on the timescale of fluorescence. Also, the influence of translational diffusion is negligible. From time-resolved fluorescence depolarisation studies the reorientation correlation time of the BODIPY group can be obtained, as well as its local orientation restriction. This is exemplified in Fig. 4A, which displays the anisotropy decay of 564/570-BODIPY-C5-GM1 (p) and 581/591BODIPY-C5-GM1 (p) in GM1 micelles at 294.8 K. The anisotropy decay of the latter fits to: r(t) = 0.169 exp(−t/14.7 ns) + 0.183, and it shows that the reorienting motions of the BODIPY group are slow. The orienting restrictions of the BODIPY group are very high, with an estimated order parameter (Heyn, 1979) value of 0.70. The slow motion, as well as the high order is compatible with the large head groups of GM1 , which lead to crowding around the BODIPY group. The somewhat faster reorientation of the 564/570-BODIPY group might be ascribed to its slightly smaller size, as compared to 581/591-BODIPY. 4.3. Electronic energy transfer between BODIPY-labelled gangliosides At sufficient dilution of the probes in a lipid system, fluorescence lifetime and depolarisation experiments provide local information of the system, as has been exemplified in the previous section. With increasing concentration of BODIPY-labelled gangliosides however, energy migration among the chemically identical BODIPY:s can occur, as well as energy transfer between chemically different BODIPY:s. The influence of donor and acceptor concentrations is illustrated by the donor–acceptor pair BODIPY-FL and the 564/570BODIPY solubilised in GM1 micelles (cf. Fig. 5B). The donor group was excited at 470 nm and the emission anisotropy and the fluorescence spectra were determined for the BODIPY-FL and 564/570-BODIPY groups. Clearly, the relative fluorescence intensity of the acceptor increases when the mole fraction of donor/acceptor increases from 0.1 to 1.0 mol%, while anisotropy levels of the donor and the acceptor decrease. The decreased r-level of the donor indicates the occurrence of donor–donor energy migration, while the correspondingly larger decrease for the acceptor group is explained by a more rapid donor–acceptor energy transfer. Moreover, the efficiency of energy transfer in the lipid vesicles (Fig. 5A) is higher than in the micelles (Fig. 5B). This is compatible an inherent affinity for aggregation of gangliosides in lipid mixtures, which has previously been observed (Marushchak et al., 2007). The energy transfer process can also be demonstrated by using donor–acceptor pairs solubilised in large unilamellar vesicles of DOPC (cf. Fig. 5A). Here BODIPY-FL-C5-GM1 (p) was the donor, while the acceptor molecule was 564/570-BODIPY-GM1 (p). The vesicles were prepared from a lipid mixture, which is commonly used to

Fig. 4. (A) Time-resolved anisotropy decay of 564/570-BODIPY-C5-GM1 (p) (black) and 581/591-BODIPY-C5-GM1 (p) (red) solubilised in GM1 micelles. In both depolarisation studies the temperature was 294.8 K and the mole fraction of probe 0.033 mol%. The steady-state anisotropy values are 0.286 and 0.316 for BODIPY-FL-C5-GM1 (p) and 581/591-BODIPY-C5-GM1 (p), respectively. (B) Electronic energy transfer from BODIPY-FL-C5-GM1 (p) to 564/570-BODIPY-C5-GM1 (p) (red) and from BODIPY-FL-C5-GM1 (np) to 564/570-BODIPY-C5-GM1 (p) (black). Both donor–acceptor pairs were solubilised large unilamellar vesicles composed of DOPC:eSM:cholesterol in the molar ratio 2:1:1. The excitation wavelength was 470 nm and the emission was collected using 590 nm long pass filter, respectively. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of the article.)

I. Mikhalyov et al. / Chemistry and Physics of Lipids 159 (2009) 38–44

43

depolarisation processes are for several reasons much faster in the vesicle systems than what was found for the GM1 micelles. While only local reorienting motions contribute to the depolarisation of the BODIPY-FL group in the micelles, the energy migration and transfer processes contribute in the vesicle system. Furthermore the lipid head group packing should be denser in the micelles. The plateau value of 0.03 indicates that the residual order is very low, and corresponds to an effective order parameter of 0.29. 4.4. Is BODIPY-labelled GM1 a representative probe of GM1 ?

Fig. 5. Electronic energy transfer from BODIPY-FL-C5-GM1 (p) to 564/570-BODIPYC5-GM1 (p) in: (A) large unilamellar vesicles of DOPC:eSM:Chol in the molar ratio of 2:1:1; (B) GM1 micelles. The fluorescence spectra and anisotropies were obtained for 1 mol% of each probe. The excitation wavelength of the BODIPY-FL and the 564/570BODIPY group was 470 nm (solid lines) and 530 nm (dotted lines), respectively (A). The micelles doped with 1 mol% (solid lines) and 0.1 mol% (dotted lines) of each probe and excited at the 470 nm (B) of each probe were examined for the same wavelength settings (B). All studies were performed at 277 K.

mimic “rafts” (Bacia et al., 2005). This lipid composition was a mixture of DOPC:eSM:cholesterol in the molar ratio of 2:1:1 (Sarah et al., 2003). The emission anisotropy obtained upon excitation of the donor group (BODIPY-FL) at 470 nm shows rather high and constant r-values (≈0.18) in the wavelength region of its emission, whereas the anisotropy drops to a much lower level (≈0.02) in the wavelength region of the acceptor’s fluorescence. This is compatible with energy transfer from the BODIPY-FL group to the 564/570-BODIPY group, as well as the observed change of the fluorescence spectral intensities. The lower anisotropy is expected since reorienting motions as well as excitation transfer contribute to the depolarisation. For the direct excitation of the acceptor (@ 530 nm) the value of the emission anisotropy is similar (r ≈ 0.12) to that obtained for the donor group. At first one might even expect the values to be more similar, because the two BODIPY groups have a similar chemical structure and are located in the same region within the polar head group. However, the larger Förster radius of energy migration for 564/570-BODIPY as compared to that of BODIPY-FL is likely to enhance this process more efficiently, whereby a somewhat lower anisotropy level is obtained. The obtained time-resolved fluorescence anisotropy upon exciting the donor (BODIPY-FL-C5-GM1 (p)), while monitoring the acceptor (564/570-BODIPY-C5-GM1 (p)) is exemplified in Fig. 4B. The anisotropy decay {r(t)} is well described by the expression r(t) = 0.09 exp(−t/0.6 ns) + 0.14 exp(−t/5.3 ns) + 0.03. One should note that the anisotropy decay is faster for the BODIPYFL-GM1 (p)–564/570-BODIPY-GM1 (p) pair as compared to that for BODIPY-FL-GM1 (np)–564/570-BODIPY-GM1 (p). This is explained by a somewhat shorter average distance for the former case when both the donor and acceptor groups are located in the polar region. The

At sufficient dilution of fluorophore-labelled lipids, the probe reports on the local reorienting motions and order of the fluorophore moiety in a lipid membrane. This dilution corresponds to molar ratios of typically less than one labelled lipid molecule per 1000 non-labelled ones. At higher concentrations the average distance between randomly distributed fluorophore-labelled lipids often becomes comparable to the Förster radius of electronic energy migration, whereby the obtained fluorescence steady-state anisotropy is lowered and the decay rate of the fluorescence anisotropy is increased. However, if the labelled lipids would have a mutual affinity for aggregation to each other, the influence of electronic energy migration on the fluorescence anisotropy would be detectable, even at very low concentrations. For this reason, it is important to examine the influence of the label, which is possible by applying fluorescence depolarisation studies. As pointed out above, gangliosides in water are known to form micelles at very low concentrations (Curatolo, 1987). A possible method to reveal an eventual affinity/aggregation between the BODIPY-labelled GM1 molecules could be to study them in mixtures with non-labelled GM1 at known molar ratios. Because of the rather large value of the Förster radius for energy migration, two or more labelled GM1 in the same micelle would be able to exchange excitation energy quite efficiently, which could be detected by steady-state fluorescence depolarisation experiments. In the present work, labelled GM1 was mixed with nonlabelled gangliosides in the 564/570-BODIPY-C5-GM1 (p):GM1 ratios of 1:100, 1:1000, 1:3000 and 1:6000 for which the correspondingly obtained steady-state anisotropy values were 0.182, 0.290, 0.299 and 0.301 at 294 K. One can independently calculate the ratio at which the presence of two or more labelled lipids in a micelle becomes negligible. The probability distribution for randomly distributed molecules among the micelles obeys Poisson statistics (Tachiya, 1975). The probability of two or more molecules interacting within the same micelle is less than 0.2% for ratios higher than 1:3000. In the calculation we used the aggregation number 170, which has recently been determined for GM1 (Sachl et al., submitted for publication). The low occupation probability of two BODIPYlabelled GM1 in the same micelle for ratios of 1:3000 and higher is compatible with a constant value of the steady-state anisotropy. Thus it is highly probable that the BODIPY groups have a negligible influence on the GM1 affinity/aggregation. 4.5. Applicability of BODIPY-labelled gangliosides—an outlook The present results strongly indicate that the presence of BODIPY-groups on the GM1 lipid can be neglected with respect to its natural lipid-lipid interaction. Thus, BODIPY-labelled GM1 lipids should be suitable for probing, non-labelled GM1 with respect to spatial distributions in lipid membrane systems, as well as specific interaction sites. Such information can be obtained from depolarisation studies of systems labelled with one class of BODIPY molecule. Depending on the concentration of BODIPY-labelled lipids, information could be obtained about local motions and order, as well as spatial distributions. The latter information can also be extracted from electronic energy transfer studies between ganglio-

44

I. Mikhalyov et al. / Chemistry and Physics of Lipids 159 (2009) 38–44

sides, or lipids that have been labelled with two spectroscopically different BODIPY:s. Such an approach would be applicable in the study of the presumed existence of rafts in biological membranes (Andrey et al., 2001). Another strategy would be to study the interaction between GM1 and some specific target molecule by means of energy transfer between a BODIPY-labelled GM1 and a spectroscopically dissimilar labelled target molecule. For instance, it is thought that specific interactions occur between GM1 lipids and Alzheimer ␤-peptides in the primary steps of amyloid formation (Brown and London, 2000). Hence, it could be possible to study this interaction by using fluorophore-labelled A ␤ peptides and BODIPY-labelled GM1 . Then by preparing Cys-mutants of the peptides and labelling them with FL-BODIPY, this group can serve as the donor of electronic energy to 564/571-BODIPY- as well as 581/591-BODIPY-gangliosides. The GM1 lipid is cone-shaped, which explains why these lipids prefer to form micelles in water. This property might be useful because mixed with other lipids, one expects that the GM1 lipids would pack themselves into the more curved regions of lipid structures formed in mixtures of different lipids. For instance, the border of pores in a lipid bilayer (Abrami et al., 2000; Stanley et al., 1998) are more curved than the planar lipid bilayer. Acknowledgements This work was financially supported by the Swedish Research Council, the Royal Swedish Academy of Sciences and the Kempe Foundations (L. B.-Å. J). References Abrami, L., Fivaz, M., van der Goot, F.G., 2000. Adventures of a pore-forming toxin at the target cell surface. Trends Microbiol. 8, 168–172. Acquotti, D., Sonnino, S., Masserini, M., Gasella, L., Fronza, G., Tettamanti, G., 1986. A new chemical procedure for the preparation of gangliosides carrying fluorescent or paramagnetic probes on the lipid moiety. Chem. Phys. Lipids 40, 71–86. Aman, A.T., Fraser, S., Merritt, E.A., Rodigherio, C., Kenny, M., Ahn, M., Hol, W.G.J., Williams, N.A., Lencer, W.I., Hirst, T.R., 2001. A mutant cholera toxin B subunit that binds GM1-ganglioside but lacks immunomodulatory or toxic activity. Proc. Natl. Acad. Sci. 98, 8536–8541. Andrey, A.V., Samsonov, V., Mihalyov, I., Cohen, F.S., 2001. Characterization of cholesterol-sphingomyelin domains and their dynamics in bilayer membranes. Biophys. J. 81, 1486–1500. Bacia, K., Schwille, P., Kurzchalia, T., 2005. Sterol structure determines the separation of phases and the curvature of the liquid-ordered phase in model membranes. PNAS 102, 3272–3277.

Bagatolli, L.A., Sanchez, S.A., Hazlett, T., Gratton, E., 2003. Giant Vesicles, Laurdan, and Two-Photon Fluorescence Microscopy: evidence of lipid lateral separation in bilayers. Methods Enzymol. 360, 481–500. Bergström, F., Mikhalyov, I., Hägglöf, P., Wortmann, R., Ny, T., Johansson, L.B.-Å., 2002. Dimers of dipyrrometheneboron difluoride (BODIPY) with light spectroscopic applications in chemistry and biology. J. Am. Chem. Soc. 124, 196–204. Bogen, S.-T., de Korte-Kool, G., Lindblom, G., Johansson, L.B.-Å., 1999. Aggregation of an ␣-helical transmembrane peptide in lipid phases studied by time resolved fluorescence spectroscopy. J. Phys. Chem. B 103, 8344–8352. Bokvist, M., Lindström, F., Watts, A., Gröbner, G., 2004. Two types of Alzheimer’s beta-amyloid (1-40) peptide membrane interactions: aggregation preventing transmembrane anchoring versus accelerated surface fibril formation. J. Mol. Biol. 335, 1039–1049. Brown, D.A., London, E., 2000. Structure and function of sphingolipid- and cholesterol-rich membrane rafts. J. Biol. Chem. 275, 17221–17224. Curatolo, W., 1987. The physical properties of glycolipids. Biochim. Biophys. Acta 906, 111–136. García-Sáeza, A.J., Schwille, P., 2008. Fluorescence correlation spectroscopy for the study of membrane dynamics and protein/lipid interactions. Methods 46, 116–122. Grechishnikova, I.V., Johansson, L.B.-Å., Molotkovsky, J.G., 1996. Synthesis of new bifluorophoric probes adapted to studies of donor–donor electronic energy transfer in lipid systems. Chem. Phys. Lipids 81, 87–98. Heyn, M.P., 1979. Determination of lipid order parameters and rotational correlation times from fluorescence depolarisation experiments. FEBS Lett. 108, 359–364. Marushchak, D., Gretskaya, N., Mikhalyov, I., Johansson, L.B.-Å., 2007. Selfaggregation—an intrinsic property of GM1 in lipid bilayers. Mol. Membr. Biol. 24, 102–112. Mikhalyov, I., Gretskaya, N., Bergström, F., Johansson, L.B.-Å., 2002. Electronic ground and excited state properties of dipyrrometheneboron difluoride (BODIPY): dimers with application to biosciences. Phys. Chem. Chem. Phys. 4, 5663–5670. Neubauer, H., Gaiko, N., Berger, S., Schaffer, J., Eggeling, C., Verdier, L., Seidel, C.A.M., Volkmer, G.C.A., 2007. Orientational and dynamical heterogeneity of Rhodamine 6G terminally attached to a DNA Helix revealed by NMR and single-molecule fluorescence spectroscopy. J. Am. Chem. Soc. 129, 12746–12755. Sarah, S.L., Veatch, L., Keller, S.L., 2003. Separation of liquid phases in giant vesicles of ternary mixtures of phospholipids and cholesterol. Biophys. J. 85, 3074–3083. Sonnino, S., Kirschner, G., Ghidoni, R., Acquotti, D., Tettamanti, G., 1985. Preparation of GM1 ganglioside molecular species having homogeneous fatty acid and long chain base moieties. J. Lipid Res. 26, 248–257. Sonnino, S., Cantù, L., Corti, M., Acquotti, D., Venerando, B., 1994. Aggregative properties of gangliosides in solution. Chem. Phys. Lipids 71, 21–45. Stanley, P., Koronakis, V., Hughes, C., 1998. Acylation of Escherichia coli hemolysin: a unique protein lipidation mechanism underlaying toxin function. Microbiol. Mol. Biol. Rev. 62, 309–333. Svennerholm, L., 1974. Gangliosides. Isolation. Methods Carbohyd. Chem. 6, 464–474. Tachiya, M., 1975. Application of a generating function to reaction kinetics in micelles kinetics of quenching of lumuniscence probes in micelles. Chem. Phys. Lett. 33, 289–292. Yip, C.M., McLaurin, J., 2001. Amyloid-beta peptide assembly: a critical step in fibrillogenesis and membrane disruption. Biophys. J. 80, 1359–1371. Zander, C., Enderlein, J., Keller, R.A., 2002. Single-Molecule Detection in Solution—Methods and Applications. VCH-Wiley, Berlin/New York. Zhang, R.G., Westbrook, M.L., Westbrook, E.M., Scott, D.L., Otwinowski, Z., Maulik, P.R., Reed, R.A., Shipley, G.G., 1995. The 2.4 Å crystal structure of cholera toxin B-subunit pentamer-choleragenoid. J. Mol. Biol. 251, 550–562.