447

Alzheimers Disease: NMR Studies of Asialo (GM1) andTrisialo (GT1b) Ganglioside Interactions with A(1-40)Peptide in a Membrane Mimic Environment*

Pravat K. Mandal1 and Jay W. Pettegrew1,2

(Accepted August 21, 2003)

Amyloid peptide (A) is the major protein constituent of neuritic plaques in Alzheimers disease(AD). This peptide is an amphipathic molecule that perturbs membranes and binds to raft-likemembranes composed of gangliosides. Ganglioside GM1 binds tightly with A and it is specu-lated that GM1 inhibits A from undergoing -helix to -sheet conformational changes. Althoughthe role of gangliosides in conformational changes of A have been studied, the specic natureof these interactions have not been reported. In the present report multidimensional NMR stud-ies of ganglioside-A interactions were conducted in sodium dodecyl sulphate (SDS) micelles, amembrane-mimicking environment. These studies reveal that asialoGM1 binds specically withA in a manner which could prevent -sheet formation, but that ganglioside GT1b does not bindA. Plausible pathways for the involvement of gangliosides in amyloidogenesis are discussed.

KEY WORDS: A peptide; Alzheimers disease; ganglioside; NMR; SDS micelle.

INTRODUCTION

Alzheimers disease (AD) is a devastating neu-rodegenerative disease affecting up to 15 million indi-viduals worldwide and is associated with progressivedeposits of amyloid plaques (1) in the brain. The prin-cipal component of amyloid deposits is a 3942 aminoacid residue peptide, A (2,3), a product of proteolyticprocessing of a much larger amyloid precursor protein(APP) encoded by a gene on chromosome 21 (4). In thisreport, A refers to the 140 amino acid residue peptideunless designated otherwise. A link between A and AD

neuropathological lesions is demonstrated by the toxicityof aggregated A to neuronal cells in culture (5) and inaged brain (6). However, the molecular mechanisms ofthe neurotoxic action of A remain largely unknown.

A growing number of observations indicate that Amay alter the physicochemical properties of neuronalmembranes, including membrane uidity (7) and perme-ability to ions and nonelectrolytes (8). These ndingsstrongly suggest that at least some of the pathophysio-logical effects of A may be mediated by A-membraneinteractions. Interest in studies of the interaction betweenA and constituents of brain membranes has been fur-ther stimulated by the identication of an A-gangliosideGM1 bound form in AD brain (9,10).

Gangliosides (11,12) are sialic acidcontaining gly-cosphingolipids and consist of two main components: ahydrophobic ceramide unit, which anchors the gangliosideto the plasma membrane, and a hydrophilic oligosaccharidechain, to which one or more sialic acid groups (N-acetyl-neuraminic acid) are attached. Gangliosides are abundant

0364-3190/04/02000447/0 2004 Plenum Publishing Corporation

Neurochemical Research, Vol. 29, No. 2, February 2004 ( 2004), pp. 447453

*Part of the work will be presented (platform) at the Society forNeuroscience Meeting, Nov. 812, 2003.

1 Laboratory of Neurophysics, Department of Psychiatry, Universityof Pittsburgh Medical School, Pittsburgh, Pennsylvania.

2 Address reprint requests to: Jay W. Pettegrew, M.D., University ofPittsburgh, A712 Crabtree Hall/GSPH, 130 De Soto street, Pittsburgh,PA 15261. Tel: 412-648-8640; Fax: 412-624-9936; E-mail: pettegre@pitt.edu

448 Mandal and Pettegrew

components of neuronal membranes and are involved inimportant neurobiological events such as neurodifferentia-tion, synaptogenesis, and synaptic transmission (13). Apermeabilizes ganglioside-containing membranes and thusdisturbs ion homeostasis (14,15). Although differentresearch groups (10,1419) have reported extensive stud-ies of A-ganglioside interactions investigated by differ-ent methodologies, crucial specic binding site interactionsbetween A and gangliosides have not been addressed.

In this report multidimensional NMR studies wereconducted on gangliosides (asialoGM1 and GT1b) Apeptide interactions in a membrane mimic SDS envi-ronment. A plausible model of amyloidogenesis arisingfrom A interacting with different-size gangliosides isdescribed, which could potentially guide future drugdesign as well as treatment strategies for AD research.

EXPERIMENTAL PROCEDURE

Materials. Uniformly 15N-labeled A was purchased from Recom-binant Peptide Technologies (Atlanta, GA, USA); asialo ganglioside(asialoGM1) and trisialo ganglioside (GT1b) were purchased from Sigma(St. Louis, MO, USA.).

Preparation of A solution in SDS. 15N-labeled A peptide wasdissolved in SDSd25 solution and sonicated for 30 s, yielding a clearsolution. The nal molar ratio of peptide/SDSd25 was 1:200. The pHof the solution was 5.6. Three mg GT1b and 2 mg asialoGM1 wereused in each experiment. Each ganglioside was separately dissolvedin SDSd25 micelle (50 l) and sonicated for 20 s before addition tothe A-SDS solution. The nal molar ratio of ganglioside/A was 7.0for GT1b and 6.1 for asialoGM1.

NMR Experiments and Data Analysis. Homo-nuclear Overhauserenhancement spectroscopy (homo-NOESY), total correlation spectro-scopy (TOCSY), heteronuclear single quantum spectroscopy (HSQC), 15Nltered 3D NOESY and 15N longitudinal and transverse relaxation time(T1 and T2), and heteronuclear nuclear Overhauser enhancement (Het-NOE) NMR experiments were performed in an SDS micelle environ-ment for sequence-specic assignment and structural and dynamic studiesof A-ganglioside interactions. All NMR experiments were performed at30C using a Bruker DRX 500 MHz NMR spectrometer. NOESY experi-ments with 150 ms mixing time and TOCSY experiments were performedwith 60-ms spinlock time. Het-NOE experiments were accomplished withand without saturation of amide protons utilizing a train of 120 pulsesat 5 ms intervals for 3 s. Relaxation experiments (15N T1, T2, Het-NOE)were performed using standard pulse sequences (20) and NMR datawere processed using nmrPipe (21) and analyzed using PIPP (22) andSPARKY (23) programs. Homo-NOESY, TOCSY, HSQC, 15N ltered3D NOESY and 15N T1, T2, Het-NOE for the asialoGM1-A complexwere performed 10 days after addition of asialoGM1. This time windowallowed stable interaction between A and asialoGM1.

RESULTS

A in SDS micelles at pH 5.6 is in a random coil(residue 114), -helixI (1523), kink (2527) and

-helixII (2836) conformations as reported previously(24). The NOE connectivity of A based on our NMRdata is similar to that reported previously [PDB code1BA4] (24) and the structure was not recalculated.Dynamic data for A in a membrane mimic environ-ment is reported.

Figure 1 is a diagrammatic representation of thetwo gangliosides (asialoGM1 and GT1b) examined inthis report. Both gangliosides have four sugar moietiesin common; however, GT1b has three additional sialicacids compared to asialoGM1.

Figure 2A represents the HSQC spectra of A inSDS with asialoGM1 (red) and without asialoGM1(blue). After addition of asiloGM1, the HSQC spectrawere monitored over time. Appreciable chemical shiftswere observed for A residues E3, D7, H13, H14, Q15,and G25-K28 after addition of asialoGM1. Amide pro-tons for residues in the helical II region G29 to V40show peak splitting with equal intensity after 4 days inthe presence of asialoGM1. However, after 10 days, splitpeaks from each residue in the region (G29V40)merged into one new peak, which did not shift furthereven after 40 days. In contrast, except for V40, therewas no appreciable chemical shift of the backboneamide proton of A with addition of GT1b (Fig. 2B)even after 40 days.

Figure 3A shows the change in chemical shift (1Hand 15N) of all A residues. Signicant chemical shifts(1H and 15N) were observed for residues G29 to V40.Overall correlation times (m) were calculated based onthe 15N T1 and T2 data (data not shown) of each amideresidue for A in control (without asialoGM1) as wellas with asialoGM1. The increase of m value from 8 0.2 ns (control) to 10 0.4 ns in the asialoGM1-Acomplex reects the slower global tumbling of the lat-ter because of association of asialoGM1 with A

Figure 3B represents the difference of Het-NOE ofeach A residue with and without asialoGM1. NegativeHet-NOE for residues G25 to V40 indicates highermobility in this domain as a result of interaction withasialoGM1, this is supported by signicant changes inchemical shif (Fig. 3A) of amide protons (both 1H and15N chemical shifts).

Figure 4 is a representation of the A residues thatare perturbed in the presence of asialoGM1 in an exper-imental model membrane. The A structure (24) is takenfrom the PDB data bank (1BA4). The diagrammatic rep-resentation of the membrane and ganglioside is notintended to indicate specic conformational details, butto represent the relative size of the sugar moieties in rela-tionship to A. The residues affected in A representedin Figure 4 were not affected in the presence of GT1b.

Alzheimers Disease: NMR Studies of Asialo (GM1) and Trisialo (GT1b) Ganglioside Interactions 449

Fig. 1. A, Structure of asialo ganglioside (GM1). B, Structure of trisialo-ganglioside (GT1b).

DISCUSSION

A peptides have been reported to exert manyeffects on cell function. It has been proposed that Aexerts these effects by initially disrupting the membranelipid environment that in turn could affect the activity ofvarious proteins involved in signal transduction and ionuxes (25). A peptides may alter the neuronal mem-brane lipid environment by inducing lipid peroxidationand/or changing lipid uidity (7). It has been shown pre-viously that A inuences both annular and bulk uid-ity of synaptic plasma membranes (26). With particularrelevance to the results reported here, synaptic plasmamembranes have a relatively high concentration of gan-

gliosides with respect to other cellular membranes, andthe loss of synapses is highly correlated with the degreeof dementia in AD subjects (27). This study focused oninvestigating the interactions of two gangliosides(asialoGM1 and trisialo GT1b) (Fig. 1) with A peptide;the gangliosides have the same sugar and ceramide unit,only differing by three additional sialic acids in GT1b.Although it has been reported by circular dichroism (CD)(14,15,27,28) and surface plasma resonance (SPR) (29)that gangliosides induce conformational changes in Apeptide in phosphate buffer saline (PBS) medium, thespecic molecular details of these interactions were notreported. Based on the CD data it was assumed that GM1induces more helical content in A peptide than GT1b.

450 Mandal and Pettegrew

Fig. 2. A, HSQC spectra of A, in which blue denotes the control (without asialoGM1) and red denotes the A with asialoGM1 after 10 days.Residues marked with star (*) sign show appreciable chemical shift resulting from asialoGM1. B, HSQC spectra of A, in which blue denotesthe control (without GT1b) and red represents A with GT1b after 10 days.

Fig. 3. A, Chemical shift changes of the A amide protons and nitrogens (1H shift change denoted by O and 15N shift change denoted by )resulting from addition of asialoGM1. B, Plot of the Het-NOE difference () ([AasialoGM1 control]) of each amino acid residue.

Alzheimers Disease: NMR Studies of Asialo (GM1) and Trisialo (GT1b) Ganglioside Interactions 451

Fig. 4. Schematic diagram of A-asioloGM1 interaction on theextracellular side of the plasma membrane. The diagrammaticrepresentation of the membrane and ganglioside is not intended toindicate specific conformational details, but to represent the relativesize of the sugar moieties in relationship to A. A is representedby a backbone trace of an NMR determined structure in the PDBdata bank (1BA4). A residues whose amide protons are perturbedby asialoGM1 are shown in red; A residues whose amide protonsare not perturbed by asialoGM1 are shown in blue.

Although ganglioside-induced conformational changes inA are beginning to be demonstrated, to the best of ourknowledge there are no structural studies concerningganglioside-A interactions either in PBS or in a mem-brane mimic environment.

Based on the amide proton chemical shift change(Fig. 3) (identied by red color), it is likely that asiloGM1

docks between region I (E3V24) and region II(G29V40) of A. The residues (G29V40) in the helix(region II) after the kink undergo major changes in chem-ical shift and exhibit higher mobility after addition ofasialoGM1 as shown by negative Het-NOE (Fig. 3B). Asa result of interactions with asialoGM1, this structuredhelical region (G29V40) is converted to an unstructuredrandom coil, as evidenced by the loss of characteristic-helical NOE pattern, that is, I32H M35NH. It isimportant to note that the asialoGM1-A sample remainsstable at room temperature. The perturbation of the amideproton chemical shifts of A may indicate that sugar unitsof asialoGM1 are inserted between A regions I and II.Detailed NOE analysis between asialoGM1 and A arein progress. Insertion of asialoGM1 between regions I andII of A would prevent these two regions from comingclose together and would likely prevent -sheet forma-tion. The bulky sugar moieties due to three sialic acidsprevent GT1b from inserting between the two A regions.

In the membrane system and the analytical tech-nique used by Ariga et al. (29), A had a lower afnityfor asialoGM1 gangliosides than gangliosides withseveral sialic acids. In the membrane system and NMRanalytical technique used in our study just the oppositewas found.

There are two potential reasons for the disagreementbetween our ndings and the ndings by Ariga et al. (29).First, the membrane system used by Ariga et al. (29) wascomposed of DPPC (dipalmitoylphosphatidylcholine) anda ganglioside. We looked at a membrane system com-posed of SDS plus ganglioside in solution. Therefore theconformation of a ganglioside molecule exposed outsidethe membrane environment could be different for aDPPC versus SDS membrane. Next, one needs to con-sider the conformation of A, which will inuence itsinteractions with the exposed part of a ganglioside. In theanalytic system used by Ariga et al. (29), there is no solu-tion and the interaction between A and ganglioside isunder solid state conditions, whereas the analytical NMRmeasurements in this study were conducted underaqueous conditions. It would not be surprising if the con-formation of A and the exposed ganglioside are verydifferent under these very different experiments condi-tions. Only further studies carefully investigating the roleof the membrane using membranes composed of differ-ent phospholipid compositions and investigating the con-formational changes in the exposed ganglioside and Apeptide will address these concerns. The NMR dataexplains two previous important ndings: (i) Why GM1does not bind the shorter A(128) peptide (16). In ourmodel the sugar moiety of asialoGM1 becomes sand-wiched between the two regions of the A peptide. In

452 Mandal and Pettegrew

the case of A(128), the absence of the main bindingregion G29V40 would prevent GM1 binding withA(128). (ii) Why gangliosides follow the order (GT1b GD1a GD1b GM1) to induce -sheet formation (28)as evidenced from CD measurements. We believe, basedon our NMR data, as the size of the gangliosides increase(GT1b GD1a GD1b GM1) with the addition ofsialic acids, it is difcult to accommodate the bulky gan-glioside between regions I and II of A and -sheet for-mation proceeds. The smaller asialoGM1 and GM1 areable to bind A and thereby prevent -sheet formation.As reported previously, the amount of GM1 decreaseswhile GT1b increases with aging (3033). Whenever theloss of protective ganglioside GM1 and the increase inunprotective GT1b reaches a signicant level, the pos-sibility of -sheet formation of A peptide is elevated,which could lead to the formation of senile plaques. Itmay be noted that other metabolites such as glycerophos-phocholine (34,35) and cholesterol (3638) are elevatedin AD brain. The possible role of these metabolites inaltering GM1-A interactions also needs careful and thor-ough investigation.

ACKNOWLEDGMENT

Authors acknowledge the nancial support from the grants(5R01HD039799 and 5R01MH058308). Thanks to R. J. McClure forsuggestions. Thanks to Virgil Simplacu for arranging NMR time andProfessor Chien Ho for permitting us to use CMU NMR facility.

REFERENCES

1. Iversen, L. L., Mortishire-Smith, R. J., Pollack, S. J., and Shearman,M. S. 1995. The toxicity in vitro of beta-amyloid protein. Biochem.J. 311:116.

2. Glenner, G. G. and Wong, C. W. 1984. Alzheimers disease: Initialreport of the purication and characterization of a novel cere-brovascular amyloid protein. Biochem. Biophys. Res. Commun.120:885890.

3. Masters, C. L., Simms, G., Weinman, N. A., Multhaup, G., McDon-ald, B. L., and Beyreuther, K. 1985. Amyloid plaque core proteinin Alzheimer disease and Down syndrome. Proc. Natl. Acad. Sci.USA 82:42454249.

4. Kang, J., Lemaire, H. G., Unterbeck, A., Salbaum, J. M., Masters,C. L., Grzeschik, K. H., Multhaup, G., Beyreuther, K., and Muller-Hill, B. 1987. The precursor of Alzheimers disease amyloid A4protein resembles a cell-surface receptor. Nature 325:733736.

5. Lambert, M. P., Barlow, A. K., Chromy, B. A., Edwards, C.,Freed, R., Liosatos, M., Morgan, T. E., Rozovsky, I., Trommer,B., Viola, K. L., Wals, P., Zhang, C., Finch, C. E., Krafft, G. A.,and Klein, W. L. 1998. Diffusible, nonbrillar ligands derivedfrom Abeta1-42 are potent central nervous system neurotoxins.Proc. Natl. Acad. Sci. USA 95:64486453.

6. Geula, C., Wu, C. K., Saroff, D., Lorenzo, A., Yuan, M., andYankner, B. A. 1998. Aging renders the brain vulnerable toamyloid beta-protein neurotoxicity. Nature Med. 4:827831.

7. Muller, W. E., Koch, S., Eckert, A., Hartmann, H., and Scheuer,K. 1995. beta-Amyloid peptide decreases membrane uidity. BrainRes. 674:133136.

8. Arispe, N., Rojas, E., and Pollard, H. B. 1993. Alzheimer dis-ease amyloid beta protein forms calcium channels in bilayer mem-branes: Blockade by tromethamine and aluminum. Proc. Natl.Acad. Sci. USA 90:567571.

9. Yanagisawa, K., Odaka, A., Suzuki, N. and Ihara, Y. 1995. GM1ganglioside-bound amyloid beta-protein (A beta): A possibleform of preamyloid in Alzheimers disease. Nature Med. 1:10621066.

10. Kakio, A., Nishimoto, S., Yanagisawa, K., Kozutsumi, Y., andMatsuzaki, K. 2002. Interactions of amyloid beta-protein withvarious gangliosides in raft-like membranes: Importance of GM1ganglioside-bound form as an endogenous seed for Alzheimeramyloid. Biochemistry 41:73857390.

11. Paola, B. and Sandro, S. 1997. Dynamics and spatial organizationof surface gangliosides. Trends Glycosci. Glycotechnol. 9:433445.

12. Kohji, K. and Yutaka, S. 2001. Involvement of lipid raft signallingin ganglioside-mediated neural function. Trends Glycosci. Glycotech-nol. 13:587594.

13. Nagai, Y. 1995. Functional roles of gangliosides in bio-signaling.Behav. Brain Res. 66:99104.

14. McLaurin, J. and Chakrabartty, A. 1996. Membrane disruptionby Alzheimer beta-amyloid peptides mediated through specicbinding to either phospholipids or gangliosides: Implications forneurotoxicity. J. Biol. Chem. 271:2648226489.

15. McLaurin, J., Franklin, T., Fraser, P. E., and Chakrabartty, A.1998. Structural transitions associated with the interaction ofAlzheimer beta-amyloid peptides with gangliosides. J. Biol. Chem.273:45064515.

16. ChooSmith, L. P. and Surewicz, W. K. 1997. The interactionbetween Alzheimer amyloid beta (1-40) peptide and gangliosideGM1-containing membranes. FEBS Lett. 402:9598.

17. Yanagisawa, K. and Ihara, Y. 1998. GM1 ganglioside-boundamyloid beta-protein in Alzheimers disease brain. Neurobiol.Aging 19:S65S67.

18. Ariga, T. and Yu, R. K. 1999. GM1 inhibits amyloid beta-protein-induced cytokine release. Neurochem. Res. 24:219226.

19. Kakio, A., Nishimoto, S., Kozutsumi, Y., and Matsuzaki, K. 2003.Formation of a membrane-active form of amyloid beta-protein inraft-like model membranes. Biochem. Biophys. Res. Commun.303:514518.

20. Farrow, N. A., Muhandiram, R., Singer, A. U., Pascal, S. M.,Kay, C. M., Gish, G., Shoelson, S. E., Pawson, T., Forman-Kay,J. D. and Kay, L. E. 1994. Backbone dynamics of a free andphosphopeptide-complexed Src homology 2 domain studied by15N NMR relaxation. Biochemistry 33:59846003.

21. Delaglio, F., Grzesiek, S., Vuister, G. W., Zhu, G., Pfeifer, J., andBax, A. 1995. NMRPipe: A multidimensional spectral processingsystem based on UNIX pipes. J. Biomol. NMR 6:277293.

22. Garrett, D. S., Powers, R., Gronenborn, A. M., and Clore, G. M.1991. A common sense approach to peak picking two-, three- andfour-dimensional spectra using automatic computer analysis ofcontour diagrams. J. Magn. Reson. 95:214220.

23. Goddard, T. D. and Kneller, D. G. 1994. SPARKY 3, Universityof California, San Francisco.

24. Coles, M., Bicknell, W., Watson, A. A., Fairlie, D. P., and Craik,D. J. 1998. Solution structure of amyloid beta-peptide(1-40) in awater-micelle environment: Is the membrane-spanning domainwhere we think it is? Biochemistry 37:1106411077.

25. Roth, G. S., Joseph, J. A., and Mason, R. P. 1995. Membranealterations as causes of impaired signal transduction inAlzheimers disease and aging. Trends Neurosci. 18:203206.

26. Avdulov, N. A., Chochina, S. V., Igbavboa, U., OHare, E. O.,Schroeder, F., Cleary, J. P., and Wood, W. G. 1997. Amyloidbeta-peptides increase annular and bulk uidity and induce lipidperoxidation in brain synaptic plasma membranes. J. Neurochem.68:20862091.

Alzheimers Disease: NMR Studies of Asialo (GM1) and Trisialo (GT1b) Ganglioside Interactions 453

27. Iqbal, K. and Grundke-Iqbal, I. 2002. Neurobrillary pathologyleads to synaptic loss and not the other way around in Alzheimerdisease. J. Alzheimers Dis. 4:235238.

28. Matsuzaki, K. and Horikiri, C. 1999. Interactions of amyloidbeta-peptide (1-40) with ganglioside-containing membranes. Bio-chemistry 38:41374142.

29. Ariga, T., Kobayashi, K., Hasegawa, A., Kiso, M., Ishida, H., andMiyatake, T. 2001. Characterization of high-affinity bindingbetween gangliosides and amyloid beta-protein. Arch. Biochem.Biophys. 388:225230.

30. Op Den Velde, W. and Hooghwinkel, G. J. 1975. The brain gan-glioside pattern in presenile and senile dementia. J. Am. Geriatr.Soc. 23:301303.

31. Svennerholm, L. and Gottfries, C. G. 1994. Membrane lipids,selectively diminished in Alzheimer brains, suggest synapseloss as a primary event in early-onset form (type I) anddemyelination in late-onset form (type II). J. Neurochem. 62:10391047.

32. Svennerholm, L. 1994. Ganglioside loss is a primary event inAlzheimer-disease type-I: Biological function of gangliosides.Prog. Brain Res. 101:391404.

33. Svennerholm, L., Bostrom, K., and Jungbjer, B. 1997. Changesin weight and compositions of major membrane components ofhuman brain during the span of adult human life of Swedes. ActaNeuropathol. (Berl.) 94:345352.

34. Pettegrew, J., Kopp, S., Minshew, N., Glonek, T., Feliksik, J., Tow, J.,and Cohen, M. 1987. 31P nuclear magnetic resonance studies ofphosphoglyceride metabolism in developing and degenerating brain:Preliminary observations. J. Neuropathol. Exp. Neurol. 46:419430.

35. Nitsch, R. M., Blusztajn, J. K., Pittas, A. G., Slack, B. E., Growdon,J. H., and Wurtman, R. J. (1992). Evidence for a membrane defectin Alzheimer disease brain. Proc. Natl. Acad. Sci. USA89:16711675.

36. Kuo, Y. M., Emmerling, M. R., Bisgaier, C. L., Essenburg, A. D.,Lampert, H. C., Drumm, D., and Roher, A. E. 1998. Elevated low-density lipoprotein in Alzheimers disease correlates with brainabeta 1-42 levels. Biochem. Biophys. Res. Commun. 252:711715.

37. Yanagisawa, K. and Matsuzaki, K. 2002. Cholesterol-dependentaggregation of amyloid beta-protein. Ann. NY Acad. Sci.977:384386.

38. Puglielli, L., Tanzi, R. E., and Kovacs, D. M. 2003. Alzheimersdisease: The cholesterol connection. Nature Neurosci. 6:345351.