Proc. Nat. Acad. Sci. USAVol. 71, No. 9, pp. 3701-3705, September 1974

13C Nuclear Magnetic Resonance Spectroscopy of Native andRecombined Lipoproteins

(lipid-protein interactions/hydrophobic binding/high density lipoproteins/liposome structure)

GERD ASSMANN*, ROBERT J. HIGHETt, EDWARD A. SOKOLOSKIt, AND H. BRYAN BREWER, JR.*

National Heart and Lung Institute, National Institutes of Health, Bethesda, Maryland 20014

Communicated by Donald S. Fredrickson, September 17, 1973

ABSTRACT 13C nuclear magnetic- resonance data onnative and recombined lipoproteins are reported. [Methyl-'3Clphosphatidylcholine, [methyl-'3Cjsphingomyelin, 1,2-[dioleoyl-1-"Cl-sn-phosphatidylcholine and cholesteryl-[1-13C]oleate were enriched with 90% carbon-13 in respec-tive molecules by chemical synthesis and used for recom-bination experiments with high density lipoprotein apo-proteins. Relaxation times for these specifically enrichedlipids in organic solvents, 2H20, and lipid-protein com-plexes isolated by ultracentrifugal flotation, were mea-sured. The results show that both the phosphatidylcholineand sphingomyelin polar headgroups have the same hy-drophilic environment in either sonicated lipid particles orreassembled lipoproteins, and suggest that ionic interac-tion of lipid and apolipoproteins is of minor importance inthe formation of plasma lipoprotein complexes. Our ex-periments indicate that 13C nuclear magnetic resonancespectroscopy will contribute to the understanding of lipid-protein interaction in lipoproteins and membranes.

From a theoretical standpoint, nuclear magnetic resonance(NMR) techniques appear to have great potential for un-ravelling the details of the lipid-protein interaction involvedin plasma lipoproteins and cellular membranes. Proton NMRspectroscopy has been used in the past to investigate themolecular organization of human serum low (LDL) and high(HDL) density lipoproteins (1-4). Results thus far, however,have been relatively modest because of difficulties in inter-pretation of the chemical shifts and the structural complexityof serum lipoproteins. In order to investigate the way inwhich the covalent structure of lipoprotein-apoproteinsdirects both conformation and biological function, we haveempolyed the techniques of '3C NMR spectroscopy. Therelatively sharp resonances and the wide range of chemicalshifts from '3C nuclei are well suited for studies of lipid-lipidand lipid-protein interactions. Moreover, the '3C nucleuscauses no perturbation of the lipid structure and specificcarbon atoms in the hydrophobic and hydrophilic site of thelipid molecules can be selectively enriched with the I3C isotopeover its natural abundance level of 1.1%. Thus, the problemsof poor signal to noise and overlapping resonances in thespectra of lipoproteins and other biological systems can beavoided by chemical or biosynthetic incorporation of en-riched '3C nuclei. The information derivable from '3C relaxa-tion measurements relates closely to factors controlling mo-

Abbreviations: NMR, nuclear magnetic resonance; D, deuterium;LDL, low density lipoproteins; HDL, high density lipoproteins; Ti,spin-lattice relaxation time; LCAT, lecithin: cholesterol acyl-trans-ferase; PC, phosphatidylcholine; SPM, sphingomyelin.* Molecular Disease Branch.t Laboratory of Chemistry.

3701

lecular motions, such as overall and local molecular geometry,bonded and nonbonded interactions, solvation effects, etc. (5).In the present study, we report the chemical shifts and spin-lattice relaxation times (Ti's) for 1'C nuclei in the polar headgroup of the phosphatidylcholine and sphingomyelin mole-cules comparing intact lipoproteins, recombined lipoproteins,sonicated lipid vesicles, and lipids in organic solvents. Inaddition, the Ti's for phosphatidylcholine, enriched at thecarboxy terminal end of the alkyl chains, cholesteryl-[1-'8C]-oleate in organic solvent and as a constituent of HDL, ob-tained by recombination technique and enzymatic transfer[lecithin: cholesterol acyltransferase] from 1,2- [dioleoly-1-13C, 1-14C]-sn-phosphatidylcholine, have been measured.The 13C NMR data obtained in these studies provide in-

formation concerning the structure, environment, and dy-namics of the polar phospholipid headgroup region and thecarboxy terminal end of alkyl chains in lipoproteins andmodel membrane systems.

MATERIALS AND METHODS

NMR Spectroscopy. All NMR spectra were obtained on aVarian XL-100 spectrometer equipped with a Digilab Fouriertransform system. Typically, pulses of 40 Asec (about 60°),sampling rates of 12 KHz and Fourier transforms of 16,000points were used to prepare spectra with resolution of approxi-mately 1 Hz. Broad band noise-modulated proton decouplingwas accomplished by the Varian gyrocode decoupler. Ti'swere determined from the slope of the usual semilog plot ofdata from a 180'-r-90' sequence (6). Chemical shifts areconsidered accurate within 0.1 ppm; Ti's are consideredaccurate within 10%. The T1 measurements were made onphospholipids and reassembled lipoproteins in 2H20 (D20)buffer thoroughly deoxygenated with nitrogen. CD30H andCDC13 samples of phospholipids and cholesterol esters weredegassed by repeated freeze-pump-thaw cycles.

Preparation ofHDL Apoproteins. Plasma was collected fromtwo normal male fasting donors (K. Z., G. A.) by plasma-phoresis in EDTA and was used within 2 hr. HDL was ob-tained by ultracentrifugal flotation in KBr between densities1.063 and 1.21 g/ml and washed once by flotation at 1.210g/ml (7). The HDL preparations gave a single precipitationline on immunoelectrophoresis against anti-human wholeserum (Hyland) and anti-human HDL (8). They did notreact with antisera to human low density lipoproteins oralbumin. HDL protein (apoHDL) was obtained by delipida-tion with chloroform-methanol (9). Less than 0.5% phos-pholipid (w/w) remained after delipidation of HDL. ApoA-Iwas obtained by chromatography of apoHDL in 6 M urea onSephadex G-200 (Pharmacia, superfine) according to the

Dow

nloa

ded

by g

uest

on

Oct

ober

16,

202

0

3702 Biochemistry: Assmann et al.

procedure of Scanu et al. (10). ApoA-II was obtained bychromatography of apoHDL in 6 M urea on DEAE cellulose(Whatman, DE-52) by modification of the method of Shoreand Shore (1i, 9). The apoprotein preparation was judgedpure on the basis of polyacrylamide gel electrophoresis in 8M urea at pH 9.4 and 2.9 (12, 13), Edman amino-terminalanalysis (14) combined with identification of the phenyl-thiohydantoin amino acids by gas liquid chromatography(15, 16), and by its unique amino-acid composition (9, 17, 18).ApoA-II was reduced and carboxymethylated essentiallyas previously described (9). Cyanogen bromide cleavage ofapoA-II was performed at room temperature for 48 hr in 70%formic acid using a 500-fold molar excess of cyanogen bromideto methionine. Following cleavage, the cyanogen bromidepeptides were reduced and carboxymethylated and separatedby gel filtration on Sephadex G-75 (19).

Preparation of Lipids. Egg yolk phosphatidylcholine (PC)was prepared by the method pf Dawson (20) and stored i4absolute ethanol under nitrogen at -20° until use. Thismaterial exhibited a single spot on two dimensional thin-layerchromatography using chloroform-methanol-ammonium hy-droxide-water 180:105:7.5:7:5, in the first dimension andchloroform-methanol-ammonium hydroxide-water 120:160:5:5 in the second dimension. To visualize the lipid, plateswere sprayed with a phospholipid reagent (21) and charred.Sphingomyelin (SPM) isolated from bovine brain (Sigma)was homogenous as judged by thin-layer chromatography inseveral solvent systems. Both lipids were subsequently usedfor chemical introduction of a 1'C and 14C label into theircholine moieties. PC and SPM were demethylated by treat-ment with sodium benzene thiolate to N-dimethyl-phos-phatidyl-ethanolamine and ceramide-1-phosphoryl-N-di-methylethanplamine, respectively (22), and methylationwith ['8C]methyl iodide (90.8% isotopic purity) (Merck,Sharpe & Dohme) was carried out according to the proceduresof Stoffel, et al. (22, 23). Before methylation, ["8C]methyliodide was mixed with [14C]methyl iodide or ['Himethyliodide (Amersham/Searle Corp.) in a molar ratio of 500:1,respectivejy.

Lysolecithin (Sigma) was acylated with the chloride of[1-'8C]oleic acid (Applied Science) in chloroform with di-methylaminopyridine as base (23). Before reaction withoxalylchloride (Eastman) (24) [1-'3C]oleic acid was mixedwith [1-_4C]oleic acid (Amersham/Searle Corp.) in a molarratio of 500:1. Acylation of cholesterol with [1-18C, 1-14C]-oleoylchloride was carried out in benzene.The labeled products were isolated by silica gel column

chromatography (25) and their purity ascertained by radio-thin-layer chromatography. Phospholipase A2 (Boehringer,Crotalus adamanteus) digestion of the purified PC and sub-sequent separation of oleic acid and lysolecithin reactionproducts by thin-layer chromatography (solvent system:chloroform/methanol/H20 65/25/4) revealed their ratio ofradioactivity to be 3.3:1, respectively. The positional im-purity of phosphatidylcholine (77% of [1-_8C, 1-_4C]oleicacid in # position, 23% in a position) might have been causedby acyl migration in the lysolecithin during storage or acyla-tion reaction. [Methyl-"3C, 'H]phosphatidylcholine (specificactivity 3.5 X 106 cprm/Mmol), 1,2-[dioleoyl-1-'8C, 1-14C]-sn-phosphatidylcholine (specific activity 4 X 106 cpm/.smol),[methyl-"8C, '4C]sphingomyelin (specific activity 2.8 X 105cpm/jumol), and cholesteryl-[1-_3C,1-14C]oleate [specific ac-tivity 9 X 105 cpm/,umol] were approximately 90% enriched

in carbon-13 in the respective molecules as ascertained bymass spectrometry.

Radioactive lipids were stored as ethanol or benzene stocksolutions and aliquots evaporated to dryness under a streamof nitrogen.

Phospholipid Dispersions. To the dry phospholipids (15-50 mg) was added 50 mM NH4HCO3 buffer, pH 8.8 (1-3 ml)in H20 or nonbuffered H20. This mixture was initially dis-persed on a Vortex mixer and then subjected, in an ice bathand under nitrogen, to sonification with a microtip probe of aBranson sonifier. At a probe setting of 5, 60-sec bursts weredelivered, separated by intervals of 30 sec for a total time of15-30 min. The resulting slightly opalescent dispersions con-tained no degradation products of PC or SPM as determinedby thin-layer chromaotgraphy. The concentration of PC andSPM was estimated from radioactivity determinations of thedispersions. Lipid samples used for NMR analyses weresonicated in deoxygenated D20 buffer and transferred undernitrogen to 12-mm nuclear magnetic resonance tubes.

Ultracentrifugal Flotation of Phospholipid-Protein Complexes.Phospholipid dispersions were added dropwise to a solutionof apoprotein in 0.1 M NH4HCO, buffer, pH 8.8, and theresulting mixture incubated for 60 min at room temperaturein a Dubnoff shaking bath. In recombination experimentswith cholesteryl ester the technique devised by Hirz and Scanuwas used, employing sonication of the lipid-protein mixtureat a temperature above the liquid-crystal transition of thecholesteryl esters in the lipid mixture (26, 27). The mixturewas adjusted to a density of 1.063 g/ml by the addition ofsolid KBr and was spun for 20 hr in a 65 rotor at 60,000 rpm.The supernatant fraction contained excess, unbound lipidwhile the infranatant contained the phospholipid-apoproteincomplex and any unbound protein. The infranatant wasadjusted to a density of 1.250 g/ml by the addition of solidKBr and was spun for 40 hr at 60,000 rpm. The lipid-apo-protein complex was analyzed by '8C NMR spectroscopy,circular dichroism spectroscopy, gel column chromatography(Bio-gel A-1.5m, 200-400 mesh), and the molar lipid to proteinratio determined (28). All ultracentrifugal supernatant andinfranatant fractions were analyzed for phospholipid andcholesteryl ester by determination of radioactivity; proteinwas measured by the procedure of Lowry, et at. (29) using abovine-serum albumin standard.

Reaction of Lecithin: Cholesterol Acyltransferase (LCAT, EC2.3.1.43). In order to enrich specifically the carboxy terminalfatty-acid chain of cholesteryl esters in naturally occurringhigh density lipoprotein we made use of the plasma enzymelecithin: cholesterol acyltransferase. It catalyzes the transferof fatty acids from the 2 position (30) of lecithin to the choles-terol of plasma liproproteins (preferentially HDL). [Dioleoyl-1-18C, 1-14C]-n-phosphatidylcholine was sonicated in H20(1.5 mM) and added in a volume of 3 ml to 100 ml of humanplamsa. Incubation was performed for 12 hr at 37°. HDL wasisolated by flotation between KBr densities 1.063 and 1.21g/ml. After dialysis, appropriate aliquots were separated forpaper electrophoresis (31), delipidation (9), gel chroma-tography (28) and 'IC NMR analysis.

RESULTSThe proton-decoupled "IC spectrum of HDL at naturalabundance is shown in Fig. la. Although the HDL containapproximately equal amounts of lipid and protein on a weightbasis, the prominent signals arise from the lipid constituents.The lack of clearly resolved protein carbons is due to the

Proc. Nat. Acad. Sci. USA 71 (1974)

Dow

nloa

ded

by g

uest

on

Oct

ober

16,

202

0

13C NMR Spectroscopy of Lipoprotein 3703

TABLE 1. Chemical shift and relaxation times of nativelipoproteins, recombined lipoproteins, and phospholipids

Chemical Relaxationshift* (ppm) time (sec)

[Methyl-13C]sphingomyelinin CD30Hin CDC13in D20recombined withapoHDL

recombined with apoA-I

recombined with apoA-II

recombined with Cys-carboxymethylatedapo A-II

54.654.454.6

54.7

0.800.110.31

0.29

54.7

54.7

54.7

[Methyl-'3C]phosphatidylcholine (egg yolk)in CD30H 54.7in CDCks 55.4in D20 54.6recombined with apo

A-II 54.6recombined withCOOH-terminalCNBr peptide of apoA-II 54.6

recombined with apoA-I

recombined with apoHDLt

54.6

55.0

1,2-[dioleoyl-1-'3C]phosphatidylcholine

in CD30H a carboxyl 174.6

,B carboxyl 174.3in CDC13 a carboxyl 173.5

, carboxyl 173.1in D20 a carboxyl 173.9

,B carboxyl 174.2in HDL a,,B carboxyl 174.2recombined with apoHDL

Cholesteryl [1-13C]oleatein CDC13in HDL (obtained byLCAT reaction)

recombined with apo

HDLt

0.32

0.29

1.310.250.45

0.50

0.45

0.50

a carboxyl 9.8,6 carboxyl 9 .4a carboxyl 4, carboxyl 3.3a carboxyl 2. 0,6 carboxyl 1.9

a,# carboxyl 1 .8

a,,6 carboxyl 174. 2 a,# carboxyl 1. 9

173.9

171.9

171.9

7.4

1.5

0.5

Reassembly experiments were done as described in Methods.* Chemical shift data were measured versus dioxane external

and are reported relative to tetramethylsilane (TMS) by use ofthe relation 8TMS = 8dioxane + 67.4.

t Chemical shift data and relaxation measurements refer to thesame experiment, in which apo-HDL was recombined with[methyl-"3C]phosphatidylcholine and cholesteryl [1-13C]oleateaccording to the technique of Scanu (25, 26).

relatively low molar protein concentration and the fact thatthe majority of protein carbons have short spin-spin relaxa-tion times and consequently broad resonances (32). The mostprominent lipid resonances can be assigned to the cholinemoieties of respective phospholipids (mostly phosphatidylcholine and sphingomyelin), the fatty-acid envelope andcertain cholesterol carbon atoms (33). It should be noted thatchemical shifts for defined carbon atoms in native HDL are

b I.c1

.].I

d

*:i

11:

F 1(a) ICNI\ sLSite triin of niative I1) at ntatral a1b)u-dnce ;eak 1 ( ) p)Iwd7 ( 1, . p) wd2ereassigpned to thecarboxvl (crl-olos. of panhdl~l\1( oline and clodeso eate,reslpectlively (ccIRolltx oher ITS0oInuIIt' have previouly beena2ssiginied 2I( 7,0)fttI ne, 4 Ot Ihi 12-nin ubet 10 itg of Ei )T-

rOtlinO pr1011). (b) Loio1l 10 (dee) ro-iherogramn of nalive1-11)1 enr1iched1 ith hlle-teiI lLC, 1'Cjoleae and I,2-[ -bdJ,//1 ;( f',t1(', phopht)}Itn~l\ileho in. {I' adlio-tdino lay1e!erchronuatogai n of 1, [diodo! 1::( ',1-''(m-so-pI )51)h3tid1(h()-line. utsed for LAT reactilon (Solvent Nvstcll: chloroform1uIllteth-111n11 water (6_ 25 4 , (origin left, fro it rIiglt. 1d) Had oihimi-layer(^l11'>111t~~~t f.)gl';llll f > 1 1 ) 1j~~1itin X 1 1^ 1l (^ 1.'1Bi131(}1 ~~ l,C1rolua(to1g1tran1 ofHI ) lipi,1dR eX1t1 'Ct, enrichedwii(t cholesterlI-

1-i:{, .'4J~oleatelb\- LCAT reacltion (szolvnit 'vsteni: petrodcuimether diethvl et her acetic acid 9t Mtt ) origfileeft, front right.(o ) ' ( NMIs1e(Irtn'P@11111 of naitive I I I )L enriched with1.2I-[dolooql-

1 C] (olea t e (2 ) dI I I 1(I UI S ref 0P

subject to certain variations, when based on comparison withrespective molecules in different solvents. This could bedemonstrated for the carboxyl carbon of cholesterol oleatewhich gives different chemical shifts in native HDL as com-pared to organic solvent (Table 1).The unequivocal assign-ment of the cholesteryl ester carboxyl carbon and the choline-carbon resonance might in the future allow quantitation of themost abundant lipids in HDL. The 13C NMR spectrum ofHDL in which cholesteryl oleate is 18C enriched in the fatty-acid carboxyl atom by LCAT reaction, using 1,2-[dioleoyl-1-13C,1-'4C]-sn-phosphatidylcholine as substrate, is shown inFig. le. The spectrum is simplified to two well-defined reso-nances under conditions where no natural abundance 13Cresonances can. be detected. These resonances were assignedto the PC and cholesteryl ester fatty-acid carboxyl carbons bycomparison of the peak height with the relative radioactivityin the respective lipids (Fig. ld). In addition, cholesteryl[1-'3C]oleate, recombined with apoHDL in the presence of

Proc. Nat. A cad. Sci. USA 71 (1974)

Dow

nloa

ded

by g

uest

on

Oct

ober

16,

202

0

3704 Biochemistry: Assmann et al.

[methyl-"C]phosphatidylcholine according to the technique ofScanu et al. (26, 27), gives essentially identical chemicalshifts (Table 1) though a slightly shorter relaxation time.

Chemical shifts and relaxation times of [methyl-"C]-sphingomyelin, [methyl-"C]phosphatidylcholine and 1,2-[dioleoyl-1-"3C, 1-_4C]-sn-phosphatidylcholine in CD30H,CDC13, D20 and recombined with individual HDL-apopro-teins are reported in Table 1. 'IC NMR relaxation experi-ments on reassembled lipid-proteins complexes were done withsamples isolated by ultracentrifugal flotation between KBrdensities 1.603-1.25 g/ml. Detailed physical studies on thelipid-protein reassembled particles are presented in theaccompanying paper (28). Temperature of the samples wasapproximately 300; slight variation from sample to samplecannot be excluded because of the uncertainty of controllingthe heat output of the decoupler system.

DISCUSSION

The molecular details of the structural organization of theserum lipoprotein classes are poorly understood. Basic prin-ciples of protein-protein, protein-lipid and lipid-lipid inter-actions have not yet been resolved. Such information isexpected to be derived from the physicochemical analyses ofnormal intact particles and their individual and reassembledconstituents. The principal HDL lipid components are phos-pholipids and cholesteryl esters which together account for85-90% of the total HDL lipids (34). The protein moiety isnow known to consist of several components (10, 11, 35).Two of these, designated as apoA-I and apoA-II (36) com-prise 85-90% of the total HDL protein content. Three otherdistinct peptides designated as C-I, C-Il, and C-III (37),account for the rest of the protein. The primary sequences ofapoA-II (38), C-I (39) and C-III (40) have been described.

Recently, it has been demonstrated that study of the spin-lattice relaxation time of "IC provides a promising probe intothe internal mobility of organic molecules (5). Studies ofdipalmitoyl- and dioleoylphosphatidylcholine at naturalabundance of "IC have yielded relaxation times for the car-boxyl carbons, the methylene envelope and the terminalmethyl carbons of fatty acids in D20 (41). The results arequalitatively consistent with previous information on thenature of the lipid vesicles in D20, a bilayer with more mobilealkyl chains showing substantially longer relaxation timesthan the polar portions of the molecules (42-44). However,the spectra obtained below the transition temperature didnot allow the resolution of the carbon atoms of the fatty-acidchain, nor the differentiation of the two acyl carboxyl carbons.In order to overcome such difficulties, and to allow the studyof lipids in the presence of the complexing proteins, we havestudied specifically labeled lipids enriched with C-13.

Since lecithin and sphingomyelin are the most abundantphospholipids in lipoproteins and cholesteryl esters the pre-dominant neutral lipid in HDL, we have chosen to enrichspecifically those lipids in the choline head group moiety andthe carboxy-terminal fatty-acid carbon atom, respectively.Using synthetic 'IC carboxyl-labeled lecithin, we were

able to assign the carboxyl resonances from the two non-equivalent lecithin fatty-acid chains in organic solvents andD20. We further demonstrated that the relaxation times ofthe two are similar in all circumstances. The relatively longrelaxation times in the most mobile circumstances in meth-anolic solution are substantially diminished in the micellesdispersed in chloroform and further diminished in the D20bilayer system. Sonicated dispersions of PC have been re-

peatedly used for studying bilayer systems. T,'s in this reportare qualitatively consistent with a previous finding, thatpacking of the molecules is highest in the region of the glycerolresidues and the first two chain methylenes near them (45,46). An intriguing result was the failure to obtain resolutionof the carboxyl resonance doublet of lecithin when reas-sembled with apoHDL or incorporated into native HDL.Selective labeling of both a and # fatty-acid chain carboxylcarbon will be necessary to explain which one of the resonanceseither shifts or broadens in the presence of HDL apoproteins.The a-fatty-acid carboxyl ",C enrichment is not high

enough to account fof the intensity of the "IC resonance re-maining in the presence of protein, and we assume that thisresonance is due to the f3-fatty-acid carboxyl 13C enrichment.There was no significant alteration of the relaxation timeswhen PC D20 vesicles, the apoHDL-PC complex and nativeHDL were compared. Since, in the presence of protein, therelaxation times were almost unchanged, one might suggestthat the # carboxyl group of PC is not involved in protein-lipid interaction. This suggestion is supported by the interest-ing finding that after complete hydrolysis of the PC in HDLby a phospholipase A2 the hydrodynamic behavior of HDLand its morphological characteristics are not altered, whereastreatment with phospholipase C and D lead to marked struc-tural alterations (47).Even more striking was the composition of chemical shift

data and relaxation times obtained with cholesteryl-[1-"3C]-oleate in a CDC13 solution of pure synthetic material and innative HDL, in which the carboxyl 13C carbon of cholesterylesters was enriched by LCAT reaction or by recombinationtechnique. In the presence of the protein, the cholesteryl esternot only had a relaxation time substantially shorter than it didin chloroform solution, but the carboxyl resonance was shiftedupfield by 2 ppm. The marked distribution in mobility of thisatom probably was the result of a specific interaction,for example, electron transfer from protein nitrogen to thepositive center of the carboxyl group. Since cholesteryl estersin native HDL are largely derived from an intramolecularprocess through the activity of the enzyme LCAT, similarexperiments using fatty acids "IC labeled in different positionsshould allow more precise determination of the location anddynamics of the cholesteryl ester moiety in native lipoproteins.Our results confirm and extend the previous observation

that the 13C spin-lattice relaxation times of lecithin andsphingomyelin enriched with 13C in the choline head groupmoiety are characteristic of both the chemical structure andthe steric interaction between the molecules in differentsolvents. The T, values of the [methyl-"C]choline resonancein egg yolk PC and SPM are consistently larger in methanolthan T, values for the corresponding nuclei in chloroform. InCDC1L, PC exists as inverted spherical micelles composed of60-70 molecules, whereas it exists in methanol as a trimer(48, 49). Thus, the relatively short N+Me3 T, values in CDC1Lreflect probably the tight packing of the head groups in theinverted micellar structure. The N+Me3 T, values of eggyolk PC as well as SPM in D20, were almost identical to thoseobtained with the reassembled lipoproteins. This findingsuggests that the polar head group of both molecules in D20bilayers have the same aqueous environment as the polarhead groups in reassembled lipoproteins and indicate that theprotein moiety is not restricting the molecular freedom of thecholine head groups. Based on proton NMR data, it has beenpreviously suggested that the polar choline head groups inHDL are unrestrained in motion (50). These studies provided

Proc. Nat. Acad. Sci. USA 71 (1974)

Dow

nloa

ded

by g

uest

on

Oct

ober

16,

202

0

c'C NMR Spectroscopy of Lipoprotein 3705

quantitative data regarding the mobility of these moieties. It.should be noted that our studies might have failed to detecta small percentage of choline carbons that were constrainedin their motion, due to the experimental error (4t10%) in-volved in measurement of relaxation time. Furthermore thisdiscussion assumes short correlation time3 (cooro << 1) for thesystems studied (6). The dependence of relaxation times on

temperature and viscosity of solutions must still be investi-gated.

It has been previously pointed out that apoC-III and apoA-II contain several pairs of sequentially adjacent amino acidswhose side chains are oppositely charged (38). The ionicinteraction of these oppositely charged side chains with thezwitterionic polar head group has been suggested as a mech-anism of protein-lipid binding (51). The present data suggestthat this mechanism is of minor importance for apoA-I andapoA-II. We suggest that the binding of PC and SPM isprimarily hydrophobic in character. This concept is, in part,supported by recent findings that the extension of the apoC-III peptide 48-79 by seven additional residues to apoC-III(41-79) was accompanied by a significant increase in PCbinding, though the additional residues did not contain a

zwitterionic pair of amino acids (52).Firm conclusions regarding the specificity and nature of

interaction and dynamics of individual lipid and apolipo-

protein constituents cannot be made at the present time.Since the lipid constituents of lipoproteins can be 13C labeledat almost any carbon atom in the fatty-acid chains, theglycerol and polar head groups moiety and several carbonatoms of the cholesterol molecule are accessible to chemicalsynthesis, quantification of lipid-lipid and lipid-proteininteraction is in the future expected from 'IC NMR analysis.Additional physical information of the interaction of phos-pholipids with apolipoproteins can be obtained by phos-phorus-31-NMR spectroscopy, as detailed in the accompany-ing paper (53).

We are grateful to Prof. Dr. W. Stoffel, Institut fur Physio-logische Chemie der Universitat Koln, for experimental and con-ceptual advice in the present studies. (See accompanying paper.)We further wish to thank Mrs. Anne Flory, Miss RosemaryRonan, and Mrs. Ann LaRue for excellent technical support.

1. Leslie, R. B., Chapman, D. & Scanu, A. M. (1969) Chem.Phys. Lipids 3, 152-158.

2. Steim, J. M., Edner, 0. J. & Bargoot, F. G. (1968) Science162, 909-911.

3. Chapman, D., Leslie, R. B., Hirz, R. & Scanu, A. M. (1969)Biochim. Biophys. Acta 176, 524-536.

4. Chapman, D., Leslie, R. B., Hirz, R. & Scanu, A. M. (1969)Nature 221, 260-261.

5. Levy, G. C. (1973) Accounts Chem. Res. 6, 161-169.6. Farrar, T. C. & Becker, E. D. (1971) in Pulse and Fourier

TransformNMR (Academic Press, New York), p. 20.7. de Lalla, 0. F. & Gofman, J. W. (1959) Methods Biochem.

Anal. 1, 459-478.8. Levy, R. I. & Fredrickson, D. S. (1965) J. Clin. Invest. 44,

426-441.9. Lux, S. E., John, K. M. & Brewer, Jr., H. B. (1972) J. Biol.

Chem. 247, 7510-7518.10. Scanu, A., Toth, J., Edelstein, C., Koga, S. & Stiller, E.

(1969) Biochemistry 8, 3309-3316.11. Shore, B. & Shore, V. (1969) Biochemistry 8, 4510-4516.12. Reisfield, R. A. & Small; P. A., Jr. (1966) Science 152, 1253-

1255.13. Chrambach, A. & Rodbard, D. (1971) Science 172, 440-451.14. Edman, P. (1970) in Protein Sequence Determination, ed.

Needleman, S. B. (Springer-Verlag, New York), pp. 211-255.

15. Pisano, J. J. & Bronzert, T. J. (1969) J. Biol. Chem. 244,5597-5607.

16. Pisano, J. J., Bronzert, T. J. & Brewer, H. B., Jr. (1972)Anal. Biochem. 45,43-59.

17. Edelstein, C., Lim, C. T. & Scanu, A. M. (1972) J. Biol.Chem. 247, 5842-5849.

18. Scanu, A. M., Lim, C. T. & Edelstein, C. (1972) J. Biol.Chem. 247, 5850-5855.

19. Lux, S. E., John, K. M., Ronan, R. & Brewer, H. B., Jr.(1972) J. Biol. Chem. 247, 7519-7527.

20. Dawson, R. M. C. (1958) Biochem. J. 70, 559-570.21. Dittmer, J. C. & Lester, R. L. (1964) J. Lipid Res. 5, 126-

127.22. Stoffel, W., LeKim, D. & Tschung, T. S. (1971) Hoppe-Sey-

ler's Z. Physiol. Chem. 352, 1058-1064.23. Stoffel, W., Zierenberg, 0. & Tunggal, B. D. (1972) Hoppe-

Seyler's Z. Physiol. Chem. 353, 1962-1969.24. Wood, T. R., Jackson, F. L., Baldwin, A. R. & Longenecker,

H. E. (1944) J. Amer. Chem. Soc. 66, 287-289.25. Rouser, G., Kritchevsky, G., Heller, D. & Lieber, E. (1963)

J. Amer. Oil Chem. Soc. 40, 425-454.26. Hirz, R. & Scanu, A. M. (1970) Biochim. Biophys. Acta 207,

364-367.27. Scanu, A. M. & Tardieu, A. (1971) Biochim. Biophys. Acta

231, 170-174.28. Assmann, G. & Brewer, H. B., Jr. (1974) Proc. Nat. Acad.

Sci. USA 71,989-993.29. Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall,

R. J. (1951)J. Biol. Chem. 193,265-275.30. Glomset, J. A. (1968)J. Lipid Res. 9, 155-167.31. Lees, R. S. & Hatch, F. T. (1963)J. Lab. Clin. Med. 61, 518-

528.32. Allerhand, A., Doddrell, D., Glushko, V., Cochran, D. W.,

Wenkert, E., Lawson, P. J. & Gurd, F. R. N. (1971) J.Amer. Chem. Soc. 93, 544-546.

33. Hamilton, J. A., Talkowski, C., Williams, E., Avila, E. M.,Allerhand, A., Cordes, E. H. & Camejo, G. (1973) Science180, 193-195.

34. Scanu, A. M. (1966) J. Lipid Res. 7,295-306.35. Rudman, D., Garcia, L. A. & Howaild, C. H. (1970) J. Clin.

Invest. 49, 365-372.36. Kostner, G. & Alaupovic, P. (1971) FEBS Lett. 15, 320-324.37. Alaupovic, P. (1971) Atherosclerosis 13, 141-146.38. Brewer, H. B., Jr., Lux, S. E., Ronan, R. & John, K. M.

(1972) Proc. Nat. Acad. Sci. USA 69, 1304-1308.39. Shulman, R., Herbert, P., Wehrly, K., Chesebro, B., Levy,

R. I. & Fredrickson, D. S. (1972) Circulation XLVI: II(Abstract) 246.

40. Brewer, H. B., Jr., Shulman, R., Herbert, P., Ronan, R. &Wehrly, K. (1972) Advan. Exp. Med. Biol. 25, 280-282.

41. Levine, Y. K., Birdsall, N. J. M., Lee, A. G. & Metcalfe,J. C. (1972) Biochemistry 11, 1416-1421.

42. Metcalfe, J. C., Birdsall, N. J. M., Feeney, J., Lee, A. G.,Levine, Y. K. & Partington, P. (1971) Nature 233, 199-201.

43. Birdsall, N. J. M., Lee, A. G., Levine, Y. K. & Metcalf,J. C. (1971) Biochim. Biophys. Acta 241, 693-696.

44. Hubbell, W. L. & McConnell, H. M. (1971) J. Amer. Chem.Soc. 93, 314-326.

45. Finer, E. G., Flook, A. G. & Hauser, H. (1972) Biochim.Biophys. Acta 260, 49-58.

46. Lee, A. G., Birdsall, N. J. M., Levine, Y. K. & Metcalfe, J. C.(1972) Biochim. Biophys. Acta 255, 43-56.

47. Scanu, A. M. (1972) Ann. N.Y. Acad. Sci. 195, 390-406.48. Price, H. I. & Lewis, W. C. M. (1929) Biochem. J. 23, 1030-

1035.49. Elworthy, P. H. & Macintosh, D. S. (1961) J. Pharmacol.

13, 633-640.50. Scanu, A. (1972) Biochim. Biophys. Acta 265, 471-508.51. Brewer, H. B., Jr., Lux, S. E., Scanu, A. M., Ronan, R. &

John, K. M. (1972) in Chemistry and Biology of Peptides, ed.Meienhofer, J. (Academic Press, New York), pp. 705-712.

52. Sparrow, J. T., Gotto, A. M. & Morrisett, J. D. (1973)Proc. Nat. Acad. Sci. USA 70, 2124-2128.

53. Assmann, G., Sokoloski, E. A. & Brewer, H. B., Jr. (1974)Proc. Nat. Acad. Sci. USA 71, 549-553.

Proc. Nat. Acad. Sci. USA 71 (1974)

Dow

nloa

ded

by g

uest

on

Oct

ober

16,

202

0