Background In our laboratory we measure the activity and specificity of phospholipid transfer proteins by measurements of lipid transport from donor membranes to acceptor membranes (Figure 1). Once it became clear that one of the lipids we routinely used in the preparation of donor vesicles was no longer commercially available (Nhexadecanoyl- lactosylceramide), we became interested in evaluating other glycolipids for their use in an agglutinin-dependent precipitation of unilamellar vesicles.

Figure 1. General scheme of proteincatalyzed lipid transport. The inclusion of a glycolipid (hexagons, green) in the donor membrane permits their selective agglutination.

Reagents
Unilamellar vesicles are prepared by rapid injection into warm buffer of a solution of lipids in
ethanol / dimethyl sulfoxide. The resulting concentration is 2.5 mM, based on total lipid present.
Vesicles may be briefly subjected to bath sonication if some turbidity is observed. Finally the vesicles
are filtered through a 1-mm glass fiber filter (Acrodisc) and 0.45-mm Teflon filter (Millex).

Donor vesicles contain egg phosphatidylcholine (PtdCho) (88 mol%), egg phosphatidate
(PtdOH) (2 mol%), and the indicated glycolipid (10 mol%); a trace (<0.5 mol%) of ³H-labeled PtdCho is
added to monitor precipitation. Acceptor vesicles contain egg PtdCho (98 mol%) and egg PtdOH (2
mol%); a trace (<0.5 mol%) of ¹⁴C-labeled cholesteryl oleate is added to monitor recovery. Lipids were
purchased or obtained from Avanti Polar Lipids, Sigma-Aldrich, Amersham-Pharmacia, and Perkin
Elmer.

The standard assay buffer is 10 mM HEPES-Na, 1 mM EDTA-Na, and 50 mM NaCl (pH 7.4). A
2 mg@ml^-1 solution of bovine plasma albumin (fatty-acid free, Sigma A-6003) is prepared in assay buffer;
this solution may be stored at -20°. Pure Ricinus communis agglutinin (RCA, 120 kDa) was isolated
from castor beans (Cawley et al., 1978), stored at -75°C, and diluted in assay buffer.

Protocol
Incubations are carried out in 1.5-ml disposable, conical centrifuge tubes. To these tubes,
placed in an ice bath, are added, in the following order, buffer (220 ml), albumin (20 ml), acceptor
vesicles (120 ml, 300 nmol), and, lastly, donor vesicles (40 ml, 100 nmol). The total volume is 400 ml.
After capping and gentle vortexing, the tubes are placed in water bath maintained at 37°. All
incubations are performed in duplicate or triplicate.

After 30 min in the water bath, the tubes are returned to the ice bath. Agglutinin (150 ml, variable protein) is added, followed by gentle vortexing to mix. After standing for 30 min, the tubes are allowed to warm to room temperature for 30 min. Next, the tubed are centrifuged at 15,000×g for 5-6 min. At this point a small white pellet of precipitated donor vesicles should be observed. Finally, an aliquot of the supernate (500 ml) is transferred very carefully to a 7-ml scintillation vial containing 5 ml water-miscible cocktail. The different radiolabeled lipids in the donor and acceptor membranes permit the calculation of donor precipitation and acceptor recovery, both expressed as a percent of the initial vesicle radioactivity (Kasper and Helmkamp, 1981).

Results and Discussion
As shown in Figure 2, the inclusion of either N-lactosyl-phosphatidylethanolamine (LacPtdEtn)
or N-hexadecanoyl-lactosylceramide (C₁₆-LacCer) as the glycolipid in the donor vesicle permitted
excellent separation of these membranes from the acceptor vesicles. The half-maximal amount of RCA
required for donor agglutination was 8 mg for LacPtdEtn and 16 mg for C₁₆-LacCer. In contrast, when
N-dodecanoyl-galactosylceramide (C₁₂-GalCer) was incorporated into the donor vesicles, it failed to
promote their subsequent agglutination. RCA preferentially binds to the D-galactosyl-(b1®4)-D-glucosyl
moiety on lipids and proteins (Nicolson and Blaustein, 1972). Furthermore, both lactose and Dgalactose
effectively inhibit the agglutination of human erythrocytes (Nicolson and Blaustein, 1972;
Olsnes et al., 1974). It seems likely that the D-galactosyl moiety must be localized sufficiently away
from the membrane surface for optimal interaction with carbohydrate binding site of RCA.

What was unexpected in our investigation was the behavior of N-octanoyl-lactosylceramide (C₈-
LacCer) to support agglutination (Figure 2). While the half-maximal amount to precipitate donor
vesicles increased to 26 :g, there was also a sharp decrease in acceptor vesicle recovery. We
suggest that a spontaneous flux of C8-LacCer from donor vesicles to acceptor vesicles occurs during
the course of incubation at 37°. Following this redistribution of C₈-LacCer, the subsequent addition of
RCA led to the agglutination and precipitation of both populations of vesicles. That this is a reasonable
conclusion is supported by rates of spontaneous transfer of glycolipids between PtdCho membranes
(Jones et al., 1990). Using their results for a series of semisynthetic, long-chain GlcCer molecules
(Figure 3), we generate an empirical expression for the dependence of spontaneous transfer on acyl
chain length. For a lipid such as C₈-LacCer, the half-time for transfer would be 2-3 min at 45°. Under
our assay conditions of 30 min at 37°, we would anticipate considerable redistribution of this shortchain
glycolipid from donor vesicles to acceptor vesicles.

References

1. Cawley, D.B., Hedblom, M.L., and Houston, L.L. (1978) Arch. Biochem. Biophys. 190, 744-755
   
2. Jones, J.D., Almeida, P.F., and Thompson, T.E. (1990) Biochemistry 29, 3892-3897
   
3. Kasper, A.M., and Helmkamp, G.M., Jr. (1981) Biochemistry 20, 146-151
   
4. Nicolson, G.L., and Blaustein J. (1972) Biochim. Biophys. Acta 266, 543-547
   
5. Olsnes, S., Saltvedt, E., and Pihl, A. (1974) J. Biol. Chem. 249, 803-810
   

Figure 2. Separation of donor and acceptor vesicles by selective agglutination.

Figure 3. Transfer of GlcCer between PtdCho vesicles.