THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 261, No. 33, Issue of November 25, pp. 15767-15771,1986 0 1986 by The American Society of Biological Chemists, Inc. Printed in U. S. A.

Purification and Characterization of a Novel Glycan- Phosphatidylinositol-specific Phospholipase C from Trypanosoma brucei”

(Received for publication, May 28, 1986)

Judith A. Fox$, Michael Duszenkog, Michael A. J. Fergusonq, Martin G. Low)), and George A. M. Cross From the Rockefeller University. New York, New York 10021 and the 11 Oklahoma Medical Research Foundation, Oklahoma City, Oklahoma 73164

A novel membrane-bound glycan-phosphatidylino- sitol-specific phospholipase C, which catalyzes the con- version of membrane form variant surface glycopro- teins to soluble variant surface glycoproteins, with the release of sn- 1,Z-dimyristylglycerol, has been isolated from Trypanosoma brucei. The activity was solubi- lized from trypanosome membrane fractions in non- ionic detergent and purified by anion exchange chro- matography on DEAE-cellulose followed by chroma- tography on phosphatidylinositol-Sepharose. The en- zyme constitutes about 0.1% of the total cellular pro- tein and has an apparent molecular weight of 39,800. The enzyme shows a head group specificity for mole- cules containing carbohydrate covalently linked to gly- can-phosphatidylinositol, but can also act on the mon- oacyl derivative of membrane form variant surface glycoprotein. It shows no specific ion requirements but is stimulated by thiol-reducing agents and inhibited by ions that thiols chelate.

Trypanosoma brucei, the causative agent of African sleeping sickness, evades the immune response of its mammalian host by the sequential expression of immunologically distinct var- iant surface glycoproteins (VSGs’) (reviewed in Boothroyd, 1985), which form a dense coat on the exterior surface of the cell membrane. In addition to the removal of an N-terminal signal sequence, VSGs undergo an unusual post-translational modification in which a short, extremely hydrophobic car- boxyl-terminal amino acid segment is cleaved and replaced with a glycolipid moiety (Boothroyd et al., 1980; Boothroyd et al., 1982; Ferguson and Cross, 1984; Ferguson et al., 1985a, b), which serves as the membrane anchor. The glycolipid moiety

* This work was supported in part by National Institutes of Health Grant AI 21531. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ Recipient of National Research Science Award AI 07185. Present address: Physiologisch chemisches Institut der Univer-

sitat, Tubingen, D-7400, FRG. 1 Present address: Oxford University, S. Parks Rd., Oxford OX1

3 QU, England. The abbreviations used are: VSG, variant surface glycoprotein;

mfVSG, membrane form VSG; GPI-PLC, glycan-phosphatidylinosi- tol-specific phospholipase C; PI, phosphatidylinositol; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; TLCK, N“-p-tosyl-L-lysine chloromethyl ketone; NOG, n-octylglucopyrano- side; EGTA, [ethylenebis(oxyethylenenitrilo)]tetraacetic acid; DTT, dithiothreitol; CAPS, 3-[cyclohexyl-amino]-l-propanesulfonic acid MES, 2-( N-morpho1ino)ethanesulfonic acid.

of membrane form VSG (mfVSG) contains dimyristylphos- phatidylinositol glycosidically linked to glucosamine, and it is coupled to the polypeptide via additional carbohydrate, phos- phate, and ethanolamine, which forms an amide bond with the carboxyl-terminal amino acid residue (Ferguson et al., 1985a, b; Holder, 1983).

Osmotic shock or mechanical disruption of the trypanosome releases soluble VSG by the action of an endogenous enzyme (Cardoso de Almeida and Turner, 1983). This enzyme has been previously characterized as a phospholipase C that cleaves a phosphodiester bond of mfVSG, releasing soluble VSG and dimyristylglycerol (Ferguson et al., 1985a). We report here the purification and characterization of this unu- sual, calcium-independent, glycan-phosphatidylinositol-spe- cific phospholipase C (GPI-PC). Although the physiological role of the GPI-PLC has not been elucidated, we show that this enzyme exhibits substantial specificity for molecules con- taining glycosylated phosphatidylinositol (PI).

EXPERIMENTAL PROCEDURES

Materials-Epoxy-activated-Sepharose 4B, concanavalin A-Seph- arose, Triton X-114, n-octylglucopyranoside, Dalton VII-L SDS- PAGE molecular weight standards, neutral lipid, and phospholipid standards were purchased from Sigma. Extractigel D was purchased from Pierce, [9,10-3H]myristate and [(~holine-methyl)-~HJphospha- tidylcholine from New England Nuclear, Whatman DE52, Silica G, and Silica 60 tlc plates from Fisher. [3H]PI and [32P]PI-4-phosphate and 4,5-bisphosphate were prepared as described previously (Low et al., 1986a). Bee venom phospholipase A, was purchased from Boeh- ringer Mannheim. Silver-staining kit and Bio-Rad Protein Assay were from Bio-Rad.

Growth of T. brucei-Trypanosomes of the Molten0 Trypanozoon antigen type (MITat) 1.4 (otherwise known as variant clone 117) of T. brucei strain 427 were purified from infected rat blood as previously described (Cross, 1975).

Preparation of [9,10-3HlMyristate-labeled mfVSG-lO’O purified trypanosomes were biosynthetically labeled by incubation with [3H] myristate for 2 b (Ferguson and Cross, 1984). mfVSG was purified by a slight modification of the procedure of Clarke et al. (1985). Trypanosomes were lysed and extracted with 0.1% trifluoroacetic acid as described (Clarke et al., 1985). The pooled trifluoroacetic acid extracts were passed through a 0.45-pm Millex filter (Millipore) and injected in 1 ml aliquots onto a 5-pm reversed-phase C-18 preparative Ultrapore column (Altex), which had been equilibrated in 0.1% tri- fluoroacetic acid using Varian 5500 high pressure liquid chromatog- raphy. The protein elution program was as follows: a linear gradient from 0-30% isopropyl alcohol in 15 min, hold for 15 min followed by a linear gradient to 40% isopropyl alcohol in 80 min. mNSG eluted reproducibly at 54 min. mfVSG-containing fractions were neutralized with ammonium bicarbonate.

Preparation of Pronase Fragment and Dimyristyl-PI from mfVSG- The pronase fragment (aspartyl-ethanolamine-glycolipid moiety) of mfVSG and dimyristylphosphatidylinositol were prepared according to published procedures (Ferguson et al., 19851a, b) by pronase

15767

15768 T. brucei Glycan-phosphatidylinositol Phospholipase C

digestion or nitrous acid deamination of intact mfVSG. Product purity was judged by tlc (Ferguson et al., 1985a, b).

Preparation of PI-Sepharose-25 mg of mixed phosphoinositides (Sigma) were dissolved in 5 ml of chloroform and dried down under nitrogen. The phospholipid was then sonicated into vesicles in 20 ml of 0.1 M Na,HCO, pH 9.5. 3 g of epoxy-activated sepharose were rehydrated and washed with 300 ml of deionized water. The resin was added to the phospholipid solution and mixed end-over-end for 20 h at room temperature. The remaining unreacted epoxy groups were blocked by the addition of 2 ml of ethanolamine. The PI-coupled Sepharose was washed free of coupling and blocking buffers and stored in deionized water with 0.5% (w/v) azide.

Assay of GPI-PLC Actiuity-GPI-PLC was incubated with [3H] myristate-labeled mfVSG in 25 mM Tris-HC1, pH 8.0, with 0.5% Nonidet P-40 for 10 min. The reaction was terminated by the addition of 10 pl of acetic acid. [3H]myri~tyldiacylglycerol was extracted with 1 ml of toluene. Organic and aqueous phases were separated by centrifugation and 0.5 ml of the organic phase was removed for liquid scintillation counting.

Protein Assay, SDS-PAGE, and Isoelectric Focus&-Protein de- terminations were performed using the Bio-Rad protein assay with IgG as standard. SDS-PAGE was performed in Laemmli buffers (Laemmli, 1970) using a 15% acrylamide running gel with a 2001 acrylamide/bis ratio. Isoelectric focusing was performed at 10 "C on an LKB Multiphor apparatus for 1.5 h. A 1-mm thick isoelectric- focusing gel was cast on Gelbond (LKB) using an Ultrathin gel mold (LKB) with a pH range from 8 to 9.5. The electrophoresis chamber contained 1 liter of NaOH, and the system was placed under a N, atmosphere to protect the pH gradient from atmospheric CO,. The pH gradient was determined by placing 5-mm slices of the gel in N2- sparged 20 mM KC1 and measuring the pH. GPI-PLC was localized by activity assay of 5-mm slices as described above.

Detergent Solubilization of GPI-PLC-Trypanosome ghosts were prepared by osmotic lysis of 2.2 X 10'' trypanosomes in 22 ml of 10 mM sodium phosphate buffer, pH 8.0, containing 0.1 mM TLCK and 0.1 mM phenylmethylsulfonyl fluoride at 37 "C. Following a 5-min incubation at 37 "C, the preparation was cooled to 0 "C and centri- fuged at 50,000 X g for 1 h. The membrane pellet was resuspended in the same buffer, sonicated on ice for 60 s at 80% power, and centri- fuged at 100,000 X g for 1 h. The resulting pellet was homogenized on ice in a loose-fitting dounce homogenizer in 1% NOG, 25 mM Tris-HC1, pH 8.0, with 0.1 mM phenymethylsulfonyl fluoride, 0.1 mM TLCK, 48 pg/ml aprotinin, 20 pg/ml leupeptin. The homogenate was then centrifuged at 100,000 X g for 1 h.

RESULTS

Purification and Properties of Glycan-Phosphatidylinositol- specific Phospholipase C-The purification of GPI-PLC from the membrane fraction of 2.2 X 10" trypanosomes is sum- marized in Table I. The nonionic detergent NOG was most effective in solubilizing the enzyme activity, which was inhib- ited both by anionic and cationic detergents (data not shown). NOG was found to be particularly useful since it does not absorb a t 280 nm, facilitating UV monitoring of column

TABLE I Purification of GPI-PLC

with 25 nM [3H]mfVSG for 10 min and analyzed as described under Samples from each step of the purification were incubated in 1 ml

"Experimental Procedures." Specific activity is expressed as pmol of mfVSG consumed/min/mg GPI-PLC. These assays were performed well below substrate saturation and are not Vms. values.

Step

Whole cell Detergent-solu-

bilized mem- branes

Supernatant after centrif- ugation

DE52 PI-Sepharose

Total protein

mg ~-

154 43

1 4

1.62 0.162

Specific activitv Purification Yield

pmollminlmg %

25.4 3.6 (IOO]

80.3 12.2 100

1000 148 150 2450 342 35

elution, is easily removed by Extractigel D, and its concentra- tion can be determined by carbohydrate assay. The solubilized GPI-PLC was passed through a DE52 anion exchange column, as shown in Fig. 1. The enzyme was slightly retarded on this column, eluting with the equilibration buffer. Subsequent washing of the column with 1 M NaCl eluted other proteins but no additional mNSG-hydrolyzing activity. Typically, we recovered 120-150% of the applied activity from the DE52 column. The increase in total activity units could be explained by the removal of an endogenous, membrane-associated in- hibitor or one introduced during sonication (such as metal ions). Although GPI-PLC could be solubilized without soni- cation, extraction was more efficient if the trypansome ghosts were completely disrupted. Fractions with GPI-PLC activity were pooled, and detergent was removed on an Extractigel D column of a volume equal to that of the pooled fractions. The enzyme was then absorbed onto a 5-ml PI-Sepharose column, as shown in Fig. 2. GPI-PLC did not bind to epoxy-activated- Sepharose which had been inactivated with water (data not

Fraction number

FIG. 1. GPI-PLC purification by DE52 chromatography. The detergent-solubilized GPI-PLC activity (10 ml) was chromato- graphed at 30 ml/h on a 20 ml DE52 column equilibrated in homog- enizing buffer at room temperature, collecting 2-ml fractions. After all the activity had eluted, the column was washed with 1 M NaCl to remove additional protein. Azm ",,, U. Activity (dpm released from a IO-min incubation with 23,000 dpm [3H]mfVSG) 0- - -0.

t 0 1

J 1000

- "" *"" " "-0" " 0 - I 4 1

I O 20 30 40 50 60 Froctlon number

FIG. 2. GPI-PLC chromatography on PI-Sepharose. Deter- gent-free GPI-PLC, obtained by Extractigel treatment of DE52- purified enzyme, was adsorbed onto a 5-ml PI-Sepharose column at 4 "C. The column was washed sequentially with 20 ml of 25 mM Tris- HCI, pH 8.0, 10 ml of 0.5% NOG, 0.5 mg/ml PI in the same buffer, and with 10 ml of 0.5% NOG in the initial buffer, collecting 4-ml fractions. The column was brought to room temperature, the flow reversed, fraction size reduced to 1 ml, and the enzyme eluted with 1% NOG, 25 mM Tris-HC1, pH 8.0. Protein concentration U. Activity (dpm released from a 10-min incubation with 23,000 dpm [3H]mfVSG) 0- - -0.

T. brucei Glycan-phosphatidylinositol Phospholipase C 15769

shown), suggesting that the adsorption of the activity to the PI-modified resin was due to specific interaction, due either to hydrophobicity or active site affinity, with PI. GPI-PLC was purified 95-fold from the membrane fraction of osmoti- cally lysed trypanosomes. The enzyme purification was effec- tively 340-fold over total trypanosome protein with a 35% yield of activity (Table I). GPI-PLC is extremely stable, maintaining full activity at 4 “C for 1-2 months.

SDS-PAGE of GPI-PLC revealed a major band when vis- ualized either by Coomassie Blue or by silver staining with an apparent molecular weight of 39,800, by reference to mo- lecular weight standards (Fig. 3) and a PI of pH 8.6-8.8 (data not shown). Following purification through PI-Sepharose, the enzyme preparation had contaminants in the molecular weight range of bovine serum albumin (66,000) or VSG (54,000-65,000). Higher specific activity GPI-PLC correlated with a decrease in contamination by the higher molecular weight impurities (data not shown). Neither anti-VSG anti- body or concanavalin A affinity columns removed the impur- ities, suggesting they were not VSG. Correspondingly, the failure of the enzyme to bind to concanavalin A suggests that it does not contain glycan with exposed mannose as do most trypanosome glycoproteins. When the preparation was sub- jected to phase separation with the detergent Triton X-114, these contaminants were found in the aqueous phase while the enzyme activity partitioned with the hydrophobic phase.

GPI-PLC is a n Integral Membrane Protein-Phase sepa- ration with Triton X-114 separates molecules on the basis of hydrophobicity. This detergent condenses above 20 “C to form detergent-enriched and aqueous, relatively detergent-free phases, which can be separated by centrifugation. Integral membrane proteins interact with the detergent and settle in

66.000

45,000

36,000

29,000 24,000

20300

14,200

39,800

the lower phase, while soluble proteins associate with the upper, aqueous phase (Bordier, 1981). GPI-PLC was subjected to phase separation with Triton X-114. Greater than 90% of the enzyme activity associated with the detergent phase (data not shown). This result was expected on the basis of the detergent requirement for solubilization of activity and the apparently irreversible hydrophobic interaction of the major- ity of GPI-PLC activity with several chromatography resins, including octyl and phenyl-Sepharose, AcA 44, Bio-Gel P-60, and Affi-Gel501.

Effects of Ions and Thiol-reducing Agents on GPI-PLC Activity-The activity of GPI-PLC exhibited small but sig- nificant inhibition in the presence of calcium ions (Table 11). These data suggest that the enzyme is calcium-independent. GPI-PLC activity showed marked stimulation in the presence of the thiol-reducing agent, DTT, and acute sensitivity to mercury and zinc ions (Table 11), perhaps indicating the presence of an active site thiol.

p H Dependence of GPI-PLC Actiuity-Evaluation of the pH dependence of GPI-PLC activity showed two maxima, one a t pH 7.0 and the other at pH 10.0 (Fig. 4). I t is possible that the maximum seen at pH 7.0 reflects the titration of a specific amino acid residue, such as histidine, or a conforma- tional change in either the enzyme or substrate. The observed increase in activity through pH 10.0 could reflect rate accel- eration by the increased concentration of hydroxide ion nu- cleophiles or titration of an amino acid of the enzyme that acts as a base in intramolecular catalysis. Ferguson et al. (1985b) demonstrated that, at pH 8.0, a major product of mfVSG hydrolysis was the cyclic inositol-phosphate rather than the free phosphate. Whether the enzyme is restricted to intramolecular catalysis by an inositol hydroxyl a t all pH values requires further investigation.

Substrate Specificity of GPI-PLC-The activity of the en- zyme toward a variety of substrates was determined (Table 111). GPI-PLC rapidly hydrolyzed mfVSG and the pronase fragment of mfVSG, which contains only the terminal amino acid and the glycolipid anchor. In addition, the enzyme cata- lyzed the hydrolysis of the monoacyl derivative of mfVSG to release monomyristylglycerol (Fig. 5 ) , indicating that it can act as a lyso-phospholipase C. In contrast to the activity against these glycan-PI derivatives, there was little activity against PI itself at 1,000-fold the concentration where rapid hydrolysis of mfVSG is observed. This was observed with both [3H]dimyristyl-PI prepared from [‘HJmfVSG by nitrous acid deamination and also with [”HIPI prepared from rat liver microsomes (Table 111). Similarly, [32P]phosphatidylinositol- 4-phosphate, [“P]phosphatidylinositol 4,5-bisphosphate or

TABLE I1 Effects of cation and sulfhydryl reagents on GPI-PLC

GPI-PLC was incubated at room temperature with the reagents indicated. [3H]mfVSG was added to each incubation to initiate the reaction. The samples were incubat,ed at 37 “C for 10 min and ana- lyzed as described under “Experimental Procedures.” All assays were performed in triplicate.

Addition Activitv % of control

None 1 mM CaClz 10 mM EGTA

None 1 2 3 4 5 10 mM dithiothreitol

FIG. 3. SDS-PAGE of GPI-PLC purification. SDS-PAGE was performed using a 15% polyacrylamide gel and silver-stained. 1, molecular weight standards; 2, detergent homogenate before centrif- ugation, 30 pg; 3, solubilized GPI-PLC, 30 pg; 4, DE52-purified GPI- PLC, 20 pg; 5, PI-Sepharose purified GPI-PLC, 15 pg.

10 p M HgClz 1 pM HgC12 100 p M ZnClZ 10 p M ZnClz

dpm f 10 min 9,200 6,400

13,000 22,000 4,200 0 2,800 0 2,500

100 70

140 240 100

0 67 0

59

15770 T. brucei Glycan-phosphatidylinositol Phospholipase C

I I

I ' O ~ ~ 9000

i '

G 4 5 6 7 8 9 1 0 i l

PH FIG. 4. pH dependence of GPI-PLC activity. The pH profile

of enzyme activity was determined in a mixed buffer system composed of 50 mM CAPS/MES/TRIS, using GPI-PLC purified 100-fold from cells through DE52 followed by, on this occasion, Affi-Gel501 rather than PI-Sepharose. The pH values were measured at 37 "C. All assays were performed as described under "Experimental Procedures" in triplicate with a control containing no enzyme at each pH.

TABLE 111 Substrute specificity of GPI-PLC

GPI-PLC activities with mfVSG and derivatives were assayed as described under "Experimental Procedures." The PI assays were performed as described in Low et at., 1986a. All assays were performed in triplicate. __ .~

Substrate Incubation time % cleavage

mfVSG" 10 min

Pronase fragment" 10 min

Dimyristyl-PIn 10 min 30 min 16 hb 16 h

Pi' 60 min

PId 60 min 16 h

64

77

0 0 0 0

0

0 1.0

These three substrates were all of the same specific activity and at a concentration of 8-10 nM. "Pronase fragment" is the aspartyl- ethanolamine-glycolipid moiety of mfVSG.

Concentration of dimyristyl-PI was 20 nM. Concentration of PI was 20 p ~ . Concentration of PI was 100 WM.

["H]phosphatidylcholine were not significantly hydrolyzed by the GPI-PLC (data not shown). This unique headgroup spec- ificity suggests that the enzyme cannot be classified as either a phosphatidylcholine- or PI-specific phospholipase C. The exact structural features of the substrate molecule required for recognition by the GPI-PLC bears further study.

DISCUSSION

In this communication we report the purification a glycan- phosphatidylinositol-specific phospholipase C that releases soluble VSG from the bloodstream form of T. brucei. Purifi-

5000

3000

IO00

I 2 3 4 5 6 7 8 9 1 0 1 1 1 2 cm

FIG. 5. GPI-PLC hydrolyzes monoacyl-mNSG. [3H]mfVSG was incubated overnight a t 37 in 0.1% (w/v) sodium deoxycholate, 1 mM CaC12, 100 mM Tris, pH 7.5, with 4 pl of bee venom phospho- lipase A2. Two 0.3-ml aliquots of the incubation were diluted with 0.7 ml of 0.5% Nonidet P-40, 25 mM Tris-HC1, pH 8.0, buffer. One Fg of GPI-PLC, partially purified by passage through DE52, was added to one of the two mixtures, and both were incubated an additional hour a t 37 "C. A third incubation of diacyl-mfVSG with GPI-PLC in buffer of identical composition was performed simultaneously as a positive control. After 1 h at 37 "C, 10 pl of glacial acetic acid was added to each assay followed by 2 ml of toluene to extract neutral lipids. 0.7- ml samples of the toluene phases were dried under nitrogen gas and redissolved in chloroform with 25 pg each of monomyristin, dimyris- tin, and free myristate. The neutral lipid products were analyzed by tlc on a Silica G plate twice developed in petroleum ether/diethyl- ether/acetic acid in a 70302 ratio. The lipid standards were identified by iodine staining: a, monomyristin; b, dimyristylglycerol; c, myristic acid. 10-mm sections were scraped and mixed with Aquasol-2 for liquid scintillation counting. A, chromatogram of toluene phase from phospholipas~ A,-digested mfVSG. B, chromatogram of toluene phase from mfVSG treated first with phospholipase A:! and then with GPI- PLC. C, c h r o m a ~ ~ a m of toluene phase from mfVSG digested with GPI-PLC.

cation involved the use of anion exchange chromatography followed by chromatography on PI-Sepharose. The resulting enzyme has an apparent molecular weight of 39,800. At this time, the enzyme appears to be at least 30% pure. The enzyme therefore constitutes approximately 0.1% of the total cell protein, equivalent to IO5 molecules/cell.

GPI-PLC appears to be an integral membrane protein, based on the requirement for detergent to achieve solubiliza- tion, phase separation with Triton X-114, and its relative hydrophobicity. The memhrane-ass~iated character of the enzyme is consistent with its specificity for molecules having a glycolipid anchor embedded in the plasma membrane. GPI- PLC exhibits no significant activity toward phosphatidylcho- line, PI, and pol~hosphoinositides, but releases dimyristyl- glycerol from intact mNSG and monomyristylglycerol from lyso-mNSG. The carboxyl-terminal fragment of mNSG (pre- pared by pronase treatment) is also hydrolyzed. However, dimyristyl-PI (prepared by nitrous acid deamination of mfVSG) was not detectably hydrolyzed. In addition, the try- panosomal GPI-PLC hydrolyzes the precursor for an insulin- dependent modulator of CAMP phosphodiesterase, which con- tains carbohydrate-glucosamine coupled to PI (Saltiel and Cuatrecasas, 1986; Saltiel et al., 1986).' By contrast, the

A. Saltiel, unpublished observations.

T. brucei Glycan-phosphatidylinositol Phospholipase C 15771

Staphylococcus aureus PI-specific phospholipase C hydrolyzes both PI and its glycosylated forms (Ferguson et al., 1985b). These data are extended by the recent report (Steiger et al., 1986) that trypanosoma1 GPI-PLC can solubilize the mem- brane form of acetylcholine esterase. Pending further defini- tion of substrate specificity with both modified mfVSG and synthetic analogues of the glycolipid anchor, we conclude that the GPI-PLC from T. brucei is a novel type of phospholipase C with a distinct headgroup specificity for glycan-PI.

The physiological role of this mfVSG-solubilizing enzyme is unclear, as is the mechanism by which its activity is regulated. The calcium-independence of the GPI-PLC sug- gests that it is not directly controlled by calcium ion flux, as previously suggested (Bowles and Voorheis, 1982). Consider- ing its narrow substrate specificity for glycan-PI, the enzyme could be required for VSG turnover and/or for the release of the VSG coat during transformation of the trypanosome from bloodstream to procyclic forms in the midgut of the insect vector (Glossina). During transformation in uitro, the VSG coat is lost and GPI-PLC activity declines (Bulow and Over- ath, 1985). Moreover, the GPI-PLC-catalyzed production of diacylglycerol, a known activator of the calcium and phospho- lipid-dependent protein kinase C, may initiate other biological activities involved in transformation.

The VSGs belong to a family of molecules that are anchored to membranes via covalently bound P I (reviewed in Low et al., 198613). These include the Thy-1 antigen (Low and Kin- cade, 1985), 5”nucleotidase (Low and Finean, 1978), alkaline phosphatase (Low and Finean, 1978; Ikezawa et al., 1976), acetylcholine esterase (Low and Finean, 1977), decay accel- erating factor (Davitz et al., 1986), and a precursor to an insulin-dependent CAMP-phosphodiesterase modulator (Sal- tiel and Cuatrecasas, 1986). During the processing of VSG and Thy-1, hydrophobic peptides at the carboxyl-termini that could adequately function for membrane attachment are re- moved and replaced with phosphatidylinositol-containing an- chors (Boothroyd et al., 1982; Boothroyd et al., 1980; Ferguson and Cross, 1984; Ferguson et al., 1985a, b; Low and Kincade, 1985; Tse et al., 1985). This carboxyl-terminal post-transla- tional modification may identify certain membrane molecules

for processing by the specific endogenous enzyme described here, providing a potential site of regulation for cellular func- tions.

REFERENCES Boothroyd, J. C. (1985) Annu. Rev. Microbiol. 39, 475-502 Boothroyd, J. C., Cross, G. A. M., Hoeijmakers, J. H. J., and Borst,

P. (1980) Nature 288, 624-626 Boothroyd, J. C., Gurnett, L. P., and Cross, G. A. M. (1982) J. Mol.

Biol. 157,527-546 Bordier, C. (1981) J. Biol. Chem. 256, 1604-1607 Bowles, D. J., and Voorheis, H. P. (1982) FEBS Letters 139, 17-21 Bulow, R., and Overath, P. (1985) FEBS Letters 187, 105-110 Cardoso de Almeida, M. L., and Turner, M. J. (1983) Nature 302,

Clarke, M. W., Olafson, R. W., and Pearson, T. W. (1985) Mol.

Cross, G. A. M. (1975) Parasitology 71, 393-417 Davitz, M. A., Low, M. G., and Nussenzweig, V. (1986) J. Enp. Med.

Ferguson, M. A. J., and Cross, G. A. M. (1984) J . Biol. Chem. 259,

Ferguson, M. A. J., Haldar, K., and Cross, G. A. M. (1985a) J. Biol.

Ferguson, M. A. J., Low, M. G., and Cross, G. A. M. (1985b) J. Biol.

Holder, A. A. (1983) Biochem. J . 209, 261-262 Ikezawa, H., Yamanegi, M. Taguchi, R., Miyashita, T., and Ohyabu,

Laemmli, U. K. (1970) Nature 227, 680-685 Low, M. G., Carroll, R. C., and Cox, A. C. (1986a) Biochem. J., in

Low, M. G., Ferguson, M. A. J., Futerman, A. H., and Silman, I.

Low, M. G., and Finean, J. B. (1977) FEBS Letters 82, 143-146 Low, M. G., and Finean, J. B. (1978) Biochim. Biophys. Acta. 508,

Low, M. G., and Kincade, P. W. (1985) Nature 318, 682-687 Saltiel, A. R., and Cuatrecasas, P. (1986) Proc. Natl. Acad. Sci. U. S.

Saltiel, A. R., Fox, J. A,, Sherline, P., and Cuatrecasas, P. (1986),

Steiger, A., Cardosa de Almeida, M.-L., Blatter, M.-C., Brodbeck, U.,

Tse, A. G. D., Barclay, A. N., Watts, A., and Williams, A. F. (1985)

349-352

Biochem. Parasitol. 17, 19-34

163, 1150-1161

3011-3015

Chem. 260,4963-4968

Chem. 260, 14547-14555

T. (1976) Biochim. Biophys. Acta 450, 154-164

press

(1986b) Trends Biochem. Sci. 11,212-215

565-570

A. 83, 5793-5796

Science 233,967-972

and Bordier C. (1986) FEBS Letters 199, 182-186

Science 230, 1003-1008