Insulin stimulation of phospholipid methylation in isolated rat ...

Proc. Natl. Acad. Sci. USA Vol. 81, pp. 1089-1092, February 1984 Biochemistry

Insulin stimulation of phospholipid methylation in isolated rat adipocyte plasma membranes (phosphatidylethanolamine/phosphatidylcholine/S-adenosyl-L-methionine)

KATHLEEN L. KELLY, FREDERICK L. KIECHLE, AND LEONARD JARETT Department of Pathology and Laboratory Medicine, University of Pennsylvania School of Medicine, Philadelphia, PA 19104

Communicated by Paul E. Lacy, November 10, 1983

ABSTRACT Partially purified plasma membranes prepared from rat adipocytes contain N-methyltraitsferase(s) that utilize(s) S-adenosyl-L-methionine to synthesize phosphatidylcholine from phosphatidylethanolamine. The incorporation of [3H]methyl from S-adenosyl-L-[methyl-3H]methionine into plasma membrane phospholipids was linear with incubatiop time and plasma membrane protein concentration and was inhibited in a dose-dependent manner by both S-adenosyl-L-homocysteine and 3-deazadenosine. The addition of insulin to plasma membranes stimulated the methylation of endogenous phosphatidylethanolamine, as evidenced by an increase in the levels of phosphatidyl-N-monomethylethanolamine, phosphatidyl-N,N-dimethylethanolamine, and phosphatidylcholine. The effect of insulin was rapid and concentration-dependent, with 100 microunits/ml providing near maximal stimulation. The incorporation of [3H]methyl into phospholipids of control and insulin-stimulated plasma membranes was enhanced by the addition of exogenous methyltransferase substrates phosphatidylethanolamine, phosphatidyl-N-monomethylethanolamine, and phosphatidyl-N,N-dimethylethanolamine. The stimulatory effect of insulin on adipocyte plasma membrane phospholipid methylation may have a physiological role in insulin action.

The conversion by methylation of phosphatidylethanolamine to phosphatidylcholine was first reported in rat liver microsomes by Bremer and Greenberg (1). After this report, investigators examined the possibility that enzymatic methylation of phospholipids plays a role in the transduction of receptormediated signals through the membranes of a variety of cells (2). In addition to agents such as f-agonists (3), benzodiazepine (4), dopamine (5), concanavalin A (6), and IgE antibody (7), several peptide hormones have been shown to alter phospholipid methylation, including glucagon (8), vasopressin (9), angiotensin (10), bradykinin (11), and nerve growth factor (12). This process is rapid and has been localized to a stimulation of plasma membrane methyltransferase activity in certain systems (13). The activity of phospholipid methyltransferases in liver endoplasmic reticulum of alloxan-treated diabetic rats is reduced as compared to control (14), suggesting that insulin may induce the opposite effect, or increase phospholipid methylation. However, no direct evidence has been reported to date to support the hypothesis that insulin, after binding to its receptor, alters phospholipid methylation. We report here that partially purified plasma membranes prepared from rat adipocytes contain a phosphatidylethanolamine methyltransferase (PtdEtn MeTase) system and that insulin stimulates this enzyme system in a concentration-dependent manner. The effects of insulin were rapid and were observed as early as 15 sec after the addition of the hormone. The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

MATERIALS AND METHODS Materials. Male Sprague-Dawley rats (120 g) were obtained from Ace Animals (Boyertown, PA). Collagenase, albumin (fraction V), L-a-phosphatidylethanolamine (type III), L-a-phosphatidyl-N-monomethylethanolamine, L-aphosphatidyl-N,N-dimethylethanolamine, and most other chemicals were purchased from Sigma. Insulin was a gift from R. Chance of Eli Lilly. S-Adenosyl-L-[methyl-3H]methionine was purchased from New England Nuclear. OCS organic counting scintillant was from Amersham. The lots of collagenase and albumin were chosen as in ref. 15. Preparation of Plasma Membranes. Adipocytes prepared from rat epidydimal fat pads were isolated by using the collagenase digestion technique of Rodbell (16) as modified by Jarett (15). Highly enriched plasma membranes were prepared in the absence of EDTA as described by Kiechle et al. (17) with minor modifications. The cells were homogenized in 10.0 mM 4-morpholenepropanesulfonic acid/0.25 M sucrose buffer, pH 7.4. The pellet containing plasma membranes and mitochondria obtained after centrifugation at 20,000 x g was fractionated on a discontinuous sucrose gradient of 0.8 M, 1.06 M, and 2.02 M sucrose. Plasma membranes were isolated at the interface between 0.8 M and 1.06 M sucrose, washed, resuspended, and frozen at -70'C in 50 mM phosphate buffer (pH 7.4) until use. An endoplasmic reticulum-enriched microsomal fraction was obtained from the 20,000 x g supernatant, as described (15). Protein concentration was determined by the method of Lowry et al. (18). Assay of PtdEtn MeTase Activity. Hirata et al. (19) have identified two methyltransferases in the synthesis of phosphatidylcholine from phosphatidylethanolamine, which together constitute PtdEtn MeTase. Without differentiating between the two methyltransferases, the activity of PtdEtn MeTase was assayed as the incorporation of [3H]methyl from S-adenosyl-L-[methyl-3H]methionine into membrane phospholipids by using a modification of the method of Hirata et al. (3). The methylation reaction was initiated by incubating adipocyte plasma membranes (0.2-2.0 mg/ml of protein) in 50 mM phosphate buffer supplemented with 5% bovine serum albumin (pH 7.4) containing S-adenosyl-L-[methyl-3H]methionine (0.2-20 AM, 0.2-2.0 fmol/cpm). The addition of 5% bovine serum albumin decreased insulin degradation by plasma membranes, as measured by trichloroacetic acid precipitation of insulin (20) and enhanced the stimulatory effect of insulin on methylation (data not shown). In designated experiments, one of three phospholipids, phosphatidylethanol-

amine, phosphatidyl-N-monomethylethanolamine, or phosphatidyl-N,N-dimethylethanolamine, was resuspended to

2.0 mM by sonication and added at a final concentration of 0.4 mM in 0.1% propylene glycol to the membrane incubation buffer. After this preincubation period, insulin in 0.1% Abbreviation: PtdEtn MeTase, phosphatidylethanolamine methyltransferase.

1089

1090

Biochemistry: Kelly et aL

bovine serum albumin or 0.1% bovine serum albumin alone (control) was added to the membrane preparation such that the final incubation volume was 0.1-0.5 ml. Aliquots of the membrane suspension were taken prior to and at designated time points after the addition of insulin or bovine serum albumin, and phospholipids were extracted by the method of Bligh and Dyer (21) as modified by Lapetina and Michell (22). For the measurement of total phospholipid methylation, an aliquot of the chloroform phase was transferred to a radioactivity-assay vial, evaporated to dryness, and resuspended in 3 ml of OCS scintillation cocktail for determination of radioactivity. To separate and identify the various methylated phospholipids, an aliquot of the chloroform phase was evaporated to dryness under a stream of argon, the residue was resuspended in a small volume of chloroform/methanol, 1:1 (vol/vol), and applied to a Whatman LK-5D silica gel plate. Phosphatidylethanolamine, phosphatidyl-N-monomethylethanolamine, phosphatidyl-N,N-dimethylethanolamine, and phosphatidylcholine were identified by cochromatography with authentic standards as described by Hirata et al. (3) in a solvent system of propionic acid/propanol/chloroform/water, 2:3:2:1 (vol/vol). Quantitative analysis of the methylated derivatives of phosphatidylethanolamine was accomplished by visualizing the standards after exposure to iodine, scraping the TLC plate into fractions corresponding to known standards, and assaying the samples in a liquid scintillation counter.

RESULTS Presence of PtdEtn MeTase activity in Rat Adipocyte Plasma Membranes. The initial experiments in this study showed the presence of PtdEtn MeTase in plasma membranes prepared from rat adipocytes. The incorporation of [3H]methyl into phospholipids of control plasma membranes was linear for at least 20 min of incubation at 370C (data not shown). There was a linear relationship between adipocyte plasma membrane concentration from 0.05-2.0 mg/ml and the incorporation of [3H]methyl into total organic-extracted phospholipids (data not shown). The activity of PtdEtn MeTase in subcellular fractions of liver has been reported to be highest in endoplasmic reticulum (23). We compared the specific activity of PtdEtn MeTase in the plasma membrane fraction of rat adipocytes to that of an endoplasmic reticulum-enriched microsomal fraction prepared by ultracentrifugation of a 20,000 x g supernatant of adipocyte homogenate (15). PtdEtn MeTase activity, expressed per mg of membrane protein, in the plasma membrane fraction was 65% of that in the endoplasmic reticulumenriched microsomal fraction (data not shown). Contamination of the plasma membrane fraction with endoplasmic reticulum was 12-20% as determined by cytochrome c reductase activity (15). In contrast, contamination of the endoplasmic reticulum fraction with plasma membranes was 512% as determined by 5'-nucleotidase and adenylate cyclase activity (15). Because the recovery of total protein in the plasma membrane fraction was greater than in the endoplasmic reticulum-enriched microsomal fraction, there was more PtdEtn MeTase activity per mg of protein in the plasma membrane fraction than could be accounted for by contamination with endoplasmic reticulum, based on these enzyme marker contamination figures and protein recovery. Plasma membranes were prepared in the absence of any divalent-cation chelator, and the addition of Mg2+ (0.4-4.0 mM) to the assay buffer had no effect on phospholipid methylation (data not shown). Two inhibitors of phospholipid methylation, S-adenosylhomocysteine (24) and 3-deazadenosine (25), inhibited the incorporation of [3H]methyl into phospholipids in a concentration-dependent manner (Fig. 1). The concentration of S-adenosylhomocysteine required for

Proc. NatL. Acad. Sci. USA 81 (1984) 100 90 CO

80

E 70 +

..60

0

50

v

+o40 0

30

W

20

1 20

10 -

9

I

-

-9 -8 -7 -6 -5 -4 -3 -2

Inhibitor, log M

FIG. 1. Inhibition of adipocyte plasma membrane PtdEtn MeTase activity by S-adenosylhomocysteine and 3-deazadenosine. Plasma membranes (1.28 mg/ml) were incubated for 10 min at 370C with S-adenosyl-L-[methyl-3H]methionine (0.06 ,uM; 0.2 fmol/cpm) in the absence (control) or presence of the methyltransferase inhibitor 3-deazadenosine (e) or S-adenosylhomocysteine (A). The data are expressed as the incorporation of [3H]methyl into total organic extracted phospholipids in the presence of inhibitors, as compared with the incorporation of [3H]methyl in the absence of inhibitors (control).

50% inhibition of phospholipid methylation was 10 nM, whereas 1 mM 3-deazadenosine was required for the same inhibitory effect. Insulin Stimulation of PtdEtn MeTase Activity in Adipocyte Plasma Membranes. Insulin increased the incorporation of [3H]methyl into total organic-extracted phospholipids in a concentration-dependent manner from 10 to 1000 microunits of insulin per ml (Fig. 2). Insulin at 100 microunits/ml stimulated PtdEtn MeTase activity to 51 ± 5% above control activity. This physiological concentration of insulin provided near maximal stimulation. In subsequent experiments, insulin at a physiological concentration of 100 microunits/ml was used to further characterize the stimulatory effect of insulin on phospholipid methylation. The increase in PtdEtn MeTase activity caused by insulin was maintained for up to 10 min (data not shown). The stimulatory effect of insulin on phospholipid methylation was rapid and apparent as early as 15 and 30 sec after addition of insulin at 100 microunits/ml (Fig. 3). The stimulation of [3H]methyl incorporation into phospholipids by insulin results from an increase in all three methylated products: phosphatidyl-N-monomethylethanolamine, phosphatidylN,N-dimethylethanolamine, and phosphatidylcholine. The radioactivity recovered in these three phospholipids represents 90 ± 0.8% of the total organic-extracted radioactivity. The Effect of Phospholipid Substrates on Control and Insulin-Stimulated PtdEtn MeTase Activity. Several phospholipid substrates were added to the incubation mixture to determine their effect on control and insulin-stimulated PtdEtn MeTase activity. Each of the three phospholipid substrates tested, phosphatidylethanolamine, phosphatidyl-N-monomethylethanolamine, and phosphatidyl-N,N-dimethylethanolamine increased control PtdEtn MeTase activity (Table 1). The addition of phosphatidylethanolamine alone slightly increased control PtdEtn MeTase activity by 25 ± 5%, whereas the addition of phosphatidyl-N-monomethylethanolamine or phosphatidyl-N,N-dimethylethanolamine increased control PtdEtn MeTase activity by 91 ± 16% and 85 ± 11%, respectively. In the presence of exogenous phospho-

Biochemistry: Kelly

Proc. NatL. Acad. Sci. USA 81 (1984)

et aL

1.47

Table 1. Effect of exogenously added phospholipids on adipocyte plasma membrane PtdEtn MeTase activity [3H]Methyl incorporation, % of control Insulin Control Addition 141 ± 6* (6) 100 None 158 ± 15t (12) 125 ± 5 (12) PtdEtn 246 ± 20 (14) 191 ± 16 (15) PtdMeEtn 241 ± 19f (9) 185 ± 11(12) PtdMe2Etn Results are expressed as percent PtdEtn MeTase activity in plasma membranes incubated in the presence of 0.1% bovine serum albumin, without the addition of phospholipids (0.56 ± 0.05 nmol/ mg per 10 min). The number of determinations for each condition is given in parentheses. Significance was determined for the effect of insulin (100 microunits/ml) as compared to its respective control, using Student's paired t test and unpaired t test. PtdEtn, phosphatidylethanolamine; PtdMeEtn, phosphatidyl-N-monomethylethanolamine; PtdMeEtn, phosphatidyl-NN-dimethylethanolamine. *P < 0.01. tP < 0.05. tP < 0.025.

1.2 **

E Eon 1.0 F

+

C

-Ev 0.8 V 0 C.

0.6 F

I

*

A,

5 C._ L-

0.4 F

IL-

0.2 F

"

20 100 10) Insulin. microunits/ml

C

1091

DISCUSSION

1(00

FIG. 2. Insulin stimulation of adipocyte plasma membrane PtdEtn MeTase activity. Plasma membranes (0.92 mg/ml) were incubated with S-adenosyl-L-[methyl-3H]methionine (20 AtM; 3.0 fmol/cpm) in the presence of insulin (10, 20, 100, or 1000 microunits/ml) in 0.1% bovine serum albumin or 0.1% bovine serum albumin alone (control) for 10 min at 370C. Results represent the mean + SEM of triplicate determinations. Significance was determined using a paired Student t test. *, P < 0.025; **, P < 0.01.

lipids, insulin enhanced the absolute amount of incorporation of [3H]methyl into phospholipids. The percentage stimulation caused by insulin in the presence of exogenous substrates was similar to that obtained in the absence of added phospholipids, as compared to their respective controls. In the presence of insulin and added phosphatidylethanolamine, phosphatidyl-N-monomethylethanolamine, or phosphatidyl-NN-dimethylethanolamine, PtdEtn MeTase activity was increased 33 ± 15%, 55 ± 20%, and 56 ± 16%, respectively.

These data demonstrate the presence of PtdEtn MeTase in plasma membranes prepared from rat adipocytes, which is stimulated by physiological concentrations of insulin. The amount and/or activity of this enzyme in plasma membranes could not be attributed to endoplasmic reticulum contamination based on a comparison of the specific activities of PtdEtn MeTase in each fraction and protein recovery, even after adjusting for contamination based on enzyme marker studies. Our observations on adipocyte plasma membrane PtdEtn MeTase are consistent with previous studies in other cell types, including rat reticulocytes (3), leukocytes (26), lymphocytes (27), and synaptosomes (28) that suggest a surface membrane localization. The PtdEtn MeTase activity of adipocyte plasma membranes is higher than that reported for adrenal medulla endoplasmic reticulum (19), reticulocyte ghosts (3), or erythrocyte ghosts (28) and comparable to that reported for synaptosomes (5) and liver endoplasmic reticulum (23). Like the PtdEtn MeTase associated with rat liver endoplasmic reticulum (23) and L-929 cells (13), adipocyte plasma membrane PtdEtn MeTase was found to be indepen-

B

E 0

E

C. C

9:

O C.

C O LC

-

av A.

a

2 1 Distance from origin, cm

3

4

5

6

7

8

9

FIG. 3. Distribution of [3H]methyl into adipocyte plasma membrane phospholipids. Plasma membranes were preincubated for 15 min with S-adenosyl-L-[methyl-3H]methionine (0.2 gM; 0.4 fmol/cpm) at 370C. Insulin (100 microunits/ml) in 0.1% bovine serum albumin (M) or 0.1% bovine serum albumin alone ([z) was then added for 15 (A) or 30 (B) sec as indicated. Columns: a, Phosphatidylcholine; b, phosphatidyl-N,Ndimethylethanolamine; c, phosphatidyl-N-monomethylethanolamine.

1092

Biochemistry: Kelly et aL

dent of exogenous Mg2+. After the incubation of membranes in the presence of EDTA, PtdEtn MeTase requires the addition of magnesium to restore activity (19). Divalent-cation chelators were not present during the preparation or incubation of our adipocyte plasma membranes, and the concentration of Mg2+ retained in the membranes was probably sufficient to maintain PtdEtn MeTase activity. PtdEtn MeTase in plasma membranes was inhibited by S-adenosyl-L-homocysteine, a competitive inhibitor of transmethylation reactions (25). The efficacy of 3-deazadenosine as a methyltransferase inhibitor is dependent on the intracellular conversion of this compound to 3-deazadenosylhomocysteine, an analog of Sadenosyl-L-homocysteine, and also a competitive inhibitor of S-adenosylhomocysteine hydrolase (25). In our subcellular system there is probably little conversion of 3-deazadenosine to 3-deazadenosylhomocysteine; thus, the adenosine analog 3-deazadenosine was a less-potent inhibitor of phospholipid methylation in adipocyte plasma membranes than has been reported for whole cells (29). The addition of physiological concentrations of insulin to plasma membranes enhanced the incorporation of [3H]methyl into phospholipids in a concentration-dependent manner, with a maximally effective concentration of 100 microunits/ml. The stimulatory effect of insulin at 100 microunits/ml on PtdEtn MeTase activity was apparent as early as 15 sec after the addition of the hormone and was maintained for at least 10 min after the addition of insulin. This rapid increase in phospholipid methylation was reflected in all three methylated derivatives of phosphatidylethanolamine: phosphatidyl-N-monomethylethanolamine, phosphatidyl-N,N-dimethylethanolamine, and phosphatidylcholine. The addition of phosphatidylethanolamine, phosphatidyl-Nmonomethylethanolamine, phosphatidyl-N,N-dimethylethanolamine increased the absolute amount of insulin-stimulated incorporation of [3H]methyl into phospholipids. These findings suggest that the endogenous level of phosphatidylethanolamine in the plasma membrane fractions may have been rate-limiting. Further studies are required to determine whether methyltransferase I or II is activated by insulin. Hirata and Axelrod (2) have presented evidence to suggest that alterations in phospholipid methylation in a variety of cell types are associated with changes in membrane fluidity and receptor functions. A role for phospholipid methylation has been proposed both in the coupling of receptor-effector systems (30) as well as in the formation of hormone-receptor complexes (31). After the binding of insulin to its cell surface receptor, several responses occur that are analogous to receptor activation in other systems where the participation of phospholipid methylation has been implicated as a transducing mechanism. In adipocyte plasma membranes, binding of insulin to its receptor is accompanied by covalent modification of the insulin receptor (32), stimulation of cellsurface transport (33), hyperpolarization of membranes (34), an alteration in membrane fluidity (35), and the generation of insulin mediator (36). We have shown here that insulin activates PtdEtn MeTase. It remains to be determined what role this pathway plays in the action of insulin. We thank Judy Smith, Cynthia Penn, and Janet Macaulay for their excellent technical assistance. This work was supported in part by U.S. Public Health Service Grant AM-28144. F.L.K. is a Hartford Foundation Fellow. 1. Bremer, J. & Greenberg, S. M. (1961) Biochim. Biophys. Acta 46, 205-216.

Proc. Natl. Acad. Sci. USA 81 (1984) 2. Hirata, F. & Axelrod, J. (1980) Science 209, 1082-1090. 3. Hirata, F., Strittmatter, W. J. & Axelrod, J. (1979) Proc. Natl. Acad. Sci. USA 76, 368-372. 4. Strittmatter, W. J., Hirata, F., Axelrod, J., Mallorga, P., Tallman, J. F. & Henneberry, R. C. (1979) Nature (London) 282, 857-859. 5. Leprohon, C. E., Blusztajn, J. K. & Wurtman, R. J. (1983) Proc. Nall. Acad. Sci. USA 80, 2063-2066. 6. Hirata, F., Toyoshima, S. & Axelrod, J. (1980) Proc. Natl. Acad. Sci. USA 77, 862-865. 7. Crews, F. T., Morita, Y., Hirata, F., Axelrod, J. & Siraganian, R. P. (1980) Biochem. Biophys. Res. Commun. 93, 42-49. 8. Castano, J. C., Alemany, S., Nieto, A. & Mato, J. M. (1980) J. Biol. Chem. 255, 9041-9043. 9. Prasad, C. & Edwards, R. M. (1981) Biochem. Biophys. Res. Commun. 103, 559-564. 10. Alemany, S., Varela, I. & Mato, J. M. (1981) FEBS Lett. 135, 111-114. 11. Bareis, D. L., Manganiello, V. C., Hirata, F., Maughan, M. & Axelrod, J. (1983) Proc. Natl. Acad. Sci. USA 80, 2514-2518. 12. Pfenninger, K. H. & Johnson, M. P. (1981) Proc. Natl. Acad. Sci. USA 78, 7797-7800. 13. Garg, L. C. & Brown, J. C. (1983) Arch. Biochem. Biophys. 220, 22-30. 14. Hoffman, D. R., Haning, J. A. & Cornatzer, W. E. (1981) Proc. Exp. Biol. Med. 167, 143-146. 15. Jarett, L. (1974) Methods Enzymol. 31, 60-71. 16. Rodbell, M. (1964) J. Biol. Chem. 239, 375-380. 17. Kiechle, F. L., Jarett, L., Kotagal, N. & Popp, D. A. (1980) J. Biol. Chem. 256, 2945-2951. 18. Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J. (1951) J. Biol. Chem. 193, 265-275. 19. Hirata, F., Viveros, 0. H., Dilberto, E. J., Jr., & Axelrod, J. (1978) Proc. Nail. Acad. Sci. USA 75, 1718-1721. 20. Hammon, G. T., Smith, R. M. & Jarett, L. (1982) J. Biol. Chem. 257, 11563-11570. 21. Bligh, E. G. & Dyer, W. J. (1959) Can. J. Biochem. Physiol. 37, 911-917. 22. Lapetina, E. G. & Michell, R. H. (1972) Biochem. J. 126, 1141-1147. 23. Schneider, W. J. & Vance, D. E. (1978) J. Biol. Chem. 254, 3886-3891. 24. Schiffman, E., O'Dea, R. F., Chiang, P. C., Venkatsubramian, K., Corcoran, B. A., Hirata, F. & Axelrod, J. (1979) in Modulation of Protein Function, eds. Atkinson, D. & Fox, C. F. (Academic, New York), pp. 299-313. 25. Chiang, P. K., Richards, H. H. & Cantoni, G. L. (1977) Mol. Pharmacol. 13, 939-947. 26. Bareis, D. L., Hirata, F., Schiffman, E. & Axelrod, J. (1982) J. Cell Biol. 93, 690-697. 27. Hoffman, T., Hirata, F., Bougnoux, P., Frager, B. A., Goldfarb, R. H., Herberman, R. B. & Axelrod, J. (1981) Proc. Natl. Acad. Sci. USA 78, 3839-3843. 28. Hirata, F. & Axelrod, J. (1978) Nature (London) 275, 219-220. 29. Pritchard, P. H., Chiang, P. K., Cantoni, G. L. & Vance, D. E. (1982) J. Biol. Chem. 257, 6362-6367. 30. Munzel, P. & Koschel, K. (1982) Proc. Natl. Acad. Sci. USA 79, 3692-3696. 31. Bhattacharya, A. & Vonderhaar, B. K. (1981) Proc. Natl. Acad. Sci. USA 78, 5704-5707. 32. Petruzelli, L. M., Ganguly, S., Smith, C. J., Lobb, M. H., Rubin, C. S. & Rosen, 0. M. (1982) Proc. Nail. Acad. Sci. USA 79, 6792-6796. 33. Melchoir, D. L. & Czech, M. P. (1980) J. Biol. Chem. 255, 1722-1731. 34. Zierler, K. L. (1960) Am. J. Physiol. 198, 1066-1070. 35. Luly, P. & Shinitzky, M. (1979) Biochemistry 18, 445-449. 36. Jarett, L., Kiechle, F. L., Parker, J. C. & Macaulay, S. L. (1983) Am. J. Med. 74, 31-37.