THE JOURNALOF Bro~oa~cnr.

CHEMISTRY
5, Issueof March 10, PP. X309-1813,1975
Vol. 250, No.
Printed in U.S.A.

Acetyl Phosphatidylethanolamine in the

Reconstitution of Ion Pumps*
(Receivedfor publication, July 25,1974)

AILEEN F. KNOWLES, ANNE KANDRACH, AND EFRAIM RACKER

From the Section of Biochemistry, Molecular and Cell Biology, Cornell University, Ithaca, New York 14850

H. GOBIND KHORANA
From the Departments of Biology and Chemistry, MassachusettsInstitute of Technology, Boston, Massachusetts
02139

SUMMARY Natural membranes contain complex mixtures of lipids.

Downloaded from www.jbc.org by guest, on February 13, 2011

While relatively little definitive information is available regarding
Acetyl phosphatidylethanolaminewas comparedwith phos-
the relationship between lipid structure and membrane function,
phatidylethanolamine in the reconstitution of several biolog- at least two lines of investigation are desirable. One should
ical membrane activities with the following results. focus on the function of the fatty acid side chains and the second
1. The proton pump reconstituted with the purple mem- on the role of the polar head groups. Several studies have fo-
brane of Halobacterium halobium and acetyl phosphatidyl- cused on the former question. Nutritional studies in animals
ethanolamine was quite active. However, some differences as well as in microorganisms have revealed the importance of un-
in the kinetic properties, particularly in the decay rate, were
saturated fatty acids in maintaining fluidity of the membrane.
noted between vesicles reconstituted with phosphatidyleth- Similarly, comparison of the composition of membranes of cold
anolamine and acetyl phosphatidylethanolamine. and warm blooded animals further support this conclusion (cf.
2. Acetyl phosphatidylethanolamine could not replace Ref. 1). Of particular interest are the observations of the rela-
phosphatidylethanolamine in the reconstitution of a Ca2+ tionship between the content of unsaturated fatty acids in yeast
pump with ATPase isolated from sarcoplasmic reticulum. mitochondria and their capability to catalyze oxidative phos-
However, inclusion of suitable amounts of stearylamine or phorylation (2). Further, it was also shown in an in vitro recon-
oleylamine during reconstitution yielded acetyl phosphatidyl- stitution system of the mitochondrial 32Pi-ATP exchange (3) that
ethanolamine vesicles with Ca2+ translocation activity com- unsaturation in the fatty acid side chains was essential although
parable to that of phosphatidylethanolamine vesicles. the saturated side chains could be varied widely. For the study
3. A mixture of acetyl phosphatidylethanolamine and of the role of polar head groups in phospholipids, the successful
stearylamine or oleylamine substituted for phosphatidyl- reconst.itution of defined membrane functions has provided a
ethanolamine in the reconstitution of mitochondrial hydro- promising approach. For example, a reconstituted membrane
phobic proteins to form vesicles that catalyze 32Pi-ATP ex- catalyzing oxidative phosphorylation has a minimal requirement
change. Since phosphatidylcholine is also required in this for two phospholipids, phosphatidylcholine and phosphatidyl-
system, these findings point to two functions of phosphatidyl- ethanolamine (4).
ethanolamine, one related to the specific properties of its The requirement for both phosphatidylcholine and phosphatid-
amino group, the other to a structural role of its small polar ylethanolamine raised the question of the role of the distribution
head group. A hydrophobic alkylamine can fulfill the first of charges on the one hand and of the structural role of the size
function, acetyl phosphatidylethanolamine the second. of the head group in the assembly of the phospholipid bilayer on
4. The importance of the charge was also observed in ex- the other hand.
periments with the reconstituted rutamycin-sensitive ATPase In this paper we describe the effect of a specific alteration of a
of mitochondria. After depletion of phospholipidsfrom the polar head group achieved by acetylation of phosphatidyleth-
hydrophobic proteins, ATPase activity and rutamycin sensi- anolamine on the properties of several reconstituted membranes
tivity were restored only if a phospholipid as well as the that are capable of catalyzing ion translocation and partial reac-
appropriate charge were present. tions of oxidative phosphorylation.
MATERIALS AND METHODS
Phospholipids and other reagents required for the assay of the
various ion transport systems were as described in previous refer-
ences (3-8). Stearylamine and oleylamine were donated by the
* This work was supportedby National Institutes of Health Ashland Chemical Co. Mgristylamine and palmitylamine were
Grant CA-08964 and National Science Foundation Grant GB- purchased from K & K Laboratories, Inc.; dodeiylamine and
30850 (E. R.) and National Institutes of Health Grant A-I 11479 nonylamine from Eastman Organic Chemicals. Dicetylphos-
(G. H. K.). phate was a product of Sigma.

1809
1810

Acetylation of Phosphatidylethanolamine-Three milliliters of TABLE I

acetic anhydride were added to 750 pmoles of phosphatidyletha- Acetyl phosphatidylethanolamine in the reconstitution of bacterial
nolamine purified from soybean phospholipids (4) that were dis-
rhodopsin proton pump
solved in 9 ml of dry pyridine. After 10 min at 0”, 30 ml of water
were added and incubated further for 10 min at 0”. The lipids To 5 @moles of dry soybean phosphatidylethanolamine (PE) or
were then extracted three times with about 50 ml of ether and the acetyl phosphatidylethanolamine (AC-PE), 0.2 ml of 0.15 M KC1
combined extracts were dried in vacua. The lipids were suspended and 20 ~1 of a suspension of purple membranes (6 mg per ml) were
in chloroform-methanol (4: 1) and passed through a Dowex 50 (H+) added. After adjusting the pH to about 6, the mixtures were
column (1 X 10 cm) which had been previously washed with pyri- sonicated as described previously (7) until t,he suspension was
dine and chloroform-methanol. The phospholipids were then ap-
either clear or only slightly cloudy. Clarification occurred more
plied to a Bio-Sil HA (Bio-Rad) column (1.5 X 40 cm) that had
quickly with the acetylated phospholipid. Initial rates of proton
been activated with hexane and washed with chloroform (5).
After washing with 100 ml of chloroform, acetyl phosphatidyl- uptake, extent, and tl/z were measured with samples containing
ethanolamine was eluted with a linear gradient of methanol in 25 rg of bacterial rhodopsin as described previously (8).
chloroform (150 ml of chloroform in the mixing vessel and 150 ml
of methanol in the reservoir). The acetyl phosphatidylethanola- Vesicles PH Initial rate Extent Decay, h,P
amine obtained by this procedure traveled as a single component -
on thin layer chromatograms with chloroform-methanol-water ttatoms H+/nin naloms H+ 5
(65:24:4) with mobility higher than that of phosphatidylethanol- PE 5.30 69 11.1 19
amine. Traces of pyridine in the preparation did not interfere 6.60 59 8.7 8.4
with reconstitutions but could be removed by brief treatment of 7.34 37 7.7 8.4
the sample with a small amount of Dowex 50 (H+). Unless spe- AC-PE 5.50 84 5.5 3
cifically stated otherwise, this preparation of acetyl phosphatidyl- 6.70 60 5.4 3.6
ethanolamine was used in the reconstitution experiments.
7.47 38 4.8 4.2
With synthetic dilaurylphosphatidylethanolamine, the same

Downloaded from www.jbc.org by guest, on February 13, 2011

acetylation procedure was used. However, fractionation on silica
gel was omitted since, after passage through Dowex 50, the prep-
aration revealed a single component on thin layer chromatograms. formed by sonication in the absence of detergents, both phospha-
Assays-Proton translocation (7, 9), aVZazf uptake (6, lo), and tidylethanolamine and phosphatidylcholine were required. We
V-ATP exchange were measured as described in the references. therefore returned to the cholate dialysis procedure to study re-
Reconstitutions-Two methods of reconstitution were used with
minor modifications specified in the legends of the tables and fig- constitution with acetyl phosphatidylethanolamine. As can be
ures: the cholate dialysis procedure (11) and the sonication method seen from Fig. IA, vesicles reconstituted with Ca2+-ATPase and
(8). acetyl phosphatidylethanolamine were completely inactive in
Ca2+ translocation. Act,ivity appeared when suitable amounts
RESULTS
of stearylamine were also included during reconstitution with
Acetyl Phosphatidylethanolamine in Reconstitution of Proton acetyl phosphatidylethanolamine; however, excess of stearyl-
Pump with Bacterial Rhodopsin--The least demanding ion trans- amine was inhibit,ory. As shown in Fig. lB, ATPase activity of
locating system thus far reconstituted is the bacterial rhodopsin the reconstituted vesicles was also markedly depressed by acetyl
proton pump (7). It operates in vesicles reconstituted with phosphatidylethanolamine and restored by stearylamine; how-
synthetic saturated phospholipids such as dimyristoylphos- ever, optimal reactivation of the ATPase activity and Ca*+ pump
phatidylcholine or dipalmitylphosphatidylcholine (9) which are activity did not coincide. It is of interest to note that incorpora-
characterized by known transition temperatures. As shown tion of stearylamine into phosphatidylethanolamine vesicles in-
in Table I, the proton pump was operative with either phos- hibited the Ca*+ pump activity under conditions when the ATP-
phatidylethanolamine or acetyl phosphatidylethanolamine as ase activity was either unaffected or only slightly enhanced.
the only lipid in reconstitution. The initial rate of proton up- Olcylamine was much more effective than stearylamine in re-
take was somewhat faster with the acetylated phospholipid storing Ca*+ pump activity to acetyl phosphatidylethanolamine
particularly at low pH values of the medium. The !I/* of vesicles. As shown in Fig. lA, at optimal concentrations of
decay was considerably shorter with the result that the extent olcylamine the rate of *%a*+ uptake was equal to that of phos-
of proton uptake (which represents the equilibrium state of the phatidylethanolamine vesicles reconstituted under identical con-
rate of uptake and back diffusion) was consistently lower. This ditions. The titration curve with oleylamine showed a sharp
behavior is reminiscent of the response of reconstituted vesicles to peak in comparison to that with stearylamine; excess oleylamine
valinomycin (9). This ionophorc invariably accelerated the rate inhibited markedly. Again, maximal activation of the ATPase
of proton uptake as well as of efflux. Depending on which effect activity took place in a range of oleylamine which was inhibitory
predominated, the extent of proton uptake was somet,imes in- to Ca*+ translocation. As shown in Table II, nonylamine was a
creased, sometimes decreased. It can also be seen from Table I very poor substitute for stearylamine. iit optimal concentra-
that at higher pH both the initial rate and the decay time of tions (12 pmoles/ml) it activated only little CaZf translocation
phosphatidylethanolaminc vesicles were lowered, whereas with but considerable ATPase activity. Dodecylamine was slightly
the acetylated phospholipid the decay time was slightly increased better; myristylamine and palmitylamine restored at best about
at a more alkaline pH. This explains why phosphatidylethanol- 50% of the activity normally found with phosphatidylethanola-
amine vesicles showed a marked drop in the extent of proton mine vesicles.
uptake at higher pH values, whereas the acetyl phosphatidyl- The inhibition of ATPase activity by acetyl phosphatidyleth-
ethanolamine vesicles did not. anolamine was not dependent on reconstitution but could be ob-
Acetyl Phosphatidylethanolamine in Reconstitution of Calcium served on addition of the phospholipid to the enzyme (Fig. 2A).
Pump-It was reported previously (6) that, a Ca2+ translocating As shown in Fig. 2B, stearylaminc counteracted the inhibition
pump can be reconstituted by the cholate dialysis procedure with by acetylated lipid.
the Ca2+-ATPase of sarcoplasmic reticulum (12) in the presence Acetyl Phosphatidylethanolamine in Reconstitution of 32Pi-ATP
of phospholipid mixtures or of phosphatidylethanolamine alone. Exchange and Rutamycin-sensitive A TPase of Mitochondrial Mem-
It was observed later (13) that, when the reconstitution was per- brane-It was shown previously (3) that both phosphatidyl-
 1811

TABLE II
Restoration of calcium uptake activity in vesicles reconstituted with
acetyl phosphatidylethanolamine by alkylamines
Reconstitutions and assays were performed as described in the
legend of Fig. 1. PE, phosphatidylethanolamine; AC-PE, acetyl
phosphatidylethanolamine.

Phospholipid Alkylamine ATPase

activity

Experiment 1
5 IO 15 20
PE None 0 114 1.53
p moles amine per ml AC-PE None 0 0 0.09
AC-PE Stearylamine 10 109 0.8
AC-PE Oleylamine 10 161 1.01
AC-PE Nonylamine 12 3 0.54
Experiment 2
PE None 0 246 1.2
AC-PE Palmitylamine 12 99 1.0
Experiment 3
PE None 0 147 1.07

Downloaded from www.jbc.org by guest, on February 13, 2011

AC-PE Myristylamine 8 61.3 0.81
AC-PE Dodecylamine 12 36.7 0.76

Acefyl- PE + stearylomine vated the ATPase of the hydrophobic protein complex. Addi-
tion of stearylamine decreased the rate of ATP hydrolysis and at
5 IO 15 20 high concentration eliminated rutamycin sensitivity. Nega-
5. p moles omine per ml tively charged dicetyl phosphate had no effect on the activated
ATPase but eliminated rutamycin sensitivity. When stearyla-
FIG. 1. Calcium uptake and ATPese activity of W+-ATPase
vesicles reconstituted from phosphatidylethanolamine (PE), mine and dicetyl phosphate were added together, rutamycin
acetyl phosphatidylethanolamine (AcetyZ-FE), and alkylamines. sensitivity, which was absent with either alone, was restored.
Phosphatidylethanolamine (12.5 amoles) or soybean acetyl phos- Similar effects were observed in the above system with acetyl
phatidylethanolamine (12.5 /rmoles) were mixed with various dilaurylphosphatidylethanolamine. As shown in Fig. 3, a wide
amounts of stearylamine or oleylamine. To the dry lipid mixture range of concentrations of the acetylated phospholipid activated
were added 0.3 ml of 0.4 M potassium phosphate, pH 7.4, and 0.06
ml of 10% potassium cholate, pH 8.0. The suspension was soni- the ATPase activity; however, rutamycin sensitivity was ob-
cated until clear. To the clear lipid solution was added 0.2 mg of served only at very low concentrations (10 to 20 PM), whereas at
Ca*+-ATPase (12) and the final volume was made to 0.5 ml with 0.4 concentrations above 100 PM little rutamycin sensitivity was
M potassium phosphate, pH 7.4. The reconstitution was achieved noted. In the presence of stearylamine (Fig. 4), the ATPase
by dialyzing against 200 volumes of the same buffer overnight.
Calcium uptake and ATP hydrolysis were determined as described activity was suppressed but reactivation was observed with in-
(10). A, calcium uptake activity; B, ATPase activity. creasing amounts of dicetyl phosphate.
These experiments support the proposition that the charge at
the phospholipid surface plays an important role in the operation
choline and phosphatidylethanolamine are required for the re-
of the rutamycin-sensitive ATPase.
constitution of an active 3’Pi-ATP exchange with a mixture of
hydrophobic proteins from the mitochondrial membrane. As
DISCUSSION
shown in Table III, the exchange activity was abolished when
acetyl phosphatidylethanolamine was used instead of phospha- It is apparent from the studies reported in this paper that ion
tidylethanolamine. With optimal amounts of stearylamine dur- pumps vary with respect to their phospholipid requirements.
ing reconstitution, the activity was recovered, but excess amine The amino group of phosphatidylethanolamine is not required
inhibited. Oleylamine was less effective in this system and, with for the proton pump reconstituted with bacterial rhodopsin but
nonylamine, less than 20 v0 of the control activity with phospha- is required for the calcium pump and for the 32Pi-ATP exchange
tidylethanolamine was recovered. Two control experiments are of the mitochondrial membrane. The latter, a partial reaction of
significant. Acetylated stearylamine was inactive and stearyla- oxidative phosphorylation and of the mitochondrial proton pump
mine with phosphatidylcholine alone was insufficient for the (11, 17)) is the most complex of the systems studied since a vesic-
reconstitution of active vesicles. As in the case of C& trans- ular structure provided by phosphatidylcholine together with
port, myristylamine, dodecylamine, and palmitylamine gave only amino groups provided by stearylamine are insufficient for the
partial restoration of the “Pi-ATP exchange activity. operation of the pump. “Pi-ATP exchange activity could not
We observed some time ago (14,15) that the hydrophobic frac- be detected without addition of acetyl phosphatidylethanolamine.
tion isolated from mitochondrial membranes strongly inhibited We can tentatively conclude from this finding that the relatively
the ATPase activity of added coupling factor FL However, on small ethanolamine moiety is required for the proper packing of
addition of phospholipids, an ATPase activity which was rutamy- the polar head groups in one or both of the membrane surfaces.
tin-sensitive was restored. Although from the above it would appear that it is possible to
As shown in Table IV, acetyl phosphatidylethanolamine acti- separate the contribution of the structural component of the
1812

TABLE III
Efect of acetyl phosphatidylethanolamine and alkylamines on “Pi-
ATP exchange
Phosphatidylcholine (PC) (1 pmole) and phosphatidylethanol-
amine (PE) (1.6 pmoles) or soybean acetyl phosphatidylethanol-
amine (AC-PE) were dried together under Nz with the indicated
amounts of various amines, suspended in 0.2 ml of a buffer con-
taining 5 mM N-tris(hydroxymethyl)methylglycine (Tricine)-
KOH, pH 8.2, 2 mM dithiothreitol, 0.1 mM EDTA, 25 mM sucrose,
20 mM ammonium sulfate, and 4 mg of potassium-cholate and
sonicated as described previously (4). The mixture was then
cooled to 0” and 0.1 ml (1 mg) of hydrophobic proteins was added.
After dialysis for 4 hours at 4’ against 125 volumes of a solution

IA0.2 0.4 0.6 08 IO 1.2 1.4 1.6

containing 10% methanol, 10 mM Tricine-KOH,
EDTA, 0.1 mM ATP, 1 mM dithiothreitol,

cles (0.1 ml) were reconstituted

pH 3.0, 0.2 mM
and 200 mM NaCl, the
buffer was changed and dialysis continued overnight. The vesi-
with coupling factors and an-
p. moles acetyl PE alyzed for Pi-ATP exchange as previously described (11).

[=P]ATP
Additions Amine

Downloaded from www.jbc.org by guest, on February 13, 2011

PE + PC 143
AC-PE + PC 5.3
AC-PE + PC + stearylamine 0.125 6.6
0.25 24
0.5 58 62
1.0 112 100
1.5 30
2.0 0.8 1.4
AC-PE + PC + oleylamine 0.5 66
$I 0.1 0.2 03 0.4 0.5
1 1.0
1.5
50 62
3.8
p moles stearylamine AC-PE + PC + nonylamine 1.0 16 22
AC-PE + PC + acetyl stearylamine 1.0 0
FIQ. 2. Inhibition of CW-ATPase from sarcoplasmic reticulum 2.0 0
by acetyl phosphatidylethanolamine (Acetyl-PE) and reactivation PC + stearylamine 1.0 0
by stearylamine. A, Cs?+-ATPase from sarcoplasmic reticulum (32
pg) were incubated with various amounts of soybean acetyl phos-
phatidylethanolamine (a 20 mM stock solution of acetyl phospha-
tidylethanolamine in 5Om~ N-tris(hydroxymethyl)methylglycine
(Tricine)-KOH, pH 8.0, was prepared by sonic dispersion) in a TABLE IV
volume of 0.2 ml. After preincubation for 5 min at room tempera- Eflect of acetyl phosphatidylethanolamine, stearylamine, and dicetyl
ture, the reaction was started by the addition of 1 ml of reaction phosphate on rutamycin-sensitive ATPase
mixture containing 0.1 M KCl, 50 mM Tris-Cl, pH 7.4,lO mM MgClt,
10 m&f ATP, and 50 PM CaClz. The reaction was stopped after 4 To 50 pg of the hydrophobic protein mixture isolated from bo-
min at 37”, and the inorganic phosphate released was determined vine heart mitochondria (ll), soybean acetyl phosphatidyleth-
calorimetrically. B, Cae+-ATPase (32 rg) was preincubated with anolamine, stearylamine, and dicetyl phosphate in the amounts
0.8 rmole of soybean acetyl phosphatidylethanolamine and the indicated in the table were added. ATPase assays were per-
indicated amount of stearylamine (a 50 mM stock solution of formed in a final volume of 1 ml in the presence of an ATP regen-
stearylamine in 50 mM Tricine-KOH, pH 8.0, was prepared by erating system as described previously (16) with or without 4 pg of
sonic dispersion). Preincubation and reaction conditions were
the same as described above. rutamycin present.

ATPase activity
I
phospholipid from that of the amino groups, it is not as simple Additions
to distinguish between the role of the amino groups as a donor of -Rut- $zy$i Inhibition
amycin I I
a dissociable proton from their contribution to the over-all charge
of the lipid surface. A clue that the charge distribution is critical %
was provided by observations that the rutamycin sensitivity of None. . 1.01 0.88 12.8
the reconstituted mitochondrial ATPase exhibits a rather narrow 75 nmoles of Ac-PE . 6.36 4.04 36
optimal range, apparently dependent on the charge which could + 200 nmoles of stearylamine . . . . 3.54 2.03 42
be titrated back and forward by addition of negatively or posi- + 400 nmoles of stearylamine 1.94 1.94 0
tively charged compounds. + 200 nmoles of dicetyl-P. 6.54 5.45 16.6
+ 400 nmoles of dicetyl-P.. 6.0 5.72 4.6
One of the most interesting and challenging problems in the
+ 400 nmoles of dicetyl-P + 200
biogenesis of membranes is the mechanism of the asymmetric or- nmoles of stearylamine. 5.09 2.59 49
ganization of its components. The unidirectionality of ion
 1813

pumps requires a highly specificand asymmetricassemblyof the

participating proteins. It does not appear from the limited
studiesreported in this paper that the aminogroupsof phospha-
tidylethanolamineper seareessentialfor the specificorganization
in the three systemstested.
The problem of asymmetric organization is particularly puz-
zling in view of the observationsthat the reconstituted systems
may either be representative of the orientation of the natural
membraneas in the caseof the reconstituted Ca2+pump or be
assembled in the oppositedirection asin the caseof the reconsti-
tuted rhodopsinproton pump. It seemsnot unreasonable there-
fore to proposethat factors other than the generallipid composi-
tion influence the orientation of the proteins and that specific
0 25 50 75 100 organizational components(either proteins or lipids) participate
n moles ocetyl dilouryl PE in this important function.
FIG. 3. Effect of acetyl dilaurylphosphatidylethanolamine
(acetyl dilauryl PE) on rutamycin-sensitive ATPase. Experimen- REFERENCES
tal conditions were as described in the legend of Table IV except 1. CHAPMAN, D., AND LESLIE,
R. B. (1970) in Membranes of
that acetyl dilaurylphosphatidylethanolamine instead of soybean
Mitochondria and Chloroplasts
(RACKER, E., ed) pp. 91-126,
acetyl phosphatidylethanolamine was added in the amounts indi- Van Nostrand Reinhold Co., New York
cated. 2. HASLAM, J. M., SPITHILL, T. W., AND LINNANE, A. W. (1973)
Biochem. J. 134, 947-957

Downloaded from www.jbc.org by guest, on February 13, 2011

5.0 , 1 3. KAGAWA, Y., KANDRACH, A., AND RACKER, E. (1973) J. Biol.
Chem. 240, 676-684
4. RACKER, E., AND KANDRACH, A. (1973) J. Biol. Chem. 248,5841-
5847
5. RAGAN, C. I., AND RACKER, E. (1973) J. Biol. Chem. 246,2563-
2569
6. RACKER, E. (1972) J. Biol. Chem. 247,8198-8200
7. RACKER, E., AND STOECKENIUS, W. (1974) J. Biol. Chem. 249,
662-663
8. RACKER, E. (1973) Biochem. Biophys. Res. Commun. 66,224-230
9. RACSER, E., AND HINKLE, P. (1974) J. Membr. Biol. 17, 181-188
10. KNOWLES, A. F., AND RACKER, E. (1975) J. Biol. Chem. 260,
in press
11. KAGAWA, Y., AND RACKER, E. (1971) J. Biol. Chem. 246, 5477-
5487
12. MACLENNAN, D. H. (1970) J. Biol. Chem. 246,4508-4518
13. RACKER, E., AND EYTAN, E. (1973) Biochem. Biophys. Res.
I I I I I Commun. 66, 174-178
50 100 150 200 300
14. KAGAWA, Y., AND RACKER, E. (1966) J. Biol. Chem. 241, 2467-
n moles dicetyl phosphate 2474
FIG. 4. Activation of inhibited ATPase by dicetyl phosphate. 15. BULOS, B., AND RACKER, E. (1968) J. Biol. Chem. 243, 3901-
Experimental conditions were as described in the legend of Fig. 3 3905
except that 37.5 nmol of acetyl dilaurylphosphatidylethanolamine 16. PULLMAN, M. E., PENEFSKY, H. S., DATTA, A., AND RACKER,
and 20 nmol of stearylamine were added to the ATPase complex to- E. (1960) J. Biol. Chem. 236, 3322-3329
gether with increasing amounts of dicetyl phosphate as indicated. 17. MITCHELL, P. (1966) Biol. Rev. (Camb.) 41,445