Purification of recombinant acyl-coenzyme a:cholesterol acyltransferase 1 (ACAT1) from H293 cells and binding studies between the enzyme and substrates using difference intrinsic fluorescence spectroscopy

Article abstract of DOI:10.1021/bi1013936

Acyl-coenzyme A:cholesterol acyltransferase 1 (ACAT1) is a membrane-bound enzyme utilizing long-chain fatty acyl-coenzyme A and cholesterol to form cholesteryl esters and coenzyme A. Previously, we had expressed tagged human ACAT1 (hACAT1) in CHO cells and purified it to homogeneity; however, only a sparse amount of purified protein could be obtained. Here we report that the hACAT1 expression level in H293 cells is 18-fold higher than that in CHO cells. We have developed a milder purification procedure to purify the enzyme to homogeneity. The abundance of the purified protein enabled us to conduct difference intrinsic fluorescence spectroscopy to study the binding between the enzyme and its substrates in CHAPS/phospholipid mixed micelles. The results show that oleoyl-CoA binds to ACAT1 with K_d = 1.9 μM and elicits significant structural changes of the protein as manifested by the significantly positive changes in its fluorescence spectrum; stearoyl-CoA elicits a similar spectrum change but much lower in magnitude. Previously, kinetic studies had shown that cholesterol is an efficient substrate and an allosteric activator of ACAT1, while its diastereomer epicholesterol is neither a substrate nor an activator. Here we show that both cholesterol and epicholesterol induce positive changes in the ACAT1 fluorescence spectrum; however, the magnitude of spectrum changes induced by cholesterol is much larger than epicholesterol. These results show that stereospecificity, governed by the 3β-OH moiety in steroid ring A, plays an important role in the binding of cholesterol to ACAT1.

Full text of DOI:10.1021/bi1013936

Biochemistry 2010, 49, 9957–9963 9957
DOI: 10.1021/bi1013936
Purification of Recombinant Acyl-Coenzyme A:Cholesterol Acyltransferase 1 (ACAT1)
from H293 Cells and Binding Studies between the Enzyme and Substrates
Using Difference Intrinsic Fluorescence Spectroscopy†
Catherine C. Y. Chang,*,‡ Akira Miyazaki,§ Ruhong Dong,‡ Alireza Kheirollah,‡ Chunjiang Yu, Yong Geng,‡
Henry N. Higgs,‡ and Ta-Yuan Chang*,‡
‡Department of Biochemistry, Dartmouth Medical School, Hanover, New Hampshire 03755, United States,
§Department of Biochemistry, Showa University School of Medicine, Tokyo, Japan, and Department of
Anatomy and Cell Biology, University of Illinois, Chicago, Illinois 60612, United States
Received August 28, 2010; Revised Manuscript Received October 14, 2010
ABSTRACT: Acyl-coenzyme A:cholesterol acyltransferase 1 (ACAT1) is a membrane-bound enzyme utilizing
long-chain fatty acyl-coenzyme A and cholesterol to form cholesteryl esters and coenzyme A. Previously, we
had expressed tagged human ACAT1 (hACAT1) in CHO cells and purified it to homogeneity; however, only
a sparse amount of purified protein could be obtained. Here we report that the hACAT1 expression level in
H293 cells is 18-fold higher than that in CHO cells. We have developed a milder purification procedure to
purify the enzyme to homogeneity. The abundance of the purified protein enabled us to conduct difference
intrinsic fluorescence spectroscopy to study the binding between the enzyme and its substrates in CHAPS/
phospholipid mixed micelles. The results show that oleoyl-CoA binds to ACAT1 with Kd = 1.9 μM and elicits
significant structural changes of the protein as manifested by the significantly positive changes in its
fluorescence spectrum; stearoyl-CoA elicits a similar spectrum change but much lower in magnitude.
Previously, kinetic studies had shown that cholesterol is an efficient substrate and an allosteric activator of
ACAT1, while its diastereomer epicholesterol is neither a substrate nor an activator. Here we show that both
cholesterol and epicholesterol induce positive changes in the ACAT1 fluorescence spectrum; however, the
magnitude of spectrum changes induced by cholesterol is much larger than epicholesterol. These results show
that stereospecificity, governed by the 3β-OH moiety in steroid ring A, plays an important role in the binding
of cholesterol to ACAT1.
Acyl-coenzyme A:cholesterol acyltransferase utilizes long-chain
fatty acyl-coenzyme A and cholesterol as its substrates, producing
cholesteryl esters and coenzyme A as its products. In mammals,
there are two ACAT1 genes that encode two similar but different
proteins, ACAT1 (1) and ACAT2 (2-4). Tissue distribution
studies show that ACAT1 is ubiquitously expressed, while ACAT2
is mainly expressed in intestinal enterocytes and in hepatocytes, as
reviewed in ref 5. Both ACAT1 and ACAT2 are potential drug
targets for therapeutic intervention of dyslipidemia and athero-
sclerosis (6, 7). ACAT1, ACAT2, and its close homologue di-
acylglycerol acyltransferase 1 (8) are founding members of the
membrane-bound acyltransferase (MBOAT) superfamily (9). The
MBOATs are fatty acyltransferases with multiple transmembrane
domains (TMD) and with an invariant histidine (His) embedded
in a long stretch of hydrophobic residues. There are more than
a dozen MBOAT members that include grehlin acyltransferase
(10, 11) and certain lysophospholipid acyltransferases (12 , 13).
ACAT1 is a resident enzyme in the ER (14-16); the enzyme is
homotetrameric (17) and contains nine TMDs, with the active
site His located within TMD 7 (18). The recombinant human
ACAT1 (hACAT1) expressed in Chinese hamster ovary (CHO)
cells has been solubilized in zwitterionic detergent CHAPS
and purified several thousandfold to homogeneity (19). With
the purified enzyme, kinetic studies showed that the cholesterol
substrate saturation curves, assayed either in mixed micelles or in
reconstituted vesicles, were sigmoidal, while the oleoyl-coenzyme
A substrate saturation curve was hyperbolic, suggesting that
ACAT1 is an allosteric enzyme activated by its own substrate
cholesterol (19). Upon activation by cholesterol, the enzyme
becomes promiscuous and is able to accommodate a variety of
sterols with 3β-OH at ring A as its substrates (20, 21). A major
drawback of the CHO cell expression system for purifying
ACAT1 is that only a low quantity of ACAT1 protein could
be obtained (19). The sparse amount of purified protein available
has hampered the research progress of ACAT1 structure-
function analysis. We have sought to increase hACAT1 yield by
using various other expression systems that include Escherichia
coli, yeast, and the baculovirus-infected insect cells but were
unsuccessful. In the current work, we report that hACAT1 can
be expressed in H293 cells at levels much higher than that in the
CHO cells. We next developed an efficient purification procedure
to purify the recombinant hACAT1 from H293 cells to homo-
geneity. The improved expression and purification procedure
enabled us to obtain enough purified protein to demonstrate that
hACAT1 is a highly fluorescent protein. We have exploited this
feature by performing binding studies between hACAT1 and its
substrates.
†This work is supported by NIH Grant RO1 HL60306 to T.-Y.C. and
C.C.Y.C. and NIH Grant RO1 GM069818 to H.N.H.
*Corresponding authors. Phone: (603) 650-1809. Fax: (603) 650-1129.
E-mail: catherine.chang@dartmouth.edu; ta.yuan.chang@dartmouth.edu.
1Abbreviations: ACAT, acyl-coenzyme A:cholesterol acyltransferase;
CHO, Chinese hamster ovary; CoA, coenzyme A; CHAPS, 3-[(3-cholamido-
propyl)dimethylammonio]-1-propanesulfonate; PC, phosphatidylcholine;
TMD, transmembrane domain. His, histidine; KCl, potassium chloride.
r
2010 American Chemical Society
Published on Web 10/22/2010
pubs.acs.org/Biochemistry

9958 Biochemistry, Vol. 49, No. 46, 2010
Chang et al.
EXPERIMENTAL PROCEDURES
band-pass. The excitation slit width was set at 2 nm. The
wavelength for excitation was set at 295 nm in order to minimize
the contribution of tyrosine residues to the fluorescence. Purified
HisACAT1/Flag was concentrated by using 30K Amicon centrif-
ugal filters such that the final concentration of the protein was
100-200 ng/μL (1.5-3 μM) and stored at -80 °C until usage.
Unless stated otherwise, the fluorescence measurements were per-
formed at pH 7.8 with 50 mM potassium phosphate, 0.5 M KCl,
and 8.13 mM CHAPS. Sterol (cholesterol, epicholesterol,
coprostanol, and epicoprostanol) was prepared as CHAPS/PC/
sterol mixed micelles (with CHAPS at 8.13 mM, PC at 4 mM, and
with varying concentrations of sterols as indicated) and stored at
20 °C in the dark until usage. The fatty acyl-CoAs were dissolved
in 20 mM sodium acetate at pH 6.4 as a 2 mM stock and stored at
-80 °C until usage. Before the experiment started, all reagents
were kept on ice in the dark. To begin the experiment, 50 μL of
mixed micelles without or with sterols was added into a Hellma
microcuvette (Type 105.251-QS) with 3 mm light path. Twenty-
five microliters of ACAT1 protein (in 8.13 mM CHAPS and
0.5 M KCl) was then added and rapidly mixed. Fluorescence was
monitored within 15 s of adding the enzyme. The sample com-
partment of the instrument was maintained at 20 °C. The
fluorescence emission spectrum was scanned in 2 min between
310 and 410 nm. The differences in fluorescence intensity at 10
intervals within 325-335 nm were averaged in order to calculate
the ΔF value between the ligated protein vs that of the unligated
protein.
Materials. The rabbit polyclonal antibodies (DM10, DM102)
against the N-terminal fragment (1-131) of human ACAT1 were
described previously (14). Anti-FLAG M2 antibody, anti-FLAG
M2 affinity gel, the 3ꢀ Flag peptide, CHAPS, taurocholate,
oleoyl-CoA, stearoyl-CoA, egg PC, cholesteryl oleate, cholesterol,
fatty acid-free bovine serum albumin, and protease inhibitor cock-
tails were all from Sigma. Ni-NTA His-Bind resin came from
Novagen. Fugene 6 was from Roche. The software program
PRISM 5 was from GraphPad. [3H]Cholesterol came from Perkin-
Elmer. [3H]Oleoyl-coenzyme A was synthesized as described (22).
Cell Culture. The HEK293S cell line was kindly provided
by Dr. Philip Reeves (23). Cells were grown as monolayers in
DMEM supplemented with 10% new born serum (NBS) at 37 °C
in a 5% CO2 incubator.
Construction of Plasmid HisACAT1/Flag and Its Expres-
sion in H293 Cells. To aid in enzyme purification, a Flag
octapeptide (DYKDDDDK, 1012 Da) was inserted at the
C-terminus of HisACAT1 cDNA (18), which contains a frag-
ment of 6-histidine (His) tag, and the enterokinase cleavage
recognition sequence (35 amino acids and 4057 Da) was inserted
at the N-terminus of human ACAT1 (19). This fragment
(HisACAT1/Flag) was then ligated into the mammalian expres-
sion vector pAG3-Zeo (24), using the BamHI and ApaI cloning
sites. This plasmid was transfected into HEK293S; individual
stable clones were isolated by selecting cells resistant to zeocin at
400 μg/mL.
Purification of the HisACAT1/Flag Protein. HEK293S
stably expressing HisACAT1/Flag were seeded in 55 of 145 mm
dishes for 2-3 days until the cells reached at confluency. Cells
were rinsed 2ꢀ with PBS and harvested by directly solublilizing
the cells with buffer A (1 M KCl, 2.5% CHAPS in 50 mM
KH2PO4 buffer at pH 7.8; 1 mL/dish). The detergent-solubilized
cell extracts were subjected to 100000g spin for 1 h at 4 °C; the
solubilized enzyme preparation was loaded onto a 20 mL size
His-Bind resin. The column was washed by using 4 column
volumes of 20 mM imidazole in buffer B (0.5 M KCl, 0.5%
CHAPS in 50 mM KH2PO4 buffer at pH 7.8). The fusion protein
was eluted off the column by using 2 column volumes of 500 mM
imidazole in buffer B. The eluate was then applied to a 5 mL
FLAG M2 affinity gel and allowed to flow by gravity. After
extensive buffer washes using buffer B, the HisACAT1/Flag
fusion protein was eluted off the column by 3 column volumes of
100 μg/mL FLAG peptide in buffer B. The purified fusion
protein could be stored at -80 °C for at least 6 months without
detectable loss in enzyme activity.
ACAT Enzyme Assay. To monitor enzyme activity during
enzyme purification, the enzyme solubilized in the detergent
CHAPS was assayed in preformed taurocholate/cholesterol/
phosphatidylcholine (PC) mixed micelles as described pre-
viously (25), with a final concentration of taurocholate at
9.3 mM, PC at 11.2 mM, and cholesterol at 1.6 mM. The amount
of mixed micelle solution was used to dilute the detergent presented
in the enzyme preparation, so the final CHAPS/PC molar ratio was
less than 0.4. The fatty acyl-CoA substrate saturation curves and
cholesterol substrate saturation curves were performed under the
same conditions as described previously (25).
Gel Electrophoresis and Staining. Samples were run on 8%
SDS-PAGE as described previously (25). The gels were stained
with the SilverQuest silver staining kit from Invitrogen.
RESULTS
Expression and Purification of Recombinant hACAT1
(HisACAT1/Flag) in H293 Cells. We transfected the H293
cells with the mammalian expression vector pAG3 that contained
the hACAT1 tagged with 6His at the N-terminus and Flag at the
C-terminus (designated as HisACAT1/Flag) as the insert. Various
clones that stably express the recombinant protein were then
isolated (see Experimental Procedures). Western blotting was
used to monitor the hACAT1 protein expression level. The
results show that the tagged hACAT1, with an apparent molec-
ular mass of 56 kDa in SDS-PAGE, can be amply expressed in
H293 cells (Figure 1A, rightmost two lanes). The expression level
of the HisACAT1/Flag shown in Figure 1A is representative of
several stable clones independently isolated from H293 cells.
H293 cell is a human cell line and expresses (untagged) hACAT1
endogenously (Figure 1A, the fifth and sixth lanes from the left),
with an apparent molecular mass of 50 kDa. The endogenous
hACAT1 could also be detected in the H293 cells that stably
express HisACAT1/FLAG, especially when a larger amount of
cellular protein (50 μg) was employed in the Western analysis
(Figure 1A, last lane from the left). Other results described in
Figure 1A show that the clones HisACAT1 (the first two lanes)
and the (untagged) hACAT1 (third and fourth lanes) stably
expressed in CHO cells exhibit an apparent MW of 56 or 50 kDa,
respectively, as expected (19). In a separate experiment, we moni-
tored the ACAT1 expression levels in CHO cells or in H293 cells
by using antibodies against ACAT1 (Figure 1B, top panel) or
using antibodies against Flag (Figure 1B, bottom panel). The
results show that the HisACAT1/Flag can also be detected by the
Flag antibodies as a broad, 56 kDa protein band. On the basis of
Intrinsic Fluorescence Measurement. Fluorescence mea-
surements were generated by using an ISS PC1 photocounting
fluorescence spectrophotometer (model 90095). This instrument
has continuously reproducible slits ranging from 0.4 to 32 nm

Article
Biochemistry, Vol. 49, No. 46, 2010 9959
FIGURE 1: Comparison of the tagged (HisACAT1 or HisACAT1/Flag) or the untagged (hACAT1) human ACAT1 protein levels in various
expression systems by Western blotting. (A, B) CHO cells vs H293 cells. (C) CHO cells, baculovirus infected insect H5 cells, or H293 cells.
the average of several experiments, we estimate that in H293 cells
the HisACAT1/Flag is expressed at protein levels approximately
20 times as high as the endogenous hACAT1. We had previously
reported that the relative high protein expression level of
HisACAT1 could be achieved by using recombinant baculovirus
infection of SF9 cells (26) or H5 cells (27). Here we performed
Western blotting and compared the relative ACAT1 protein
expression levels in baculovirus infected H5 cells vs that in the
H293 cells stably or transiently expressing the HisACAT1/Flag
plasmid. The results (Figure 1C) show that the His ACAT1
expression levels in baculovirus infected H5 cells and the H293 cells
stably expressing the HisACAT1/Flag are comparable (lanes 3 and
4 vs lanes 7 and 8). Additional results showed that in H293 cells the
HisACAT1/Flag expression level is higher in the stable clone than in
the cells that are transiently transfected with the HisACAT1/Flag
plasmid (Figure 1C, lanes 7 and 8 vs lanes 9 and 10).
Table 1: a
ACAT specific activity
source
(pmol min-1 mg-1
)
AC29
0.4
37.2
HEK293
hACAT1 in CHO
131.6
116.5
1199
HisACAT1 in CHO
HisACAT1/Flag in HEK293
aThe cell extracts were solubilized in 1 M KCl and 2.5% CHAPS and
assayed in cholesterol/PC/taurocholate mixed micelles.
expressing hACAT1 or expressing HisACAT1. The result also
shows that the endogenous ACAT activity in the nontransfected
H293 cells is less than 5% that of the stable clone expressing
HisACAT1/Flag.
Encouraged by the results described above, we used the
monolayers of H293 cells stably expressing HisACAT1/Flag as
the source to purify the enzyme. The purification procedure
(described in detail in Experimental Procedures) is a modification
of the procedure previously described (19). The major difference
occurs at the last purification step. Previously, we had employed
We next compared the ACAT specific enzyme activity ex-
pressed in CHO cells vs that in the H293 cells by using cell
homogenates solubilized in the detergent CHAPS as the enzyme
source. The results (Table 1) show that the enzyme activity of a
stable clone of H293 cells expressing HisACAT1/Flag is about
9-10 times higher than that of a stable clone of CHO cells either

9960 Biochemistry, Vol. 49, No. 46, 2010
Chang et al.
Table 2: Recovery of ACAT Enzyme Activity during Purificationa
whole cell extract
100%
solubilized enzyme
97
nickel column chromatography
Flag column chromatography
total recovery
21
32
7
total ACAT unit
118 nmol/min
total ACAT1 protein recovered
250 μg
a55 of 145 cm2 dishes of H293 cells stably expressing HisACAT1/Flag
were employed as the starting material.
F
IGURE 2: Assessment of the purity of HisACAT1/Flag protein by
SDS-PAGE and silver staining. Left lanes contained increasing
amounts of molecular mass standards (from Sigma); right lanes
contained increasing amounts of HisACAT1/Flag at the end of the
purification procedure.
the monoclonal antibody against hACAT1 as the affinity probe
for column chromatography and eluted the ACAT1 off the
column by using a buffer at pH 3-3.5. To avoid using the acidic
pH condition which may cause partial enzyme denaturation,
we redesigned the recombinant hACAT1 construct by placing
the Flag tag at the C-terminus of hACAT1 and employed the
monoclonal antibody against Flag as the affinity probe in the
second column chromatography step. In this manner, the enzyme
could be eluted off the column by using an eluting buffer that
contains the Flag peptide at neutral pH. Table 2 summarizes the
results of two representative experiments. The total recovery in
enzyme units is approximately 7%. The purity of the recombi-
nant hACAT1 is essentially homogeneous, as judged by SDS-
PAGE and stained with silver (Figure 2).
Intrinsic Fluorescence of hACAT1. hACAT1 contains 550
amino acids; 15 of them are tryptophan (W) and 29 are tyrosine
residues. To avoid the fluorescence contributed by tyrosine, we
set the excitation at 295 nm and monitored the emission of
fluorescence from 310 to 410 nm. Figure 3A shows the fluores-
cence emission spectrum of the purified hACAT1 dispersed in
0.5% CHAPS. The spectrum is characterized by a single peak
centered at 330 nm, due to tryptophan fluorescence of the
protein. We tested the effects of pH on ACAT1 fluorescence;
the result showed that the fluorescence intensities of the protein
gradually increased when the pH values increased from 6 to 9;
however, when the pH value increased from 9 to 10, a 24%
decrease in fluorescence intensity occurred (Figure 3A). We
noted that quenching of the protein fluorescence signal occurs
gradually. When measuring samples at 20 °C in the dark and at
various pHs as indicated (from 6 to 10), approximately 10% of
the signal is lost in 15 min (Figure 3B), presumably due to
photobleaching of the protein fluorescence. We also tested the
F
IGURE 3: Effects of pH and KCl on intrinsic fluorescence of
HisACAT1/Flag. (A) The effects of varying pH of buffer from 6 to
10 as indicated. Buffers used: 50 mM citrate for buffer 6; 50 mM
phosphate for pH 7 and 8; glycine/NaOH for pH 9 and 10. (B) The
stabilities of HisACAT1/Flag fluorescence intensity at various pH
conditions. (C) The effects of varying KCl concentrations as indi-
cated. Buffer of pH 7.8 was employed. The background fluorescence
of buffers alone (shown at the bottom) is essentially negligible.
(D) The stabilities of HisACAT1/Flag fluorescence intensity at
various KCl concentrations. The pH was at 7.8.
effect of KCl on ACAT1 fluorescence. The result showed that KCl
added at 0.25 M or higher concentrations significantly increased the
fluorescence intensities of the protein; at 1 M KCl, the intensity was
2.2-fold higher than the value in the absence of KCl (Figure 3C).
This observation correlates well with our previous observation that
KCl added at 0.5-1 M concentration increases the enzyme activity
of the detergent-solubilized ACAT1 (19). Under conditions without
or with high KCl, the fluorescence signal gradually decreases, with
8% loss in 5 min and 12% in 15 min (Figure 3D). To minimize the
quenching in fluorescence intensity, the fluorescence spectrum of
the protein was taken rapidly (within 10-15 s) after the protein was
added to the cuvette that already contains the substrate (as
described in Experimental Procedures).
Changes in Intrinsic Fluorescence of hACAT1 upon Bind-
ing to Fatty Acyl-Coenzyme A or upon Binding to Sterol.

Article
Biochemistry, Vol. 49, No. 46, 2010 9961
F
IGURE 4: Substrate saturation and binding studies of fatty acyl-
CoAs with HisACAT1/Flag as the enzyme source. (A) Oleoyl-CoA
and stearoyl-CoA substrate saturation curves. Results are represen-
tative of two separate experiments. The curves are plotted as hyper-
bolic. (B) Representative intrinsic fluorescence spectra of hACAT1 in
response to increasing concentrations of oleoyl-CoA (abbreviated
asOA). (C) Binding curvesofoleoyl-CoA and stearoyl-CoAasmoni-
tored by changes in intrinsic fluorescence of HisACAT1/Flag.
Results are the composite of two separate experiments; the curves
are plotted by using the one-site-specific binding program. The
Prism software program was used for curve plottings.
F
IGURE 5: Substrate saturation and binding studies of sterols with
HisACAT1/Flag as the enzyme source. (A) Cholesterol substrate
saturation curve. The cholesterol concentrations varied from 5 to
1600 μM, with oleoyl-CoA concentration kept at 70 μM. The curve is
plotted assigmoidal. (B) Representativeintrinsic fluorescence spectra
of HisACAT1/Flag in response to increasing concentrations of
cholesterol (CHOL). (C) Binding curve of cholesterol as monitored
by changes in intrinsic fluorescence of hACAT1; the cholesterol
concentrations varied from 0.5 to 600 μM. Results are composite of
five separate experiments. The curves are plotted by using the two-
site-specific binding program. (D) Binding curve of epicholesterol as
monitored by changes in intrinsic fluorescence of HisACAT1/Flag.
The epicholesterol concentration varied from 0.5 to 500 μM. Results
are the composite oftwo separate experiments. The curvesare plotted
by using the one-site-specific binding program. The Prism software
program was used for curve plottings.
Prior to conducting binding experiments, we first conducted
steady-state enzyme kinetic experiments using purified HisACAT1/
Flag as the enzyme source and either oleoyl- (18:1) CoA or
stearoyl- (18:0) CoA as the variable substrate, at a constant,
saturating level of cholesterol in the CHAPS/PC mixed micelles.
The result (Figure 4A) showed that the Km for oleoyl-CoA is
1.3 μM and is 6.4 μM for stearoyl-CoA; the Vmax value for oleoyl-
CoA is 2.4-fold higher than for stearoyl-CoA. These results
confirm and extend previous studies (28), demonstrating that
ACAT1 prefers oleoyl-CoA to stearoyl-CoA as the substrate. We
next conducted binding studies by monitoring the effect of
increasing concentrations of oleoyl-CoA on the intrinsic fluores-
cence of HisACAT1/Flag. The result (Figure 4B) shows that
while the peak of the spectrum remained unaltered, the peak
height is increased by approximately 32%, with respect to the
unliganded ACAT1 protein. A parallel experiment showed that
stearoyl-CoA added at increasing concentrations also caused
increases in the intrinsic fluorescence of HisACAT1/Flag. How-
ever, the maximal change in the peak of the spectrum is much
smaller (5% with respect to the unliganded protein). These results
(Figure 4C) showed that the oleoyl-CoA-induced fluorescence
change is specific and saturable. We analyzed the data in Figure 4C
for a simple bimolecular dissociation equilibrium using the Prism
program. The results show that the dissociation constant for
oleoyl-CoA is 1.9 μM. The changes caused by stearoyl-CoA were
too small to derive a reliable dissociation constant.
We had previously conducted steady-state kinetic studies and
demonstrated that cholesterol is an efficient substrate as well as
an efficient activator, while epicholesterol (which contains the
OH moiety at 3R) or enantiomeric cholesterol (which is the
mirror image of cholesterol) is neither a substrate nor an activator
for ACAT1 (21). Here we used the purified HisACAT1/Flag
as the enzyme source and conducted cholesterol substrate
saturation curve by varying the cholesterol concentration from

9962 Biochemistry, Vol. 49, No. 46, 2010
Chang et al.
activity occurred because the majority of the enzyme activity
failed to be bound to the nickel column (results not shown). In
both the CHO cell and the H293 cell expression systems, the total
yield in enzyme units at the end of the purification procedure is
approximately 7%. In the current work, we attached a Flag tag at
the C-terminus of hACAT1 and used the Flag peptide dissolved
in buffer at neutral pH to elute the ACAT1 off the affinity
column; this elution condition is much milder than the one
employed previously (19).
The expression and purification system reported here provided
us with enough pure ACAT1 protein to conduct binding studies
between the enzyme and fatty acyl-CoAs by using difference
intrinsic fluorescence spectroscopy. The results show that oleoyl-
CoA binds to ACAT1 with high affinity. The dissociation
constant for oleoyl-CoA is 1.9 μM. The Kd for oleoyl-CoA is
similar to its apparent Km (1.3 μM), suggesting that the enzyme
binds to oleoyl-CoA and reaches equilibrium relatively rapidly,
while catalysis occurs relatively slowly and constitutes the rate-
limiting step. The changes in stearoyl-CoA observed were too
small to derive a reliable dissociation constant. The Vmax for these
two substrates differed by 2.4-fold. These results suggest that
upon binding to oleoyl-CoA large structural change(s) within the
ACAT1 occur(s) to increase the catalytic efficiency of the enzyme.
We also showed that cholesterol caused significant positive
changes in the ACAT1 fluorescence spectrum in a concentration-
dependent manner. Parallel experiments show that the other
three cholesterol analogues, epicholesterol, coprostanol, and
epicoprostanol, also elicit positive spectrum changes in ACAT1,
but the magnitude of changes induced by either one of these
analogues is much smaller than cholesterol. Cholesterol and
epicholesterol share the same steroid ring structure and the same
isooctyl side chain at ring D but differ at the orientation of the
3-OH moiety at ring A. Coprostanol contains the steroid rings A
and ring B in the trans configuration (instead of a cis configura-
tion as in cholesterol). Epicoprostanol differs from coprostanol
at the orientation of the 3-OH moiety at ring A. Cholesterol and
epicholesterol interact with membrane phospholipids biophysi-
cally and affect membrane properties in very similar manners
(29, 30). Coprostanol also interacts with membrane phospho-
lipids but in a manner different from that of cholesterol (30).
In biological membranes, cholesterol, but not epicholesterol, can
directly bind to a variety of membrane proteins in a stereospecific
manner to affect their functions (29, 31-34). Consistent with
this concept, our current results suggest that sterol binding to
ACAT1 occurs in two different modes. The first is an indirect
one, mediated through its binding to phospholipids, which in
turn interacts with the transmembrane domains of ACAT1. The
second one involves the direct binding of sterol to certain residues
in ACAT1; the direct binding is governed at least in part by the
3β-OH moiety in ring A. The overall change of the ACAT1
fluorescence spectrum caused by cholesterol may be the compos-
ite of alteration caused by cholesterol’s ability to bind to ACAT1
directly and the alteration caused by cholesterol’s ability to
interact with phospholipid to affect ACAT1 structural property
indirectly. The direct, stereospecific interaction between choles-
terol and ACAT1 causes significant structural changes of the
enzyme. In the future, simultaneously monitoring the effects of
various sterols in altering the physical properties of the micelles vs
their effects in altering structural changes of ACAT1 can further
test the validity of this interpretation.
F
IGURE 6: Percent changes in intrinsic fluorescence of ACAT by
adding various sterols. Experiment was conducted as described in
Figure 5C,D. Values were averages of duplicates with deviations less
than 10% of the mean.
5 to 1600 μM, with oleoyl-CoA concentration kept at 70 μM. The
result showed that the shape of the cholesterol saturation curve
is sigmoidal, confirming our previous study when the hACAT1
without the Flag tag was used as the enzyme source (19). We next
performed the binding experiment by monitoring the effect of
increasing concentrations of cholesterol, from 10 to 600 μM, on the
intrinsic fluorescence of HisACAT1/Flag. The results (Figure 5B)
show that the peak height is positively altered by cholesterol
in a concentration-dependent manner, reaching up to 35%
with respect to unliganded ACAT1 protein. We estimate that the
cholesterol concentration that causes half-maximal spectral
change is 35 μM. We conducted a parallel experiment by using
epicholesterol as the ligand and showed that epicholesterol at
similar concentrations examined (from 0.5 to 500 μM) also
caused a positive change in the peak height of the spectrum.
However, the maximal changes in the peaks of the spectrum
caused by epicholesterol, especially at higher sterol concentra-
tions, are much smaller than those caused by cholesterol. Similar
to the result obtained when stearoyl-CoA was tested as the
ligand, the changes caused by epicholesterol were too small to
derive a reliable binding curve. We next compared the changes in
the intrinsic fluorescence of ACAT1 caused by cholesterol,
epicholesterol, coprostanol, or epicoprostanol as the ligand.
The result (Figure 6) showed that at 5 or 500 μM, the change
caused by cholesterol is much larger than the changes caused by
either one of the other three sterols.
DISCUSSION
ACAT1 and several of its homologues in the MBOAT family
are potential drug targets for treating various diseases, and it
is important to carry out structure-function analysis of these
enzymes. In the current work, we report the purification of
hACAT1 expressed in H293 cells. Based on an average of nine
experiments, we obtained 200 ( 40 μg of pure hACAT1 protein
with 120 ( 20 nmol/min enzyme units from 55 of 145 cm2 dishes
of H293 cells that stably express HisACAT1/Flag. Previously, we
purified hACAT1 from the same number of dishes of CHO cells
stably expressing the HisACAT1 to homogeneity with 4.8 nmol/
min enzyme units (19). The current method thus represents more
than 20-fold gain in enzyme units than the previous method. His
hACAT1 expressed in baculovirus infected H5 cells could also
be purified to homogeneity (27). However, the overall recovery of
enzyme activity was only approximately 1%; the large loss in
ACAT1 contains nine transmembrane domains (TMDs) (18).
Site-specific mutagenesis and disulfide cross-linking experiments

Article
Biochemistry, Vol. 49, No. 46, 2010 9963
suggested that several residues (F453, A457, H460, and F479),
located within TMD 7 or TMD 8, respectively, played critical
roles in maintaining ACAT1 enzyme activity (35). In the future,
the procedures described here will be employed to test whether
any of these residues are involved in binding to cholesterol and/or
to oleoyl-CoA. Presently, among the MBOAT enzyme family,
ACAT1 is the only one that has been purified to homogeneity.
The new information described in our current work may be
applicable to study other MBOAT members in general.
Regulation and immunolocalization of acyl-coenzyme A:cholesterol
acyltransferase in mammalian cells as studied with specific antibodies.
J. Biol. Chem. 270, 29532–29540.
15. Sakashita, N., Miyazaki, A., Takeya, M., Horiuchi, S., Chang,
C. C. Y., Chang, T. Y., and Takahashi, K. (2000) Localization of
human acyl-coenzyme A:cholesterol acyltransferase-1 in macro-
phages and in various tissues. Am. J. Pathol. 156, 227–236.
16. Khelef, N., Soe, T. T., Quehenberger, O., Beatini, N., Tabas, I., and
Maxfield, F. R. (2000) Enrichment of acyl coenzyme A:cholesterol
O-acyltransferase near trans-golgi network and endocytic recycling
compartment. Arterioscler., Thromb., Vasc. Biol. 20, 1769–1776.
17. Yu, C., Chen, J., Lin, S., Liu, J., Chang, C. C., and Chang, T. Y. (1999)
Human acyl-CoA:cholesterol acyltransferase-1 is a homotetrameric
enzyme in intact cells and in vitro. J. Biol. Chem. 274, 36139–36145.
18. Guo, Z. Y., Lin, S., Heinen, J. A., Chang, C. C., and Chang, T. Y.
(2005) The active site His-460 of human acyl-coenzyme A:cholesterol
acyltransferase 1 resides in a hitherto undisclosed transmembrane
domain. J. Biol. Chem. 280, 37814–37826.
ACKNOWLEDGMENT
We thank Dr. Gustav E. Lienhard for careful reading of the
manuscript. We thank Dean R. Madden and Larry C. Myers at
Dartmouth Medical School and members of the Chang labora-
tory for helpful discussions during the course of this work. We
thank Oneil N. Gardner for participating in the enzyme purifica-
tion work as an intern of the SURF program.
19. Chang, C. C. Y., Lee, C. Y. G., Chang, E. T., Cruz, J. C., Levesque,
M. C., and Chang, T. Y. (1998) Recombinant human acyl-CoA:
cholesterol acyltransferase 1 (ACAT1) purified to essential homo-
geneity utilizes cholesterol in mixed micelles or vesicles in a highly
cooperative manner. J. Biol. Chem. 273, 35132–35141.
20. Zhang, Y., Yu, C., Liu, J., Spencer, T. A., Chang, C. C., and Chang,
T. Y. (2003) Cholesterol is superior to 7-ketocholesterol or 7 alpha-
hydroxycholesterol as an allosteric activator for acyl-coenzyme A:
cholesterol acyltransferase 1. J. Biol. Chem. 278, 11642–11647.
21. Liu, J., Chang, C. C., Westover, E. J., Covey, D. F., and Chang, T. Y.
(2005) Investigating the allosterism of acyl coenzyme A:cholesterol
acyltransferase (ACAT) by using various sterols: in vitro and intact
cell studies. Biochem. J. 391, 389–397.
REFERENCES
1. Chang, C. C., Huh, H. Y., Cadigan, K. M., and Chang, T. Y. (1993)
Molecular cloning and functional expression of human acyl-coenzyme A:
cholesterol acyltransferase cDNA in mutant Chinese hamster ovary cells.
J. Biol. Chem. 268, 20747–20755.
2. Cases, S., Novak, S., Zheng, Y. W., Myers, H. M., Lear, S. R., Sande,
E., Welch, C. B., Lusis, A. J., Spencer, T. A., Krause, B. R., Erickson,
S. K., and Farese, R. V., Jr. (1998) ACAT-2, a second mammalian
acyl-CoA:cholesterol acyltransferase. Its cloning, expression, and
characterization. J. Biol. Chem. 273, 26755–26764.
22. Bishop, J. E., and Hajra, A. K. (1980) A method for the chemical
synthesis of 14C-labeled fatty acylcoenzyme A’s of high specific
activity. Anal. Biochem. 106, 344–350.
23. Reeves, P. J., Kim, J. M., and Khorana, H. G. (2002) Structure
and function in rhodopsin: a tetracycline-inducible system in stable
mammalian cell lines for high-level expression of opsin mutants. Proc.
Natl. Acad. Sci. U.S.A. 99, 13413–13418.
3. Anderson, R. A., Joyce, C., Davis, M., Reagan, J. W., Clark, M.,
Shelness, G. S., and Rudel, L. L. (1998) Identification of a form of
acyl-CoA:cholesterol acyltransferase specific to liver and intestine in
nonhuman primates. J. Biol. Chem. 273, 26747–26754.
24. Ikeuchi, T., Dolios, G., Kim, S. H., Wang, R., and Sisodia, S. S. (2003)
Familial Alzheimer disease-linked presenilin 1 variants enhance pro-
duction of both Abeta 1-40 and Abeta 1-42 peptides that are only
partially sensitive to a potent aspartyl protease transition state
inhibitor of “gamma-secretase”. J. Biol. Chem. 278, 7010–7018.
25. Chang, C. C., Lee, C. Y., Chang, E. T., Cruz, J. C., Levesque, M. C.,
and Chang, T. Y. (1998) Recombinant acyl-CoA:cholesterol acyl-
transferase-1 (ACAT-1) purified to essential homogeneity utilizes
cholesterol in mixed micelles or in vesicles in a highly cooperative
manner. J. Biol. Chem. 273, 35132–35141.
4. Oelkers, P., Behari, A., Cromley, D., Billheimer, J. T., and Sturley,
S. L. (1998) Characterization of two human genes encoding acyl
coenzyme A:cholesterol acyltransferase-related enzymes. J. Biol.
Chem. 273, 26765–26771.
5. Chang, T. Y., Li, B. L., Chang, C. C., and Urano, Y. (2009)
Acyl-coenzyme A:cholesterol acyltransferases. Am. J. Physiol. 297, E1–E9.
6. Das, A., Davis, M. A., Tomoda, H., Omura, S., and Rudel, L. L. (2008)
Identification of the interaction site within acyl-CoA:cholesterol acyl-
transferase 2 for the isoform-specific inhibitor pyripyropene A. J. Biol.
Chem. 283, 10453–10460.
26. Cheng, D., Chang, C. C., Qu, X., and Chang, T. Y. (1995) Activation
of acyl-coenzyme A:cholesterol acyltransferase by cholesterol or by
oxysterol in a cell-free system. J. Biol. Chem. 270, 685–695.
27. Yu, C., Zhang, Y., Lu, X., Chen, J., Chang, C. C., and Chang, T. Y.
(2002) Role of the N-terminal hydrophilic domain of acyl-coenzyme
A:cholesterol acyltransferase 1 on the enzyme’s quaternary structure
and catalytic efficiency. Biochemistry 41, 3762–3769.
28. Seo, T., Oelkers, P. M., Giattina, M. R., Worgall, T. S., Sturley, S. L.,
and Deckelbaum, R. J. (2001) Differential modulation of ACAT1 and
ACAT2 transcription and activity by long chain free fatty acids in
cultured cells. Biochemistry 40, 4756–4762.
29. Gimpl, G., Burger, K., and Fahrenholz, F. (1997) Cholesterol as
modulator of receptor function. Biochemistry 36, 10959–10974.
30. Xu, X., and London, E. (2000) The effect of sterol structure on
membrane lipid domains reveals how cholesterol can induce lipid
domain formation. Biochemistry 39, 843–849.
31. Romanenko, V. G., Rothblat, G. H., and Levitan, I. (2002) Modula-
tion of endothelial inward-rectifier Kþ current by optical isomers of
cholesterol. Biophys. J. 83, 3211–3222.
32. Simons, K., and Ikonen, E. (2000) How cells handle cholesterol.
Science 290, 1721–1726.
33. Westover, E. J., and Covey, D. F. (2004) The enantiomer of choles-
terol. J. Membr. Biol. 202, 61–72.
34. Gimpl, G. (2010) Cholesterol-protein interaction: methods and cho-
lesterol reporter molecules. Subcell. Biochem. 51, 1–45.
35. Guo, Z. Y., Chang, C. C., and Chang, T. Y. (2007) Functionality of
the seventh and eighth transmembrane domains of acyl-coenzyme A:
cholesterol acyltransferase 1. Biochemistry 46, 10063–10071.
7. Ikenoya, M., Yoshinaka, Y., Kobayashi, H., Kawamine, K., Shibuya,
K., Sato, F., Sawanobori, K., Watanabe, T., and Miyazaki, A. (2007)
A selective ACAT-1 inhibitor, K-604, suppresses fatty streak lesions
in fat-fed hamsters without affecting plasma cholesterol levels.
Atherosclerosis 191, 290–297.
8. Cases, S., Smith, S. J., Zheng, Y. W., Myers, H. M., Lear, S. R., Sande,
E., Novak, S., Collins, C., Welch, C. B., Lusis, A. J., Erickson, S. K.,
and Farese, R. V., Jr. (1998) Identification of a gene encoding an acyl
CoA:diacylglycerol acyltransferase, a key enzyme in triacylglycerol
synthesis. Proc. Natl. Acad. Sci. U.S.A. 95, 13018–13023.
9. Hofmann, K. (2000) A superfamily of membrane-bound O-acyltrans-
ferases with implications for Wnt signaling. Trends Biochem. Sci. 25,
111–112.
10. Yang, J., Brown, M. S., Liang, G., Grishin, N. V., and Goldstein, J. L.
(2008) Identification of the acyltransferase that octanoylates ghrelin,
an appetite-stimulating peptide hormone. Cell 132, 387–396.
11. Gutierrez, J. A., Solenberg, P. J., Perkins, D. R., Willency, J. A.,
Knierman, M. D., Jin, Z., Witcher, D. R., Luo, S., Onyia, J. E., and
Hale, J. E. (2008) Ghrelin octanoylation mediated by an orphan lipid
transferase. Proc. Natl. Acad. Sci. U.S.A. 105, 6320–6325.
12. Riekhof, W. R., Wu, J., Jones, J. L., and Voelker, D. R. (2007)
Identification and characterization of the major lysophosphatidyl-
ethanolamine acyltransferase in Saccharomyces cerevisiae. J. Biol.
Chem. 282, 28344–28352.
13. Shindou, H., and Shimizu, T. (2009) Acyl-CoA:lysophospholipid
acyltransferases. J. Biol. Chem. 284, 1–5.
14. Chang, C. C. Y., Chen, J., Thomas, M. A., Cheng, D., Del Priore,
V. A., Newton, R. S., Pape, M. E., and Chang, T. Y. (1995)