|
|
|
|||||||
|
|||||||||
(Received for publication, September 29, 1994) From the
Apolipoprotein A-I (apoA-I), the major protein of high density
lipoproteins, facilitates reverse cholesterol transport from peripheral
tissue to liver. To determine the structural motifs important for
modulating the in vivo catabolism of human apoA-I (h-apoA-I),
we generated carboxyl-terminal truncation mutants at residues 201
(apoA-I
The concentration of plasma high density lipoproteins (HDL) ( ApoA-I is synthesized as a
prepropeptide and is cotranslationally cleaved to proapoA-I, which then
is secreted and endoproteolytically cleaved to form the mature 243
amino acid apoA-I protein(10) . The primary amino acid sequence
of apoA-I (11, 12) has been suggested to contain
highly conserved 11-amino acid repeats, which form amphipathic
Analysis of patients with reduced plasma concentration of HDL has
revealed that accelerated catabolism of apoA-I is the most common cause
of low HDL levels(22, 23, 24, 25) .
The mechanism, however, for the increased clearance of apoA-I in these
various studies has not been delineated. Proteolysis is known to
regulate the clearance of many other plasma
proteins(26, 27, 28, 29, 30, 31) .
The amino-terminal endoproteolytic cleavage of the propeptide, however,
appears not to be important in modifying the plasma clearance of
apoA-I(32) . Although the aminoterminal domain of the mature
form of apoA-I can undergo limited proteolysis(33) , the
carboxyl-terminal domain has been shown to be particularly sensitive to
proteolysis(20, 34, 35, 36, 37, 38, 39, 40, 41) .
These in vitro studies suggest that proteolysis of apoA-I may
possibly occur in vivo. Preliminary data of a 26-kDa
proteolyzed fragment of apoA-I with an intact amino-terminal domain
showed rapid in vivo catabolism (42) . Kunitake et
al.(34) reported on the occurrence of a 26-kDa and a
14-kDa proteolytic fragment isolated from pre- To address the role of
structural motifs of apoA-I in its metabolic clearance, we have
produced carboxyl-terminal truncation mutants of apoA-I and analyzed
their metabolic clearance in vivo. These studies establish the
importance of the carboxyl-terminal domain of apoA-I in its association
with lipoprotein particles and catabolism.
Figure 1:
Construction of the expression vector
pMal-c/A-I. Recombinant apoA-I cDNA and truncation mutants were fused
upstream to the maltose binding protein coding gene (malE
The predicted secondary structure of apoA-I based on the Chou
and Fasman analysis(15, 16, 21, 48, 49, 50, 51) is
illustrated in Fig. 2. The locations of the carboxyl-terminal
truncations of apoA-I are indicated by arrows. ApoA-I
Figure 2:
Schematic representation of the secondary
structure of apoA-I. ApoA-I is composed of nine amphipathic
Intact apoA-I
and the truncation mutants were expressed in E. coli as a
fusion protein with the maltose-binding protein and an unique factor Xa
cleavage site between the two protein domains. This expression system
produced high levels of the fusion protein (30-70 mg/liter), but
digestion of the fusion protein with factor Xa in order to produce free
apoA-I resulted in nonspecific degradation of apoA-I. Interestingly,
apoA-I The purified recombinant intact apoA-I and the
carboxyl-terminal truncation mutants were characterized by
polyacrylamide gel electrophoresis, isoelectric focusing, and amino
acid analysis. In addition, the PCR-amplified cDNA constructs were
confirmed by DNA sequencing. Each of the apoA-I forms migrated as a
single band (Fig. 3) with the expected molecular masses on
polyacrylamide gel electrophoresis (r-apoA-I, 28 kDa; apoA-I
Figure 3:
SDS-gel electrophoresis of apoA-I. Human
and r-apoA-I were isolated, radioiodinated, and electrophoresed on a
10-20% SDS acrylamide gel: lane 1, h-apoA-I; lane
2, r-apoA-I; lane 3, apoA-I
Figure 4:
Isoelectric focussing of apoA-I. Purified
h-apoA-I (lane 1) and r-apoA-I (lane 2) were
electrophoresed by isoelectric focussing (pH 5-7) and stained for
protein. The major isoforms of apoA-I
The metabolic clearance of radiolabeled r-apoA-I and
h-apoA-I was analyzed in rabbits (Fig. 5). The mean plasma decay
curves of injected radiolabeled h-apoA-I (n = 6) and
r-apoA-I (n = 6) were similar (Fig. 5A), demonstrating that r-apoA-I expressed in E. coli has similar in vivo kinetic behavior as
h-apoA-I. The FCR using three-exponential equations with the SAAM31
program for r-apoA-I and h-apoA-I were 0.93 ± 0.07/day (mean
± S.D.) and 0.91 ± 0.34/day, respectively. Similar FCR
values were previously reported for r-apoA-I and h-apoA-I in
normolipemic male Japanese White rabbits(52) . Fig. 5, B-D, depicts the plasma decay curves of the three
radiolabeled truncation mutants. All truncated apoA-I mutants had a
markedly increased rate of catabolism when compared with normal apoA-I (Fig. 5A). Interestingly, the apoA-I
Figure 5:
In vivo metabolism of apoA-I.
Proteins were labeled with either
Since the disruption of the
carboxyl-terminal apoA-I altered its plasma turnover, we examined the
association of these various mutants with lipoprotein particles. The
isolated apolipoproteins were radiolabeled, incubated in human plasma
for 60 min at 37 °C, and analyzed by density gradient
ultracentrifugation. The density distributions of h-apoA-I, r-apoA-I,
apoA-I
Figure 6:
Density gradient ultracentrifugation of
apoA-I in vitro. Radioiodinated proteins were incubated in
human plasma for 60 min at 37 °C. Continuous density gradient
ultracentrifugation was performed, and radioactivity of density
fractions was determined for h-apoA-I (
We further investigated the
lipoprotein particle binding properties of the apoA-I mutants by
examining the density distribution of radiolabeled h-apoA-I, r-apoA-I,
and apoA-I
Figure 7:
Density gradient ultracentrifugation of
apoA-I in vivo. Radioiodinated h-apoA-I (top panel),
r-apoA-I (middle panel), and apoA-I
These differences in distribution within
particles defined ultracentrifugally are paralleled by differences in
the particle distribution observed by electrophoretic separation (Fig. 8). Fasting human plasma has Fat 7B-stained lipoprotein
particles of
Figure 8:
Agarose gel electrophoresis. Plasma was
subjected to agarose gel electrophoresis. Human (lane 1) and
rabbit plasma incubated with
We also analyzed the lipoprotein particle compositions of
New Zealand White rabbit plasma using FPLC. Fig. 9A
illustrates the analyses of total cholesterol, cholesteryl ester,
phospholipids, and triglycerides for each fraction. The volumes
18-22 ml represent rabbit low density lipoprotein particle
(LDL), whereas 26-32 ml represents HDL and 32-36 ml of
phospholipid-rich HDL. The comparison of these HDL fractions isolated
by gel filtration chromatography with the HDL
Figure 9:
Gel filtration chromatography. Composition
analysis of New Zealand White rabbit plasma is illustrated in Fig. 9A, showing the concentrations of total
cholesterol, cholesterol ester, phospholipids, and triglycerides in
mg/dl. Radioiodinated apoA-I
The concentration of plasma apoA-I, which is inversely
correlated with the risk for CAD(1, 2, 3) ,
is determined by its rate of production and
catabolism(55, 56) . The mechanism for the catabolism
of apoA-I has not been completely delineated. We were interested in
examining the effect of carboxyl-terminal proteolysis on the metabolic
clearance of apoA-I, since this is a possible site for in vivo proteolysis(20, 34, 35, 36, 37, 38, 39, 40, 41, 42) . In order to test the functional role of the carboxyl-terminal domain
of apoA-I, we produced recombinant apoA-I containing various
carboxyl-terminal deletions. The recombinant forms of apoA-I had the
predicted size, charge, and amino acid composition as assessed by
polyacrylamide gel electrophoresis (Fig. 3), isoelectric
focusing (Fig. 4), and amino acid analysis (Table 1). The
amino-terminal amino acid sequence of r-apoA-I was the same as h-apoA-I
purified from plasma. The in vivo clearance rate of r-apoA-I
was nearly identical to the clearance of h-apoA-I (r-apoA-I, FCR
= 0.93 ± 0.07/day; h-apoA-I, FCR = 0.91 ±
0.34/day), thus validating the use of the recombinant truncation
mutants in exploring the role of the carboxyl-terminal domain in the
metabolic clearance of apoA-I. Similar results have been reported in a
study of the in vivo conversion of recombinant proapoA-I to
apoA-I in normolipemic male Japanese White rabbits (h-apoA-I, 0.98
± 0.03/day; r-apoA-I, FCR = 1.02 ± 0.09/day) (52) . In addition to the intact r-apoA-I, we generated three
truncation mutants (Fig. 2). ApoA-I It is known that radioiodination can interfere
with the molecular properties of apolipoproteins as
apoA-I(61, 62) ; however, we used a standardized
radiolabeling method, that is reproducible and has been shown to yield
kinetics of the radiolabeled protein that are almost identical to the
kinetics obtained by endogenously labeled stable isotope
technique(63) . In addition, to the amount of radioactivity,
the concentration of the protein used for labeling, the ratio of the
iodinated versus the non-iodinated protein and the handling of
the iodinated proteins were the same for all of the labeled forms of
apoA-I in this study. The proteins were injected within less than 24 h
after radiolabeling and analyzed by polyacrylamide gel electrophoresis
to exclude self-association (64) (Fig. 3). The different
behavior of the truncation mutants versus h-apoA-I and
r-apoA-I is, therefore, not due to the radiolabeling procedure. Since the affinity of apolipoproteins for lipoprotein particles has
been shown to inversely correlate with their metabolic
clearance(65, 66) , we examined both the in vitro and in vivo lipoprotein association of the truncation
mutants for lipid. After incubating the purified apoA-I forms in
plasma, h-apoA-I and r-apoA-I associated with HDL The major sites of apoA-I degradation are
the kidney and liver(67) . It has been proposed that once
apoA-I dissociates from HDL, it is quickly removed from the circulation
by filtration through the kidney(68, 69) . Renal
tubular cells have been shown to uptake apoA-I from urine(68) .
We demonstrate that intact apoA-I does not appear within the urine,
which supports the tissue uptake of apoA-I and liberation of the iodide
radiotracers. Horowitz (70) demonstrated an increased plasma
and renal clearance of an exchangeable pool of apoA-I in individuals
with low plasma HDL concentrations. The authors suggested that these
subjects have an increased cholesteryl ester transfer protein-mediated
exchange of HDL cholesteryl ester for very low density lipoprotein
triglyceride, resulting in a triglyceride-enriched HDL. Human
triglyceride-rich HDL ApoA-I is synthesized as a
prepropeptide and normally undergoes several proteolytic events before
mature apoA-I is produced(32) . The prepeptide signal sequence
is cleaved cotranslationally in the endoplasmic reticulum(74) .
The amino-terminal propeptide is then cleaved by an endoprotease
rapidly after secretion(74, 75) . The rate of
conversion of proapoA-I to apoA-I is much faster than the clearance of
mature apoA-I(32) . The in vivo proteolytic cleavage
of apoA-I at other sites has not been fully demonstrated, but it has
been shown that apoA-I, particularly apoA-I on pre- In summary, our results are consistent with the hypothesis that
carboxyl-terminal proteolysis of apoA-I leads to rapid catabolism. The
smallest deletion of 17 residues, which disrupts the last amphipathic
Volume 270,
Number 10,
Issue of March 10, 1995 pp. 5469-5475
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES
), 217 (apoA-I
), and 226
(apoA-I
) by site-directed mutagenesis. ApoA-I was
expressed in Escherichia coli as a fusion protein with the
maltose binding protein, which was removed by factor Xa cleavage. The in vivo kinetic analysis of the radioiodinated apoA-I in
normolipemic rabbits revealed a markedly increased rate of catabolism
for the truncated forms of apoA-I. The fractional catabolic rates (FCR)
of 9.10 ± 1.28/day (±S.D.) for apoA-I, 6.34
± 0.81/day for apoA-I, and 4.42 ± 0.51/day
for apoA-I were much faster than the FCR of recombinant
intact apoA-I (r-apoA-I, 0.93 ± 0.07/day) and h-apoA-I (0.91
± 0.34/day). All the truncated forms of apoA-I were associated
with very high density lipoproteins, whereas the intact recombinant
apoA-I (r-apoA-I) and h-apoA-I associated with HDL
and
HDL
. Gel filtration chromatography revealed that in
contrast to r-apoA-I, the mutant apoA-I associated with a
phospholipid-rich rabbit apoA-I containing particle. Analysis by
agarose gel electrophoresis demonstrated that the same mutant migrated
in the pre-
position, but not within the 
position as did
r-apoA-I. These results indicate that the carboxylterminal region
(residue 227-243) of apoA-I is critical in modulating the
association of apoA-I with lipoproteins and in vivo metabolism
of apoA-I.
)is inversely related to the risk of developing human
cardiovascular disease(1, 2, 3) . ApoA-I is
the major structural protein of HDL and an important determinant of the
concentration of HDL in plasma(4) . ApoA-I binds and transports
plasma lipid, serves as cofactor for the enzyme lecithincholesterol
acyltransferase, and increases cholesterol efflux from peripheral
tissues(5, 6, 7) . In addition, apoA-I has
been reported to be an important ligand in the binding of HDL to cell
membranes(8, 9) . All these properties are important
in the ability of apoA-I to facilitate reverse cholesterol transport,
the mechanism that has been proposed to account for the protective
effect of HDL on cardiovascular disease.-helices that interact with the lipid surface of lipoprotein
particles(13, 14, 15, 16, 17, 18) .
Structural analysis revealed that the carboxyl-terminal region of
apoA-I plays an important role in lipid binding and in the interaction
of HDL with cell
membranes(9, 19, 20, 21) . HDL, which was
reduced by the presence of protease inhibitors. The amino-terminal
sequence of both peptides were intact, suggesting a carboxyl-terminal in vivo proteolysis of apoA-I.
Construction of the Expression Vector
pMAL-c/A-I
PCR was utilized to introduce appropriate DNA
sequences for constructing a fusion protein of apoA-I to the maltose
binding protein. The following 5` and 3` primers were used in the
study: 1)
5`-GCCGCCGCCGGATCCAACAACAACAACAACAACAACAACAACAACATCGAGGGTCGTGATGAACCCCCCCAGAGCCCCTGGGATCGAGTG-3`;
2) 5`-GGCGGCGGCAAGCTTTCACTGGGTGTTGAGCTTCTTAGTGTACTC-3`. The 5`
primer contained a BamHI site (boldface), a poly-Asn domain,
and a factor Xa cleavage site (underlined), which was immediately
followed by the codons for the first 11 amino acids of mature apoA-I.
The 3` primer has a HindIII site (boldface), a stop codon
(underlined), and the codons for the last 9 carboxyl-terminal amino
acids of apoA-I. The pLen-A-I plasmid(43) , which contains a
full-length cDNA sequence of apoA-I, was used as a template for PCR. A
PCR fragment of 0.8 kilobase pair was digested with BamHI and HindIII and ligated into the unique BamHI and HindIII sites of pMAL-c2 plasmid (New England Biolabs,
Beverly, MA) to produce the plasmid designated pMAL-c/A-I (Fig. 1). In addition, pMAL-c/A-I vector was modified to create
three carboxyl-terminal truncated apoA-I variants by introducing stop
codons at residues 201 (Ser), 217 (Gly), and 226 (Lys) by using PCR.
These apoA-I truncation mutants are designated apoA-I,
apoA-I, and apoA-I. The 5` end primer for
the PCR reactions was the same primer used for the full-length
construct of apoA-I. The following are the 3` primers that were used
for the truncation mutants: 1)
5`-GGCGGCGGCAAGCTTTCAGCTCAGATGCTCGGTGGCCTTGGCGTG-3` for
apoA-I, 2)
5`GGCGGCGGCAAGCTTTCAGCCTTGGCGGAGGTCCTCGAGCGCGGG-3` for
apoA-I, and 3)
5`-GGCGGCGGCAAGCTTTCACTTGAAGCTCTCCAGCACGGGCAGCAG-3` for
apoA-I. All three primers contain a HindIII site
(boldface), an in-frame stop codon (underlined), and the codons for the
9 carboxyl-terminal amino acids of each truncated apoA-I variant. The
DNA sequence of all constructs was confirmed by the dideoxy chain
termination method, using Sequenase T7 DNA polymerase (version 2, U. S.
Biochemical Corp.).
). Expression of the fusion
protein was driven by the
isopropyl-1-thio--D-galactopyranoside-inducible tac promoter (P
).
Expression and Purification of Recombinant
ApoA-I
DH5 Escherichia coli cells containing
pMal-c/A-I were grown in LB-broth containing ampicillin (100 µg/ml)
at 37 °C and were induced with 0.3 mM isopropyl-1-thio--D-galactopyranoside. Bacteria were
harvested by centrifugation at 4000 
g at 4 °C for
15 min, and the cell pellet was resuspended in 10 mM Tris (pH
7.4), 200 mM NaCl, 1 mM EDTA, 1 mM azide, 1
mM dithiothreitol, 0.5 mM phenylmethylsulfonyl
fluoride, 0.2 mg/ml lysozyme, 2.5 µg/ml of leupeptin and aprotinin.
After three cycles of freeze-thawing, the bacterial suspension was
sonicated by short pulses of 10 s or less on ice (50 watts), and
centrifuged at 9000 g at 4 °C for 30 min. The
supernatant was filtered (0.2 µm, Millipore) and affinity-purified
with amylose resin (New England Biolabs). The fusion protein was
dialyzed against 20 mM NH
HCO,
lyophilized, and redissolved in 20 mM Tris (pH 7.4), 100
mM NaCl, 2 mM CaCl, and 1 mM azide (pH 8.0). Digestion with factor Xa (New England Biolabs) was
performed using 10 mg/ml fusion protein and 0.05% factor Xa by weight
at 4 °C for 72 h, and the digestions were terminated by adding 50
mM phenylmethylsulfonyl fluoride. The digested fusion protein
was separated by electrophoresis on a 10-20% Tricine
polyacrylamide gel containing SDS. ApoA-I was isolated from unstained
gel slices, utilizing stained lanes of purified apoA-I as standard.
Following electroelution (Bio-Rad) in 20 mM NHHCO, apoA-I was lyophilized and
redissolved in distilled H0. Standard apoA-I was isolated
from normolipemic human plasma by column chromatography, as described
previously(11) . The purity of apoA-I samples were established
by SDS-gel electrophoresis, immunoblotting, and isoelectric focussing,
as described previously(11) .In Vivo Metabolic Studies
ApoA-I was iodinated in
1 M glycine, 0.1 M Tris (pH 8.5) by the iodine
monochloride method(44) , using either 
I or 
I as the radiotracer. Unbound iodine was removed by
extensive dialysis against 0.85% (w/w) NaCl, 0.01% (w/w) EDTA, 0.1 M Tris (pH 7.4). All preparations were filter-sterilized
(0.22-µm filters, Millipore) prior to injection and were used
within 48 h of radiolabeling. New Zealand White female rabbits,
2.2-2.8 kg, were injected simultaneously with I and I radiolabeled apoA-I or apoA-I truncation mutants via
the medial artery of the ear. Blood samples were collected from a
catheter of the external jugular vein at 5 min and up to 5 days after
injection and analyzed for radioactivity on a Packard Cobra 
counter (Packard Instrument Co., Downers Grove, IL). Residence times
were obtained from the area under the plasma radioactivity decay
curves, using a multiexponential computer curve-fitting
technique(45) . Fractional catabolic rates (FCR) represent the
reciprocal of residence times. Continuous density gradient
ultracentrifugation was performed as described previously(46) .
HDL density subfractions were defined the same as in humans
(HDL, d = 1.063-1.125 g/ml;
HDL, d = 1.125-1.210 g/ml; and VHDL, d = 1.21-1.25 g/ml, respectively).Amino Acid Analysis
Approximately 10 µg of the
purified proteins were electrophoresed on a SDS-polyacrylamide gel and
transferred to polyvinylidene difluoride membranes (Millipore). The
membrane pieces were subjected to NH-terminal sequence
analysis on an Applied Biosystems model 610A sequencer. Amino acid
analysis were performed after hydrolysis in 6 M HCl for 24 h
on a Beckman model 6300 amino acid analyzer. Results from glycine
determinations were excluded because of high background.Gel Filtration Chromatography
One-hundred µl
of rabbit plasma were analyzed by FPLC on two Superose-6 columns
connected in series (Pharmacia, Uppsala, Sweden). Lipoproteins were
eluted at 0.3 ml/min with phosphate-buffered saline, containing 1
mM EDTA and 0.02% (w/w) sodium azide(47) . After the
first 10 ml were collected, the next 50 ml were collected in 0.5-ml
fractions. Total cholesterol and triglyceride concentrations were
determined enzymatically (Sigma). Free cholesterol and phospholipid
concentrations were determined enzymatically (Wako Chemicals USA Inc.;
Richmond, VA).
is truncated within the carboxyl-terminal amphipathic
-helix. ApoA-I is truncated in the region between
the eighth and ninth helix, whereas apoA-I is truncated
within the eighth helix. These sites were chosen to span the region
that has previously been shown to undergo carboxyl-terminal
proteolysis(20, 34, 35) .
-helices (illustrated as rectangles) and two -sheets (zigzag lines). The locations of the three carboxyl-terminal
deletion mutants at residues 226 (Lys), 217 (Gly), and 201 (Ser) are
indicated by arrows.
was much less susceptible to degradation compared
to the other truncation mutants and r-apoA-I. The final yield of free
apoA-I was 10 µg/l of purified apoA-I, and
0.1-0.5 µg/l for apoA-I, apoA-I,
and r-apoA-I. 26 kDa; apoA-I, 25 kDa; and apoA-I,
23 kDa). The identity of each band was confirmed by immunoblot
analysis, using a polyclonal rabbit anti-human apoA-I antibody (data
not shown). Fig. 4shows the similar isoelectric focussing
pattern of the two major isoforms of h-apoA-I and r-apoA-I, indicating
that no significant charge modification of the recombinant apoA-I
occurred during either the biosynthesis by the bacteria or during the
protein purification. In Table 1, the predicted versus the observed amino acid composition of the purified apoA-I forms
expressed as mole percent is shown. These data confirm the location of
the apoA-I truncation and demonstrate the purity of the recombinant
apoA-I forms. Amino-terminal sequence analysis of all the isolated
apoA-I proteins revealed intact NH-terminal domains, thus
demonstrating the specificity of the factor Xa cleavage (data not
shown).
; lane
4, apoA-I; and lane 5,
apoA-I. The indicated molecular mass markers are
phosphorylase b (97.4 kDa), bovine serum albumin (69 kDa),
ovalbumin (46 kDa), carbonic anhydrase (30 kDa), and trypsin inhibitor
(21.5 kDa).
and apoA-I
are indicated.
mutant, which is missing only the carboxyl-terminal 17 amino
acids, was rapidly catabolized, with an approximately five times higher
FCR (FCR = 4.42 ± 0.51/day) than intact apoA-I. The FCR
for apoA-I (FCR = 6.34 ± 0.81/day) and
apoA-I (FCR = 9.10 ± 1.28/day) were even
faster, suggesting that the more the carboxyl-terminal domain was
deleted, the greater its catabolic rate. In addition, we evaluated the
urine radioactivity during the first 24 h after the injection of the
radiolabeled r-apoA-I and apoA-I. By monitoring
radioactivity, we observed a urine/plasma ratio of 0.36 for r-apoA-I versus 5.82 for apoA-I during the 24-h
collection period. By trichloroacetic acid precipitation and gel
electrophoresis, we estimate that greater than 95% of the total counts
were not protein bound. These results support the increase metabolic
turnover of the truncated apoA-I forms.
I or I
and the in vivo metabolism of radiolabeled apoA-I forms was
analyzed as described under ``Experimental Procedures.'' In A, the radioactivity decay curves for human (
) and
recombinant (
) apoA-I are illustrated using a two-log scale as
the ordinate. Each time point represents the mean
(±S.D.) for residual radioactivity from six independent
experiments. S.D. of human apoA-I is indicated with upper error
bars, and S.D. of recombinant apoA-I is indicated with lower
error bars. B-D represent the radioactive decay
curves for apoA-I (n = 3), apoA-I (n = 3), and apoA-I (n = 3), respectively, using a two-log scale as the ordinate.
, apoA-I, and apoA-I are indicated in Fig. 6. Native apoA-I and r-apoA-I were
primarily associated with HDL (d =
1.063-1.125 g/ml) and HDL (d =
1.125-1.210 g/ml) with their peak density at 1.13 g/ml. The
truncation mutants were almost exclusively associated with VHDL (d = 1.21-1.25), with a peak density at 1.23 g/ml.
Therefore, the loss of the carboxyl-terminal amphipathic helices of
apoA-I, in addition to altering its plasma clearance, also altered its
association with the HDL subfractions.
), r-apoA-I (+),
apoA-I (), apoA-I (
), and
apoA-I (
). The density fractions corresponding to
very low density lipoproteins (VLDL, d < 1.006
g/ml), low density lipoproteins (LDL, d =
1.006-1.063 g/ml), high density lipoproteins (HDL, d = 1.063-1.125; HDL, d = 1.125-1.210; VHDL, d = 1.21-1.25 g/ml), and
lipoprotein-deficient serum (LPDS, d > 1.25) are indicated
by arrows.
following injection in vivo (Fig. 7). The density distributions of h-apoA-I (Fig. 7A) and r-apoA-I (Fig. 7B) were
similar and were detected in both HDL and HDL density fractions. In contrast, apoA-I was found
primarily within VHDL (Fig. 7C), as we observed in
vitro (Fig. 6).
(lower
panel) were injected into New Zealand White rabbits. Plasma
obtained after 20 min (solid rectangle), 8 h (gray
rectangle), and 24 h (open rectangle) were separated into
the density fractions indicated by the arrows by continuous
density gradient ultracentrifugation and analyzed for
radioactivity.
and mobility (Fig. 8, lane 1).
In contrast, a pre- mobility is present in rabbits (Fig. 8, lanes 2 and 3). The radiolabeled apoA-I was present in particles with a slower pre- migration (Fig. 8, lane 5) than the mobility observed for
r-apoA-I.
I-r-apoA-I (lanes 2 and 4) and apoA-I (lanes 3 and 5) were analyzed by staining with Fat Red 7B (lanes
1-3) and autoradiography (lanes 4 and 5).
ApoA-I and r-apoA-I were analyzed by gel electrophoresis
using a buffered agarose system (Ciba Corning Diagnostics Corp., Palo
Alto, CA) and a Fat Red 7B stain.
,
HDL, and VHDL isolated by sequential ultracentrifugation (Fig. 6), suggest that HDL and HDL elute
with 26-32 ml, whereas VHDL elutes with 32-36 ml. We also
characterized the lipoprotein-associated radioiodinated r-apoA-I and
apoA-I after incubation in rabbit plasma by gel
filtration chromatography. The highest radioactivity was in the
fractions of the elution volume of 28-30 and 32-35 ml. The
r-apoA-I primarily eluted with the 26-36-ml fractions, whereas
the majority of apoA-I eluted with the 32-36-ml
fractions. Plasma obtained 20 min and 4 h after injection of
radioiodinated r-apoA-I and apoA-I into the rabbits were
also analyzed by gel filtration chromatography as shown in Fig. 9, C and D. Twenty min after injection (Fig. 9C) r-apoA-I was associated with the HDL
fraction, which elutes with the 26-32-ml fractions.
ApoA-I is almost equally distributed within the HDL and
the phospholipid-rich VHDL fraction. Four hours after injection, the
total counts of apoA-I rapidly disappear, especially
within the VHDL fraction (Fig. 9D). To verify that the
radioactivity in each fraction is associated with apoA-I, we analyzed
each fraction by polyacrylamide gel electrophoresis and immunoblot
analysis. Within 26-36 ml of the eluted volume all fractions
showed immunodetectable rabbit apoA-I with higher concentrations of
apoA-I within the 26-32-ml fraction (data not shown). The same
fractions were concentrated by trichloroacetic acid precipitation and
analyzed by polyacrylamide gel electrophoresis. The autoradiogram
showed identical distribution of r-apoA-I compared with rabbit apoA-I,
which suggests that rabbit apoA-I and r-apoA-I associate with HDL in a
similar pattern, which, in turn, is not surprising since there is a
high sequence homology between rabbit and human apoA-I(54) .
and r-apoA-I were analyzed
prior to injection (Fig. 9B), 20 min (Fig. 9C), and 4 h (Fig. 9D) after
injection into New Zealand White rabbits.
had a stop
codon in the middle of the last amphipathic helix of apoA-I. Residues
from 226 to 243 have been proposed to play a role in apoA-I binding to
cells(9, 57, 58) . For the apoA-I mutant, the truncation occurs in the hinge region between the
eighth and ninth helix. In addition, we truncated apoA-I at residue
201, the site of an apoA-I phosphorylation (59) as well as a
frameshift mutation which results in a truncated apoA-I and low levels
of circulating apoA-I(60) . In contrast to the similar kinetics
of clearance for r-apoA-I and h-apoA-I (Fig. 5A), all
of the truncation mutants were catabolized significantly faster (Fig. 5, B-D). Moreover, the greater the
carboxyl-terminal deletion, the greater the FCR (apoA-I,
FCR = 4.42 ± 0.51/day; apoA-I, FCR =
6.34 ± 0.81/day; apoA-I, 9.10 ± 1.28/day).
Urine analysis revealed that the rapid catabolism of apoA-I is associated with a dramatic increase of its free radiotracer
within the urine. and
HDL. In contrast, the truncated forms of apoA-I were
relatively lipid-poor and associated with VHDL (Fig. 6). We
defined the rabbit HDL subfractions by the same density fractions as in
humans. We obtained similar results when the lipid-protein association
of the apoA-I forms were examined in vivo (Fig. 7, A-C). Composition analysis of the rabbit VHDL revealed a
phospholipid rich apoA-I containing particle (Fig. 9), which
shows a similar composition as human VHDL(53) . The analysis of
r-apoA-I and apoA-I before and after injection into
rabbits by gel filtration chromatography revealed that
apoA-I, in contrast to r-apoA-I, eluted in the fractions
containing the phospholipid rich apoA-I containing particle. In
addition, apoA-I associated with a lipoprotein particle
that migrated in a pre- position on agarose gel electrophoresis (Fig. 8), in contrast to the position observed for
r-apoA-I (Fig. 8). has been shown to form pre- HDL
after incubation with hepatic lipase(71) . Similarly, HDL can
transform into discoidal HDL particles by cholesteryl ester transfer
protein and hepatic lipase activity(72) . This, in turn, may
lead to a more loosely bound apoA-I that is susceptible to increased
renal clearance, due to the altered surface properties of lipase
modified HDL. The predicted structure of apoA-I, as illustrated in Fig. 2, consists of amphipathic -helices that span the
surface of the highly curved discoidal HDL(14, 16) .
Sparks et al.(73) examined the structural stability
of discoidal and spherical HDL particles and concluded that discoidal
HDL is unstable and, therefore, has the tendency to interact with cell
membranes and other lipoproteins, thus facilitating cholesterol efflux.
The transition of discoidal HDL particles into less dense HDL particles
may be accompanied by dissociation of lipid poor apoA-I, which is then
rapidly catabolized. The concept of increased clearance of loosely
bound apoA-I is also supported by studies of model peptides that differ
in their affinity for HDL(64) . Synthetic peptides that have
decreased lipid association have increased catabolism in rats. In
addition, Westerlund (65) showed that carboxyl-terminal
truncation mutants of apoE, an apolipoprotein with similar structural
features of apoA-I, have different lipid association properties, which
result in more non-lipoprotein bound apoE and an increased metabolic
clearance. Similarly, our data suggest that the apoA-I truncation
mutants are relatively lipid poor and bound to a phospholipid rich
apoA-I containing particle that migrates in pre- position. The
increased metabolic clearance we observed for apoA-I truncation mutants
is likely to be related to their decreased ability to associate with
HDL and HDL. Either the smaller lipoprotein
particles itself or the increased dissociation of apoA-I from the
smaller HDL lipoprotein particles contributes to the rapid clearance of
the truncated forms of apoA-I. HDL, is
cleaved by an EDTA sensitive protease shortly after
phlebotomy(34) . This cleavage has been shown to occur at the
carboxyl-terminal domain of the protein, because the resulting 26-kDa
fragment has a complete amino-terminal domain(34) . The
carboxyl-terminal domain is also susceptible to cleavage in vitro by several different
proteases(36, 37, 38, 39, 40, 41) .
This suggests that apoA-I has a site in the carboxyl-terminal domain,
which is near the site of the apoA-I truncation mutants, that is
particularly susceptible to proteolysis. Our results suggest that
cleavage by a protease in the region between the middle of the eighth
amphipathic helix and the ninth amphipathic helix would produce an
apoA-I fragment with altered functional properties: decreased lipid
association and increased metabolic clearance. This suggests a possible
mechanism for the regulation of the catabolism of apoA-I by
proteolysis, which is known to occur for other plasma proteins (26, 27, 28, 29, 30, 31) . -helical domain markedly altered the lipoprotein association
properties and increased catabolism of apoA-I. The more the
carboxyl-terminal domain was truncated, the faster apoA-I was
catabolized. The truncated forms of apoA-I appeared on smaller and
denser phospholipid rich HDL particles, which suggests a possible
mechanism for increased clearance of truncated apoA-I.
)
We acknowledge Drs. Douglas Powell, Robert F. Hoyr,
Jr., and the other members of the Section on Laboratory Animal Medicine
and Surgery, NHLBI for help in critical catheter placement as well as
in the care of these animals; Susan Meyn and Glenda Talley for
excellent technical support, and Nysa Sanders for help in preparing the
manuscript.
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  H. Komori, H. Arai, T. Kashima, T. Huby, T. Kita, and Y. Ueda Coexpression of CLA-1 and Human PDZK1 in Murine Liver Modulates HDL Cholesterol Metabolism Arterioscler. Thromb. Vasc. Biol., July 1, 2008; 28(7): 1298 - 1303. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. G. Frank and Y. L. Marcel Apolipoprotein A-I: structure;-function relationships J. Lipid Res., June 1, 2000; 41(6): 853 - 872. [Abstract] [Full Text]
|
|
|
|
 |
|
 D. C. McManus, B. R. Scott, P. G. Frank, V. Franklin, J. R. Schultz, and Y. L. Marcel Distinct Central Amphipathic alpha -Helices in Apolipoprotein A-I Contribute to the in Vivo Maturation of High Density Lipoprotein by Either Activating Lecithin-Cholesterol Acyltransferase or Binding Lipids J. Biol. Chem., February 18, 2000; 275(7): 5043 - 5051. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. Huang, J. Sasaki, A. Matsunaga, H. Han, W. Li, T. Koga, M. Kugi, S. Ando, and K. Arakawa A Single Amino Acid Deletion in the Carboxy Terminal of Apolipoprotein A-I Impairs Lipid Binding and Cellular Interaction Arterioscler. Thromb. Vasc. Biol., January 1, 2000; 20(1): 210 - 216. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Han, J. Sasaki, A. Matsunaga, H. Hakamata, W. Huang, M. Ageta, T. Taguchi, T. Koga, M. Kugi, S. Horiuchi, et al. A Novel Mutant, ApoA-I Nichinan (Glu235->0), Is Associated With Low HDL Cholesterol Levels and Decreased Cholesterol Efflux From Cells Arterioscler. Thromb. Vasc. Biol., June 1, 1999; 19(6): 1447 - 1455. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Jauhiainen, J. Huuskonen, M. Baumann, J. Metso, T. Oka, T. Egashira, H. Hattori, V. M. Olkkonen, and C. Ehnholm Phospholipid transfer protein (PLTP) causes proteolytic cleavage of apolipoprotein A-I J. Lipid Res., April 1, 1999; 40(4): 654 - 664. [Abstract] [Full Text]
|
|
|
|
|
|
K. L. Gillotte, M. Zaiou, S. Lund-Katz, G. M. Anantharamaiah, P. Holvoet, A. Dhoest, M. N. Palgunachari, J. P. Segrest, K. H. Weisgraber, G. H. Rothblat, et al. Apolipoprotein-mediated Plasma Membrane Microsolubilization. ROLE OF LIPID AFFINITY AND MEMBRANE PENETRATION IN THE EFFLUX OF CELLULAR CHOLESTEROL AND PHOSPHOLIPID J. Biol. Chem., January 22, 1999; 274(4): 2021 - 2028. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. G. Sorci-Thomas, L. Curtiss, J. S. Parks, M. J. Thomas, and M. W. Kearns Alteration in Apolipoprotein A-I 22-Mer Repeat Order Results in a Decrease in Lecithin:Cholesterol Acyltransferase Reactivity J. Biol. Chem., March 14, 1997; 272(11): 7278 - 7284. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Sviridov, L. E. Pyle, and N. Fidge Efflux of Cellular Cholesterol and Phospholipid to Apolipoprotein A-I Mutants J. Biol. Chem., December 27, 1996; 271(52): 33277 - 33283. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Moriyama, J. Sasaki, Y. Takada, A. Matsunaga, J. Fukui, J. J. Albers, and K. Arakawa A Cysteine-Containing Truncated Apo A-I Variant Associated With HDL Deficiency Arterioscler. Thromb. Vasc. Biol., December 1, 1996; 16(12): 1416 - 1423. [Abstract] [Full Text]
|
|
|
|
 |
|
 W. S. Davidson, T. Hazlett, W. W. Mantulin, and A. Jonas The role of apolipoprotein AI domains in lipid binding PNAS, November 26, 1996; 93(24): 13605 - 13610. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. L. Gillotte, W. S. Davidson, S. Lund-Katz, G. H. Rothblat, and M. C. Phillips Apolipoprotein A-I Structural Modification and the Functionality of Reconstituted High Density Lipoprotein Particles in Cellular Cholesterol Efflux J. Biol. Chem., September 27, 1996; 271(39): 23792 - 23798. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M.N. Nanjee, J.R. Crouse, J.M. King, R. Hovorka, S.E. Rees, E.R. Carson, J.-J. Morgenthaler, P. Lerch, and N.E. Miller Effects of Intravenous Infusion of Lipid-Free Apo A-I in Humans Arterioscler. Thromb. Vasc. Biol., September 1, 1996; 16(9): 1203 - 1214. [Abstract] [Full Text]
|
|
|
|
|
|
P. Holvoet, Z. Zhao, E. Deridder, A. Dhoest, and D. Collen Effects of Deletion of the Carboxyl-terminal Domain of ApoA-I or of Its Substitution with Helices of ApoA-II on in Vitro and in Vivo Lipoprotein Association J. Biol. Chem., August 9, 1996; 271(32): 19395 - 19401. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. N. Palgunachari, V. K. Mishra, S. Lund-Katz, M. C. Phillips, S. O. Adeyeye, S. Alluri, G.M. Anantharamaiah, and J. P. Segrest Only the Two End Helixes of Eight Tandem Amphipathic Helical Domains of Human Apo A-I Have Significant Lipid Affinity : Implications for HDL Assembly Arterioscler. Thromb. Vasc. Biol., February 1, 1996; 16(2): 328 - 338. [Abstract] [Full Text]
|
|
|
|
|
|
B. R. Scott, D. C. McManus, V. Franklin, A. G. McKenzie, T. Neville, D. L. Sparks, and Y. L. Marcel The N-terminal Globular Domain and the First Class A Amphipathic Helix of Apolipoprotein A-I Are Important for Lecithin:Cholesterol Acyltransferase Activation and the Maturation of High Density Lipoprotein in Vivo J. Biol. Chem., December 21, 2001; 276(52): 48716 - 48724. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| All ASBMB Journals | Molecular and Cellular Proteomics |
| Journal of Lipid Research | ASBMB Today |