(Received for publication, April 16, 1997)
From the § Baker Medical Research Institute, Melbourne,
Victoria 3181, Australia and the National Institute
of Health and Nutrition, Tokyo 162, Japan
The protection against coronary artery disease attributed to high density lipoprotein (HDL) may be associated with several functions, including its central role in reverse cholesterol transport, possible antioxidant and antithrombotic properties and others not yet identified which may depend on specific interactions between HDL and cell receptors. Several HDL-binding proteins have been identified including two candidate liver HDL receptors, HB1 and HB2 recently purified in this laboratory. We now report the cloning, sequencing, and some properties of HB2, the most abundant of the pair. It shows significant homology with the adhesion molecules ALCAM and BEN of the immunoglobulin superfamily and the cDNA, when transfected into HepG2 or COS cells, caused specific HDL3 binding to increase by 80-100%. Further, ligand blotting of glycoproteins isolated from phorbol 12-myristate 13-acetate-treated THP-1 cells or from transfected HepG2 and Chinese hamster ovary cells also provided evidence of increased binding of HDL3 to HB2. Differentiation of THP-1 cells into macrophages resulted in a striking increase in HB2 mRNA which was attenuated if cells were cholesterol-loaded by incubation with acetylated low density lipoprotein. If the interaction between HDL and HB2 reduces the adhesion-induced inflammatory cellular events that characterize arterial wall injury, thereby achieving the protection associated with higher plasma levels of HDL, these findings may provide a clue to one mitigating effect of HDL in heart disease.
The incidence of premature artery disease is lowered in the presence of high levels of circulating high density lipoprotein (HDL)1 suggesting that HDL may protect individuals against atherosclerosis (1). Part of this protection may be attributed to the participation of HDL in the reverse cholesterol transport pathway, a mechanism which reduces the accumulation of cholesterol in the arterial wall and the narrowing which subsequently occurs (2).
More recent evidence would indicate that HDL influences many other biological events in a manner consistent with a protective function, suggesting that the antiatherogenic role of HDL is likely to be complex. HDL may have antioxidant and antithrombotic properties (3-5), it may act as a signal transductant and as a secretagogue (6), and it may possibly influence the production of some cell adhesion molecules such as VCAM and ICAM (7).
Despite these observations, the mechanism(s) responsible for the HDL-induced physiological effects remain enigmatic. The fact that HDL binds to a variety of cells with varying degrees of specificity has invited speculation about the existence of an HDL receptor. Our laboratory (8) and others (9, 10) have purified liver cellular HDL-binding proteins, thereby strengthening the evidence for the existence of HDL receptor(s) that may have important functions as determinants of HDL metabolism. To date, only SR-B1, a member of the scavenger receptor class of membrane proteins that binds LDL, modified LDL and HDL, has been shown to elicit a physiological response, viz. the transfer of cholesteryl ester into cells, particularly in steroidogenic tissues (11).
We have identified two liver plasma membrane proteins named HB1 and HB2 (HDL-binding proteins 1 and 2) which are candidate HDL receptors (8, 12). Although present in low abundance, HB2 was purified in sufficient quantities to provide amino acid sequence data. This paper describes the subsequent cloning of HB2, its complete amino acid sequence, and evidence for increased binding of HDL to HB2 in cells overexpressing HB2.
HepG2, COS, CHO, and THP-1
cells were obtained from the American Type Culture Collection
(Rockville, MD) and maintained under standard tissue culture conditions
using Dulbecco's modified Eagle's medium for HepG2 and COS,
-minimal essential medium for CHO cells, and RPMI 1640 medium for
THP-1 cells, plus 10% fetal bovine serum. THP-1 cells were
differentiated into adherent macrophages by treatment with 200 nM phorbol 12-myristate 13-acetate (PMA) for 72 h.
Human HDL3 (d 1.12-1.21 g/ml) was obtained from plasma as described previously and contained apoAI and apoAII but no apoE (8). HDL3 was radioiodinated using IODOBEADS (Pierce) to specific activities ranging from 150-210 cpm/ng of protein. HDL3 was also labeled with [3H]cholesteryl oleoyl ether (13). Briefly, HDL3 was incubated with donor particles containing egg yolk phosphatidylcholine, [3H]cholesteryl oleoyl ether, and lipoprotein-deficient serum (d > 1.21 g/ml) as a source of cholesteryl ester (CE) transfer protein for 6 h at 37 °C, and the HDL was then separated from donor particles by ultracentrifugation at 1.06 g/ml. HDL3 was labeled with [14C]CE by incubating a d > 1.12 g/ml (dialyzed) fraction of human plasma (lecithin:cholesterol acyltransferase source) with [14C]cholesterol (Amersham). The CE-labeled HDL was isolated by ultracentrifugation at d 1.21 g/ml and dialyzed.
Protein SequencingHB2, a glycoprotein of Mr 100,000 was purified from rat liver plasma membrane as described previously (12). Attempts at NH2-terminal sequencing were unsuccessful. Internal sequences were obtained following cleavage with an endoproteinase (Lys C) using an enzyme substrate ratio of 1:10 (w/w). Peptides were separated by high performance liquid chromatography using an RP300 column (30-min gradient of acetonitrile from 0-60%) in 0.1% trifluoroacetic acid as mobile phase. This approach produced low yields and limited sequence information. Better yields were obtained following digestion of HB2 with cyanogen bromide (CNBr) for 16 h at room temperature. The CNBr fragments were separated by SDS-PAGE (12% acrylamide) and transferred to Problott membranes (Applied Biosystems), and the peptides were detected by staining with Coomassie Blue. Bands on Problott membranes were excised, cut into pieces, and applied directly to the cartridge of an Applied Biosystems Model 470A sequencer equipped with an on-line Model 120A PTH analyzer. Sequences, ranging from 11 to 20 amino acid residues, were obtained from several peptides.
cDNA Cloning and Sequence AnalysisTotal RNA was
prepared from both liver and lung of male Sprague-Dawley rats by
homogenizing with guanidium isothiocyanate and centrifugation through a
cushion of CsCl (14). Total RNA was used for cDNA synthesis using
Moloney murine leukemia virus reverse transcriptase and random hexamers
and oligo(dT)16 as primers. PCR was performed with
oligonucleotides 9511 (forward) and 9519 (reverse) (based on the amino
acid sequence of internal peptides of HB2) as follows:
primer 9511, 5-CA(A/G)TAFGAFGA(T/C)GT(F/G)CC(F/G)GA(A/G)TA; 9519, 5
-T(A/G)AAFTCFTC(F/G)GG(F/G)GGNGG(A/G)TT (F = T/C, N = any).
As indicated, 5
-fluoro-dUMP (F) was incorporated into the oligonucleotides during synthesis (Bresatec, Australia) to limit degeneracy (15). Conditions for PCR were 95 °C, 2 min for one cycle,
then 30 cycles of 95 °C, 1 min, an annealing step commencing at
53 °C for 30 s, then increasing to 72 °C at a ramp rate of 0.1 °C/s; 72 °C, 2 min ending with 5-min elongation at 72 °C. The amplified product was subcloned into vector, pCRTMII,
using a TA cloning kit (Invitrogen). Double-stranded sequencing of the
fragment was performed using M13 forward and reverse primers and
Sequenase Version 2 (U. S. Biochemical Corp.).
Poly(A)+ RNA from rat lung was isolated from total RNA by Oligotex [d(T)]30 (Roche). Synthesis of cDNA from 5 µg of poly(A)+ RNA was catalyzed by Moloney murine leukemia virus reverse transcriptase with a Timesaver cDNA synthesis kit (Pharmacia) which includes EcoRI/NotI adapters. The cDNA was ligated into the EcoRI site of Lambda ZAP II vector (Stratagene) and packaged into phage particles with Gigapack III Gold packaging extract (Stratagene). The recombinant phage was used to infect Escherichia coli strain XL1-Blue MRF.
Approximately 2.5 × 105 plaque-forming units were
screened for HB2 clones after lifting onto Colony/Plaque
ScreenTM membrane (DuPont). The PCR product from primers
9511/9519 was labeled with [-32P]dCTP by random
priming. Hybridization was performed at 42 °C in 5 × SSPE,
5 × Denhardt's solution, 1% SDS, 50% formamide, and 100 µg/ml denatured salmon sperm DNA. Filters were washed three times
with 2 × SSC and 0.1% SDS solution at 50 °C for 15 min, then
washed with 0.1 × SSC and 0.1% SDS solution at 50 °C for 30 min. Thirty-three positive plaques remained after tertiary screening
from which 15 clones were selected, excised in vivo, and
used for further sequencing analysis.
The HB2 cDNA (1- 2,771 nucleotides) was introduced into the EcoRI site of the expression vector, pcDNA I/Amp (Invitrogen) for transient expression. Transfection was performed by the DEAE-dextran method for COS and CHO cells and by calcium phosphate precipitation for HepG2 cells (16). The same procedure was followed, except for the omission of the plasmid, to produce mock-transfected cells. After 24 h, the cells were removed from the flasks and dispensed into 12-well plates at appropriate seeding densities in preparation for binding experiments which were performed at 48 h after transfection. In some experiments, cells were washed and harvested, and membranes were recovered for ligand blotting analysis (see below).
Binding ExperimentsCOS or HepG2 cells in monolayers were maintained in DMEM containing 10% fetal bovine serum. Prior to binding experiments, cells were washed with serum-free DMEM, and the growth medium was replaced with DMEM containing 0.2% (w/v) bovine serum albumin and 25 mM HEPES, pH 7.4. Transfected and mock-transfected cells were incubated with 125I-labeled HDL3 (as described under "Results") with or without unlabeled HDL3. Specific binding was determined as described previously (17).
Western and Northern BlotsCell membranes were prepared as described previously (8). Immunoblotting was performed after transfer of membrane proteins (following SDS-PAGE) to nitrocellulose which was incubated with antiserum against HB2 and then horseradish peroxidase-conjugated second antibody (8).
For Northern blot analysis of rat tissues, total RNA was isolated from tissues including liver, intestine, lung, spleen, kidney, brain, ovary, and thymus, fractionated by formaldehyde-agarose gel electrophoresis, and transferred to nylon membranes. These were probed with 32P-labeled probes (see "Results"), and hybridizing bands were identified by autoradiography or bioimaging. Membranes were stripped and rehybridized with a cDNA probe for rat glyceraldehyde phosphate dehydrogenase. For detection of HB2 mRNA in human tissues, human multiple RNA blots (CLONTECH) were similarly probed as described above, except that the actin probe provided with the blots was used as an internal standard.
Ligand BlotsCells were lysed in the presence of 20 mM CHAPS, and solubilized proteins were recovered by centrifugation at 18,000 × g for 10 min. Glycoproteins were isolated with wheat germ lectin agarose (Pharmacia) and then applied to SDS-PAGE gels. Ligand blotting was performed as described previously (12).
Computer AnalysisData base searches with the nucleotide sequence of rat HB2 cDNA were performed with GenBank and EMBL data bases by the FASTA program from the Genetics Computer Group sequence analysis software. Alignments and consensus sequence derivation of peptide sequences were analyzed by GENETYX (version 9.0) from Software Development Co., Ltd. (Tokyo, Japan). Further analyses of the HB2 gene product for common motifs and potential secondary structure were performed using GENETYX.
HB2 antiserum was used to probe membranes from various rat organs as described under "Materials and Methods." As shown previously (18), the strongest expression of HB2 was present in liver, lung, and intestine. Total RNA from lung and liver were therefore subjected to reverse transcription-PCR as described under "Materials and Methods," with primers 9511 and 9519.
Under the annealing/elongation conditions described for PCR, a single product of 600 base pairs was amplified from lung, but not from liver. Sequencing of this fragment confirmed the presence of both primers as well as an additional internal peptide of HB2, in the correct reading frame. To obtain a full-length clone, a lung cDNA library was prepared and probed with the 600-base pair fragment. Of 2.5 × 105 plaque-forming units probed, 33 positive clones remained after three rounds of screening. Fifteen of these were further analyzed. As shown in Fig. 1, the cDNA of HB2 encoded a protein of 65 kDa constituted from 583 amino acids, considerably smaller than the apparent size of 100 kDa of the glycosylated form of HB2 identified by ligand blotting as previously reported by our laboratory (12). Over the entire coding region, rat HB2 showed highest homology with human ALCAM (93%) (19) and the avian protein BEN (70%) (20), which are adhesion proteins of the immunoglobulin superfamily. Alignment with ALCAM and BEN is also shown in Fig. 1. There was no significant homology with any other lipoprotein receptor including candidate HDL receptors previously reported (11, 21).
Binding to HDL3
To demonstrate that this
transmembrane protein functions as an HDL-binding protein, COS, CHO, or
HepG2 cells were transfected or mock-transfected as described under
"Materials and Methods." As shown in Fig. 2,
A and B, specific binding of
125I-labeled HDL3 was increased (at saturation)
approximately 2-fold in transfected HepG2 cells and by 1.8-fold in
transfected COS cells compared with mock-transfected cells. This
increased binding however was not associated with transfer of
cholesteryl esters from HDL to cells because no differences were
observed in cholesteryl ester uptake between mock and transfected COS
or HepG2 cells (data not shown).
Glycoproteins isolated from mock- or transiently transfected CHO or HepG2 cells were compared for their HDL binding capacity by ligand blots. As shown in Fig. 2C, untransfected HepG2 cells showed binding of HDL3 to both HB2 and HB1, but the signal for HB2 was much stronger in transfected cells. Similarly the faint signal for HB2 observed in mock-transfected CHO cells was markedly increased when membrane proteins from transfected CHO cells were incubated with HDL3 (Fig. 2D).
Expression of HB2 mRNA in Rat and Human TissuesFig. 3A shows Northern blot
analyses of HB2 mRNA expression in various rat tissues.
Strongest expression was found in the lung, then brain, liver, and
kidney. In comparison, to the rat, human mRNA to HB2
(probed with rat HB2 cDNA) was strongest in the brain,
prostate, pancreas, small intestine, and liver of human (Fig.
3B), but lower in the lung.
To determine whether blood monocytes or macrophages express
HB2, RNA from THP-1 cells or from cells differentiated by
treatment with PMA was subjected to Northern blot analysis. As shown in Fig. 4A, HB2 mRNA, hardly
detectable in untreated THP-1 cells, was strongly induced when the
cells were transformed into macrophages by PMA treatment. After
incubating THP-1 cells with 50 or 100 µg/ml acetylated LDL, a
dose-dependent reduction in HB2 mRNA
expression was observed. No changes were observed in expression of
glyceraldehyde phosphate dehydrogenase mRNA with any of the
treatments (Fig. 4B). Ligand blots of membrane proteins
revealed weak HDL3 binding for THP1 cells, but after
differentiation with PMA, strong binding was detected for
HB2 (Fig. 4C) as well as an increase in binding to HB1. Binding was also stronger to another protein
(approximately 66 kDa) which may represent degraded HB2 or
another isoform of HB2.
Cloning and sequencing of HB2, one of a pair of HDL-binding proteins previously identified and purified in this laboratory (12), have revealed its identity with a subclass of the immunoglobulin superfamily of proteins with adhesion type properties. To confirm that HB2 performed a role as an HDL-binding protein, HepG2, CHO, and COS cells were transiently transfected with HB2 cDNA. Compared with control (mock-transfected) cells, HepG2 and COS cells (transfected) demonstrated an 80-100% increase in HDL binding, establishing that HB2, together with other candidate HDL receptors (11, 21), may play an important role in HDL metabolism in the body. The level of HDL binding activity resulting from transfection with HB2 cDNA was intermediate between that of an HDL-binding protein found in endothelial cells when expressed in COS cells (21) and that of SR-B1 (11), another candidate HDL receptor. It is not surprising that untransfected cells will demonstrate variable levels of specific HDL binding in view of the presence on most cells of "specific" but low affinity HDL binding sites recently identified in this laboratory (22). This phenomenon is a characteristic of the interaction between HDL and cells and the most likely explanation for the relatively high level of specific binding observed in the mock-transfected cells in this study.
Additional evidence that HB2 is involved in the cellular recognition of HDL came from ligand blotting studies which confirmed that this membrane protein, when overexpressed in HepG2, CHO, and THP-1 cells, bound HDL3. Untransfected cells, apparently expressing low levels of HB2 compared with HB1 (Fig. 2C) showed a marked increase in HB2 that was active in binding HDL3, when transfected with HB2 cDNA. Similarly, ligand blotting revealed a significant increase in HDL binding to HB2 present in transiently transfected CHO cells, compared with controls. The up-regulation of HB2 mRNA that followed transformation of THP-1 cells into "macrophages" following PMA treatment was also associated with an increase in expression of HB2 which was active in binding HDL3 as seen in Fig. 4C.
These experiments have not yet provided definitive information about the clinical relevance of HB2, but they do support a functional role for HB2 that involves interaction with HDL. Expression of HB2 mRNA, barely detectable in blood monocytes, is up-regulated when THP-1 cells undergo differentiation to macrophages on treatment with PMA, and further, a reduction in expression of HB2 mRNA follows cholesterol loading with acetylated LDL. This suggests some association, either direct or indirect, between HB2 and cholesterol metabolism. If, as expected, sterol synthesis was decreased in the cholesterol-loaded cells, it would be consistent with a finding reported by this laboratory previously (23) that the administration of simvastatin to rats down-regulated HB2 (and HB1) expression by 50%. More studies are planned to investigate these relationships between HDL binding, HB2 levels, and cell cholesterol metabolism. Our present studies however have demonstrated that expression of HB2 is not associated with selective uptake of HDL cholesteryl ester by cells, a function which has been demonstrated for another candidate HDL receptor, SR-B1 (11).
Another function attributed to the family of membrane proteins that includes HB2, BEN, and ALCAM is adhesion. When expression of HB2 is up-regulated in cells (such as blood monocytes as described above) it is conceivable that the increased adhesion that follows may produce vascular remodelling and the initiation of atherogenesis. While many proteins which elicit injurious effects on the vessel wall are known to be produced by stimulated macrophages, increased levels of HDL or of apoAI-rich particles, may compete with binding sites on the HB2 adhesion molecules to significantly reduce the migration of macrophages into the arterial wall thereby achieving some protection against damage to the vessel. This model is entirely consistent with the protective role of HDL against premature atherosclerosis. The scavenger receptor cysteine-rich domain of CD6 is also recognized by this family of adhesion molecules (24), and interruption of this interaction by HDL may contribute to a decreased cell-cell association that is injurious to the vessel wall. Together with the observation that HDL apparently down-regulates the expression of soluble adhesion molecules (7), the combined inhibitory effect of HDL on cell adhesion may provide a formidable protective action against vascular disease. Clearly, these and other systems need to be explored, because the protective role of HDL is apparently not only limited to metabolic processes connected with lipid metabolism, such as reverse cholesterol transport, but may include mitogenic effects, secretagogue activity, and possible cellular signaling pathways (6). In fact, the tissue distribution of HB2 mRNA is consistent with any of the functions suggested above. In the human, organs that play important roles in lipoprotein transport such as the liver and intestine and steroidogenic tissue express HB2, but others such as the lung, brain, placenta, and pancreas, which may participate in the alternative protective functions listed above, also contain varying proportions of HB2 mRNA.
The structural features of HB2 are also consistent with a
"receptor" role for this membrane protein. As shown in the working model (Fig. 5) based on available structural information
on this group of membrane proteins of the IgG superfamily,
HB2 is characterized by a 32-amino acid cytoplasmic domain,
a 24-amino acid hydrophobic transmembrane domain, and approximately 500 residues of an extracellular domain terminating in the NH2
residue. Data base searches of consensus sequences for
N-glycosylation sites revealed eight potential sites, all
extracellular, at residues 95, 167, 265, 306, 361, 457, 480, and 499. It is likely that most sites are glycosylated since the protein is
mainly expressed in the higher Mr form, although
in HepG2 cells the presence of a faster migrating HB2
isoform was found by Western blotting.2 The
function of glycosylation in HDL binding is unclear, although previous
ligand blot studies (12) indicate that glycosylation is not essential
for HDL binding. The carbohydrate however may contribute to other
essential functions such as intracellular transport and influence the
capacity of HB2 (in some tissues) to reach sites for HDL
binding. Some provisional protein kinase C phosphorylation sites
(residues 8, 73, 74, 209, and 421) were found, but since all are
extracellular, their potential function as determinants of signaling
via protein kinase C is weakened. cAMP- and cGMP-dependent
protein kinase phosphorylation sites were also found. Possibly, sites
which are inactive when HB2 is membrane-bound are activated
if the protein is internalized. Further experiments are planned to
investigate the interrelationships between structure and function, as
related to HDL binding, and cellular localization of
HB2.
We are grateful to Anne Au and Ann Seward for expert technical assistance.