STRUCTURAL BIOLOGY:
LDL Receptor's b-Propeller Displaces LDL
Thomas L. Innerarity*
Much of what is known about receptor-mediated
endocytosis comes from studies of the low density lipoprotein receptor (LDLR)
pathway (1). LDLR binds cholesterol-carrying LDL,
associates with clathrin-coated pits, and is internalized into acidic
endosomes where it separates from its ligand. The ligand is degraded in
lysosomes, while the receptor returns to the cell surface. Mutations in
the LDLR gene can lead to elevated plasma cholesterol levels, resulting
in coronary heart disease and artherosclerosis (1).
Seminal observations by Rudenko et al. on page
2353
in this issue (2) shed light on a mystery of LDLR
recycling--how the LDLR releases its lipoprotein ligand in the endosome.
The LDLR has several domains (see the figure). The ligand-binding
domain contains seven imperfect repeats, each with three disulfide bonds
and a coordinated Ca2+ ion. Extracellularly, it binds two
ligands: apolipoprotein (apo) B100 (the only protein in LDL) and apoE (a
protein in other lipoproteins). The second domain (411 amino acids in
length) is analogous to the membrane-bound precursor of the epidermal
growth factor (EGF). It consists of two EGF repeats, followed by a
b-propeller region that contains the
consensus sequence Tyr-Trp-Thr-Asp, and another EGF repeat (see the
figure). LDLR with the EGF precursor domain deleted still binds apoE,
but not LDL. However, apoE is not released in the endosome, and the
ligand-receptor complex is degraded in the lysosome. Thus, the EGF
precursor domain is critical for ligand release and recycling of the
receptor, but until now the mechanism remained a mystery (3).
The solution came from the structure of the extracellular domain of
the human LDLR crystallized at pH 5.3. In this structure, the
b-propeller region of the EGF precursor
domain interacts with the main ligand-binding repeats of the LDLR (R4
and R5) (see the figure). Rudenko et al. (2)
propose that in the endosomes, the b-propeller
region displaces the bound lipoprotein ligand by acting as an alternate
substrate for the ligand-binding domain. This compelling model is
supported by other key evidence: mutations in the ligand-binding and EGF
precursor regions that abolish function, phylogenetic evidence of
conserved amino acids, and biochemical evidence that the ligand-binding
repeats associate with the EGF precursor at pH 6 but not at pH 8.
Clusters of histidines in the b-propeller
region likely act as pH-sensitive switches for the domain interactions.
The histidines carry no net charge at pH 7.3, but are partially charged
at pH 5.3 and participate in the formation of salt bridges in the
crystal structure.
Catch and release. A model for how LDLR releases LDL. A
crystal structure of the extracellular domain of LDLR at pH 5.3 (2)
shows that ligand-binding repeats R4 and R5 interact with the
b-propeller region of the EGF precursor
domain. This interaction may displace LDL from the receptor in acidic
endosomes.
CREDIT: KATHARINE SUTLIFF/SCIENCE
The interaction of the b-propeller region
with repeats 4 and 5 appears to have much in common with the
interactions between lipoprotein ligands and the ligand-binding repeats
and clarifies a controversy about lipoprotein-receptor interactions. The
interaction of the two domains of the LDLR, as shown by the crystal
structure, is based on six hydrophobic bonds and seven salt bridges
between R4/R5 and the b-propeller region.
Previous studies indicated that ionic or salt bridges are also critical
for LDLR-ligand interactions (4, 5),
with conserved acidic amino acids in the ligand-binding repeats forming
ionic interactions with positively charged amino acids in the receptor
binding site of the lipoprotein ligands. However, this ionic interaction
model has been questioned because many of the conserved acidic amino
acids in the ligand-binding repeats coordinate Ca2+ and are
completely or partially buried in a Ca2+ cage. Hence they are
presumably unavailable to bind to apoB100 or apoE. Instead, a
hydrophobic concave face on the opposite side of the Ca2+
cage was proposed to interact with the lipoproteins (6).
The crystal structure reported by Rudenko et al. (2)
resolves this conundrum by illustrating that some conserved acidic amino
acids that coordinate Ca2+ also participate in the formation
of salt bridges with basic residues of the b
propeller. Although the negative charge potential is somewhat
attenuated, the three disulfide bonds and the Ca2+
coordination lock the negatively charged side chains of R4 and R5 in
place for optimal interaction with the basic residues of the
b-propeller region (2).
Rudenko et al. (2) point out that the
ligand-binding repeats are not in contact with each other and can
accommodate different-sized ligands. ApoE (relative molecular mass
33,000) and apoB100 (relative molecular mass 550,000) differ
dramatically in size and have no common structural features or amino
acid sequence similarity with the exception of a short sequence that
serves as the receptor binding site and main proteoglycan binding site
in both (5, 7, 8).
Extensive studies on apoE show that basic residues in this region are
critical for receptor binding (4, 5,
9), and a three-dimensional structure of the 22-kD
LDLR binding domain of apoE shows that the receptor binding site is a
positively charged, amphipathic helix (10).
The larger apoB100 is less well understood, but an analogous sequence
is the likely site for receptor binding. Mutation of basic amino acids
in this site to neutral amino acids abolishes receptor binding (8).
Because only one small site common to both apoE and apoB100 appears
critical for receptor binding, it is likely these proteins have critical
interactions with only one or two ligand-binding repeats of the LDLR.
This is analogous to the b-propeller, which
only interacts with two ligand-binding repeats (R4, R5).
The model of acidic-triggered ligand release by binding to an
alternate tethered site will probably be the paradigm for other members
of the LDLR family. Will other receptors engaged in receptor-mediated
endocytosis outside of the LDLR family have a similar mechanism?
Finally, although this study provides insights into the binding of the
LDLR with its ligands, a definitive answer will only come from the
cocrystallization of a receptor-binding active fragment of apoE with the
seven ligand-binding repeats of the LDLR.
References
- M. S. Brown, J. L. Goldstein, Science 232,
34 (1986).
[Medline]
- G. Rudenko et al., Science 298,
2353 (2002); published online 28 November 2002 (10.1126/science.
1078124).
- C. G. Davis et al., Nature 326,
760 (1987).
[Medline]
- R. W. Mahley, Science 240, 622 (1988).
[Medline]
- K. H. Weisgraber, Adv. Protein Chem. 45,
249 (1994).
[Medline]
- D. Fass, S. Blacklow, P. S. Kim, J. M. Berger, Nature
388, 691 (1997).
[Medline]
- J. Borén et al., J. Clin. Invest. 101,
2658 (1998).
[Medline]
- J. Borén et al., J. Clin. Invest. 101,
1084 (1998).
[Medline]
- R. W. Mahley, S. C. Rall Jr., Annu. Rev. Genomics Hum. Genet.
1, 507 (2000).
[Medline]
- C. Wilson, M. R. Wardell, K. H. Weisgraber, R. W. Mahley, D. A.
Agard, Science 252, 1817 (1991).
[Medline]
The author is at the Gladstone Institute of Cardiovascular Disease,
San Francisco, CA 94110, USA. E-mail:
tinnerarity@gladstone.ucsf.edu
Related articles in Science:
- Structure of the LDL Receptor Extracellular Domain at
Endosomal pH
- Gabby Rudenko, Lisa Henry, Keith Henderson, Konstantin Ichtchenko,
Michael S. Brown, Joseph L. Goldstein, and Johann Deisenhofer
Science 2002 298: 2353-2358. (in Research Articles)
[Abstract]
[Full Text]
Volume 298, Number 5602, Issue of 20 Dec 2002, pp. 2337-2339.
Copyright © 2002 by The American Association for the Advancement of
Science. All rights reserved.
|
