Download Apolipoprotein D interacts with the long-form leptin receptor: a hypothalamic function in the control of energy homeostasis

Apolipoprotein D interacts with the long-form leptin
receptor: a hypothalamic function in the control of
energy homeostasis1
ZHITONG LIU, GUO-QING CHANG, AND SARAH F. LEIBOWITZ2
The Rockefeller University, New York, New York 10021, USA
SPECIFIC AIM
2. Dietary fat stimulates Apo D expression in the
hypothalamus
To identify and clone genes that are expressed in the
hypothalamus and involved in the development of
obesity through the regulation of food intake and body
fat accrual.
To confirm that dietary fat stimulates the hypothalamic
mRNA level of Apo D, we measured Apo D mRNA in
the medial hypothalamus by quantitative RT-PCR in an
additional set of rats (n⫽5– 6/group) maintained for 3
wk on either a low-fat (10%), moderate-fat (30%), or
high-fat (60%) diet. The results demonstrate that the
relative (to actin) Apo D mRNA level increases significantly as dietary fat rises from 10% to 30% (⫹19%,
P⬍0.03) and even further in rats on a 60% fat diet
(⫹25%, P⬍0.001). This increase in dietary fat concentration and Apo D mRNA was accompanied by a
significant rise in circulating levels of leptin. Body fat
pad weights (retroperitoneal, inguinal, mesenteric, and
epididymal), as well as body weight and total daily
intake, were also elevated in the high-fat diet rats.
PRINCIPAL FINDINGS
1. Identification of Apo D
We used representational difference analysis (RDA) to
identify genes that exhibit increased expression in the
hypothalamus of rats maintained on a high-fat diet,
which is known to cause obesity and stimulate hypothalamic expression of peptides involved in energy balance. We then investigated whether any of these RDA
clones encode proteins that interact with the long-form
receptor of leptin, which controls food intake and body
weight and is stimulated by a high-fat diet. We performed this experiment by screening a yeast two-hybrid
library and searching the resultant clones for DNA
sequences identical to those generated by RDA.
In the RDA experiment, the cDNA fragments prepared from the medial hypothalamus of rats (SpragueDawley, male, n⫽10/group) maintained for 3 wk on a
low-fat diet (10% fat, 3.75 Kcal/g) were subtracted
from those of the rats on a high-fat diet (60%, 5.10
Kcal/g) diet, and the resultant cDNA fragments were
amplified by PCR. After three rounds of subtraction
and amplification, distinct DNA bands were obtained in
agarose gel. The subsequent cloning and sequencing of
these DNA fragments (53 clones) revealed a clone
containing a 0.5 kb cDNA fragment of Apo D. In a
GAL4 yeast two-hybrid system, the cytoplasmic domain
immediately following the transmembrane region of rat
Ob-Rb, hereafter referred as Ob-Rbc (for carboxyl
terminal), was used as the bait to screen a rat brain
cDNA library (2⫻106 yeast colonies). Sequencing of 57
positive clones from an X-GAL filter assay revealed a
clone that contained a 0.9 kb amino-terminal truncated
cDNA fragment of Apo D.
0892-6638/01/0015-1329 © FASEB
3. Apo D interacts with Ob-Rb
To confirm the binding between Apo D and Ob-Rbc,
we tested their interaction in a different LexA yeast
two-hybrid system. In these experiments, only the yeast
harboring Apo D and Ob-Rbc fusion proteins grew, and
the resultant colonies turned blue within 3 days on the
test medium, demonstrating that Apo D interacts specifically with Ob-Rbc.
We confirmed this interaction with purified proteins
by protein-to-protein interaction experiments in vitro.
Apo D and Ob-Rbc were expressed in bacteria as a GST
fusion protein and a thioredoxin (Trx) fusion protein,
respectively, and purified. It was found that 20 ␮g of
Trx䡠Ob-Rbc coprecipitated with 1 ␮g of GST 䡠 Apo D,
but not with 1 ␮g of GST when GST and GST 䡠 Apo D
were precipitated with glutathione agarose beads. Reciprocally, 1 ␮g of GST 䡠 Apo D coprecipitated with 20
␮g of Trx 䡠 Ob-Rbc, but not with 20 ␮g of Trx, when Trx
and Trx 䡠 Ob-Rbc were precipitated with Ni-NTA SuperflowTM resin.
1
To read the full text of this article, go to http://www.
fasebj.org/cgi/doi/10.1096/fj.00 – 0530fje; to cite this article,
use FASEB J. (February 22, 2001) 10.1096/fj.00 – 0530-fje
2
Correspondence: The Rockefeller University, Box 278,
1230 York Ave., New York, NY 10021, USA. E-mail: leibow@
rockvax.rockefeller.edu
1329
5. Apo D and Ob-Rb are coexpressed in hypothalamic
neurons
Figure 1. Protein-to-protein interaction in vivo. Ob-R and its
associated proteins were precipitated by a goat anti-Ob-R
antibody (lane 1) from protein extracts made from pooled rat
hypothalamus at 4°C overnight. After washing 4 times, proteins were separated in a 10% polyacrylamide gel and assayed
for Apo D by a monoclonal anti-human Apo D antibody. Apo
D was not detected in a mock precipitation (lane 2) conducted by using normal goat IgG in the precipitation step.
By using a monoclonal anti-human Apo D antibody, we
also observed the existence of Apo D protein in neurons of hypothalamic nuclei in immunohistochemical
experiments. As shown in Fig. 2 (top panel), the
immunoreactivity for Apo D is evident in the cytoplasm
of neurons, both parvocellular and magnocellular, of
the paraventricular nucleus and in small neurons of the
arcuate nucleus. This immunoreactivity is specific to
Apo D protein, since no signal was generated when the
above antibody was preabsorbed by purified human
Apo D (not shown). Ob-R immunoreactivity is also
present in neurons of the paraventricular and arcuate
nuclei, as revealed by using a polyclonal anti-Ob-R
antibody (Fig. 2, middle panel). In this double-labeling
experiment, we found that the proteins of both Apo D
and Ob-R are clearly colocalized in the same neurons of
these hypothalamic nuclei (Fig. 2, lower panel). This
coexistence is also observed in other areas, including
4. Apo D does not interact with Ob-Ra
Since the mutation of Ob-Rb into the natural short
form, Ob-Ra, results in obesity in C57BL/ks db-/dbmice, it was interesting to investigate whether Apo D
also interacts with Ob-Ra. We expressed the cytoplasmic
domain of rat Ob-Ra (Ob-Rac) as a Trx fusion protein
(Trx 䡠 Ob-Rac) in bacteria and purified it. In a proteinto-protein interaction experiment, 1 ␮g of GST 䡠 Apo D
did not coprecipitate with 20 ␮g of Trx 䡠 Ob-Rac when
Trx 䡠 Ob-Rac was precipitated by Ni-NTA SuperflowTM
resin. Apo D did not interact with Ob-Rac in the LexA
yeast two-hybrid system. Therefore, through independent approaches, we have demonstrated that Apo D
fails to interact with Ob-Ra.
The above experiments indicate that the amino acid
sequence responsible for the interaction with Apo D is
present in Ob-Rbc but not Ob-Rac. To confirm this, we
generated a truncated Ob-Rbc (Ob-Rbt) by removing a
stretch of sequence at the NH2 terminus of Ob-Rbc. We
found that 20 ␮g of Trx 䡠 Ob-Rbt coprecipitated with 1
␮g of GST 䡠 Apo D but not with 1 ␮g of GST when GST
and GST 䡠 Apo D were precipitated with glutathione
agarose beads. Reciprocally, 1 ␮g of GST 䡠 Apo D coprecipitated with 20 ␮g of Trx 䡠 Ob-Rbt, but not with 20 ␮g
Trx 䡠 Ob-Rac, when the Trx fusion proteins were precipitated by Ni-NTA SuperflowTM resin. These experiments indicate that Ob-Rbt is sufficient for the interaction between Apo D and Ob-Rbc. However, the amino
acids or motifs on both Apo D and Ob-Rbt involved in
this interaction remain to be determined.
We also demonstrated the interaction between Apo
D and Ob-R in vivo in immunoprecipitation experiments with proteins extracted from pooled rat hypothalamus (Fig. 1). A single band of ⬃30 kDa, representing the glycosylated Apo D, was detected (Fig. 1, lane
1). In contrast, a mock precipitation generated no
signal (Fig. 1, lane 2).
1330
Vol. 15
May 2001
Figure 2. Colocalization of Apo D and Ob-R protein in
hypothalamic nuclei. The monoclonal anti-human Apo D
antibody and goat polyclonal anti-Ob-R antibody were used.
The immunoreactivity was detected by fluorescein- and cyanine-3-conjugated secondary antibodies, respectively, and examined under a confocal microscope. Apo D and Ob-R
protein are observed in the same neurons of the paraventricular (PVN) and arcuate (ARC) nuclei of the hypothalamus.
The FASEB Journal
LIU ET AL.
the hypothalamic supraoptic nucleus, cortex, and choroid plexus (not shown).
6. Hypothalamic Apo D mRNA levels correlate
positively with body fat and circulating leptin
We used quantitative RT-PCR to determine relative Apo
D mRNA levels in the medial hypothalamus of rats
exhibiting differential body fat accrual on a high-fat
diet. In this experiment, Sprague-Dawley rats were fed
ad libitum on a high-fat diet for 3 wk. Based on the
amount of body fat accumulated (measured in four
dissected depots) over this period, these subjects were
then divided into two subgroups, referred to as ‘lean’
(n⫽7) with 15–21 g body fat or ‘obese’ (n⫽8) with
26 –34 g body fat. Whereas both groups were similar in
their total daily intake, the obese rats with ⬃50%
greater body fat had significantly higher Apo D mRNA
levels in the medial hypothalamus (1.57⫾0.05 vs.
1.38⫾0.04, P⬍0.05). They also had considerably higher
levels of circulating leptin (21.5⫾3.1 vs. 9.8⫾2.0,
P⬍0.05). Across the whole group, Apo D mRNA was
found to be strongly, positively correlated with total
body fat (r⫽⫹0.87, P⬍0.01), body weight (r⫽⫹0.82,
P⬍0.01), and the level of leptin (r⫽⫹0.76, P⬍0.01),
which in turn was positively related to adiposity
(r⫽⫹0.87, P⬍0.01). Similar results were obtained in
inbred mouse strains (SWR/J and AKR/J) with a differential propensity toward obesity.
7. Apo D mRNA levels are reduced in medial
hypothalamus of ob-/ob- and db-/db- mice
We further investigated whether Apo D remains positively associated with body fat mass in mouse strains
that are obese due to a mutated leptin or Ob-Rb gene.
We first examined Apo D expression in C57BL/3j
db-/db- mice using quantitative RT-PCR. Due to their
mutational loss of the cytoplasmic portion of Ob-Rb,
Figure 3. Hypothesis. Dietary fat stimulates Apo D expression
of in hypothalamic neurons as well as leptin production.
Through its interaction with the carboxyl terminus of activated Ob-Rb, Apo D may be stimulated, probably resulting in
a conformational change. This stimulation may enable Apo D
to bind a putative ligand, which is thought to be generated
through the activation of Ob-Rb by leptin. Apo D may
transport this ligand as a paracrine signal within particular
hypothalamic nuclei or carry this ligand as a hormone to
specific locations of the body through circulation.
APO D INTERACTS WITH OB-RB
these mice presumably do not support an interaction between Apo D and the mutant Ob-Rb. The level
of medial hypothalamic Apo D mRNA in C57BL/3j
db-/db- mice was found to be considerably reduced
relative to that of the wild-type mice (0.69⫾0.01 vs.
0.88⫾0.01, P⬍0.05). To investigate the influence of
the loss of leptin itself on Apo D expression, we
compared hypothalamic Apo D mRNA in obese
C57BL/6j ob-/ob- mice to that of the lean wild-type
C57BL/6j mice. With a mutant leptin but intact Ob-Rb
in C57BL/6j ob-/ob- mice, Ob-Rb could not be stimulated, even though it may still interact with Apo D.
Similar to the result in db-/db- mice, the Apo D mRNA
level was 30% lower in the C57BL/6j ob-/ob- mice
compared to their lean wild-type littermates (0.71⫾0.01
vs. 1.06⫾0.01, P⬍0.05). These findings, indicating that a
functional leptin/Ob-Rb signaling process is required
for the up-regulation of Apo D expression, support a
possible role for hypothalamic Apo D in the control of
body fat accrual.
CONCLUSIONS AND SIGNIFICANCE
We provide strong evidence suggesting that Apo D
interacts with Ob-Rb, but not Ob-Ra, in hypothalamic
neurons in vivo. We have also found that fat intake
significantly stimulates hypothalamic expression of Apo
D, which in turn is strongly, positively correlated with
body fat pad weights and circulating levels of leptin.
However, the mutant ob-/ob- and db-/db- mice have
considerably reduced levels of hypothalamic Apo D
mRNA compared to their lean wild-type littermates.
These findings indicate that the leptin/Ob-Rb signaling pathway modulates Apo D expression in hypothalamic neurons. We propose that the reduced expression of hypothalamic Apo D in ob-/ob- and db-/db- mice,
possibly resulting in a deficiency of Apo D protein,
contributes to the development of the obesity, particularly on a high-fat diet.
Apo D belongs to the lipocalin protein family, the
members of which bind and transport small hydrophobic
ligands. However, the specific ligand to which Apo D
binds has not been unequivocally identified. Published
evidence suggests that Apo D may bind to multiple
ligands, and each ligand is specific to the tissue or cell type
where Apo D is expressed. We thus speculate that Apo D
may be activated through its interaction with a leptinstimulated Ob-Rb and may bind a specific ligand in
hypothalamic neurons, where it exerts signaling functions. This ligand may be produced after leptin stimulation and serve as a paracrine signal within particular
hypothalamic nuclei or a hormone that enters the circulation. Apo D may thus play a contributing role in the
regulation of body fat accrual. This proposed function of
hypothalamic Apo D is consistent with a previous report
that a Taq I Apo D polymorphism is linked to obesity and
hyperinsulinemia, as well as to noninsulin-dependent
diabetes mellitus, a condition commonly associated with
obesity in animals and humans.
1331