Pleiotrophin inhibits HIV infection by binding the cell
surface-expressed nucleolin
Elias A. Said
1
, Jose
´Courty
2
, Josette Svab
1
, Jean Delbe
´
2
, Bernard Krust
1
and Ara G. Hovanessian
1
1 UPR 2228 CNRS, UFR Biome
´dicale des Saints-Pe
`res, Paris, France
2 Laboratoire de Recherche sur la Croissance Cellulaire, la Re
´paration et la Re
´ge
´ne
´ration Tissulaires (CRRET), FRE CNRS 2412,
Universite
´Paris Val de Marne, Cre
´teil, France
The human immunodeficiency virus (HIV) infects tar-
get cells by the capacity of its envelope glycoproteins
gp120-gp41 to attach cells and induce the fusion of
virus to cell membranes, a process which leads to virus
entry. The receptor complex essential for HIV entry
consists of the CD4 molecule and at least one of the
members of the chemokine receptor family: CCR5 or
CXCR4 [1,2]. Contrary to the virus entry process, the
attachment of HIV particles to cells can occur even
independently of CD4. We have previously demonstra-
ted that HIV attachment is inhibited by the pseudo-
peptide HB-19 that binds specifically the C-terminal tail
of nucleolin, a cell-surface-expressed protein identified
to be implicated in HIV attachment [3–5]. Conse-
quently, we have suggested that HIV attachment is
achieved by the coordination of at least two events
implicating on the one hand heparan sulfate proteo-
glycans [6,7] and on the other hand the cell surface-
expressed nucleolin [4]. In the search for natural
ligands of nucleolin that exhibit a potential inhibitory
activity against HIV infection, other than midkine [8]
and lactoferrin [9], here we show that pleiotrophin
Keywords
binding; HIV; pleiotrophin; receptor; surface
nucleolin
Correspondence
E. A. Said, UPR 2228 CNRS, UFR
Biome
´dicale des Saints-Pe
`res; 45 rue des
Saint-Pe
`res, 75270 Paris Cedex 06, France
Fax: +33 142862042
Tel: +33 142864136
E-mail: elias.said@umontreal.ca
(Received 11 May 2005, revised 30 June
2005, accepted 18 July 2005)
doi:10.1111/j.1742-4658.2005.04870.x
The growth factor pleiotrophin (PTN) has been reported to bind heparan
sulfate and nucleolin, two components of the cell surface implicated in the
attachment of HIV-1 particles to cells. Here we show that PTN inhibits
HIV-1 infection by its capacity to inhibit HIV-1 particle attachment to the
surface of permissive cells. The b-sheet domains of PTN appear to be
implicated in this inhibitory effect on the HIV infection, in particular the
domain containing amino acids 60–110. PTN binding to the cell surface is
mediated by high and low affinity binding sites. Other inhibitors of HIV
attachment known to bind specifically surface expressed nucleolin, such as
the pseudopeptide HB-19 and the cytokine midkine prevent the binding of
PTN to its low affinity-binding site. Confocal immunofluorescence laser
microscopy revealed that the cross-linking of surface-bound PTN with a
specific antibody results in the clustering of cell surface-expressed nucleolin
and the colocalization of both PTN and nucleolin signals. Following its
binding to surface-nucleolin, PTN is internalized by a temperature sensitive
mechanism, a process which is inhibited by HB-19 and is independent of
heparan and chondroitin sulfate proteoglycans. Nevertheless, proteoglycans
might play a role in the concentration of PTN on the cell surface for a
more efficient interaction with nucleolin. Our results demonstrate for the
first time that PTN inhibits HIV infection and suggest that the cell surface-
expressed nucleolin is a low affinity receptor for PTN binding to cells and
it is also implicated in PTN entry into cells by an active process.
Abbreviations
ALK, anaplastic lymphoma kinase; AZT, azidothymidine; CHO, Chinese hamster ovary; HARP, heparin affin regulatory peptide; HB-GAM,
heparin-binding growth-associated molecule; HBNF, heparin-binding neurite-promoting factor; MK, midkine; PTN, pleiotrophin; RPTP,
receptor-type tyrosine phosphatase.
4646 FEBS Journal 272 (2005) 4646–4659 ª2005 FEBS
(PTN) that binds surface nucleolin inhibits HIV
attachment to cells by its capacity to bind surface
nucleolin as a low affinity receptor.
PTN is an 18-kDa protein which was first identified
as a heparin-binding protein that progresses mitogenic
activity in rat and mouse fibroblasts [10]. The first
purification was from bovine uterus and neonatal rat,
brain, bone and kidney. PTN is rich in basic amino
acids especially lysine in both N- and C-terminal tails.
It was also named as heparin-binding-growth-associ-
ated molecule (HB-GAM) [11], heparin-binding neur-
ite-promoting factor (HBNF) [12] or heparin affin
regulatory peptide (HARP) [13].
Biological functions of PTN are mitogenic, angio-
genic and oncogenic activities, cell motility, differenti-
ation, and synaptic plasticity [14]. Elevated serum PTN
levels have been detected in patients with testicular,
pancreatic, colon, breast and other cancers [15–17].
Consequently, the circulating levels of PTN have been
proposed to serve as a tumor marker. Interestingly,
PTN is expressed in fracture healing [18] and its gene
expression is also upregulated in newly forming blood
vessels, in OX42-positive macrophages, and in inva-
sion-independent pathways of blood-borne metastasis
[14,19,20]. PTN is also expressed in adults with inflam-
matory diseases, and proinflammatory cytokines
enhance its expression [21,22]. Finally, PTN inhibits
infectivity of human herpes simplex viruses type 1 and
2 and human cytomegalovirus [23].
Several cell surface components have been reported
as potential receptors for PTN, such as the heparan
sulfate proteoglycans of N-syndecan [24] and the chon-
droitin sulfate proteoglycan of receptor-type tyrosine
phosphatase b⁄f(RPTP b⁄f) [25,26]. In addition, ana-
plastic lymphoma kinase (ALK) has been reported to
be a receptor that transduces PTN-mediated signals
and the PTN-ALK axis can play a significant role
during development and disease processes [27]. PTN
binds the extracellular domain of ALK with a K
d
of
32 ± 9 pm[27].
PTN shows a striking structural homology with
another heparin binding growth-associated factor
called midkine, with whom it shares 45% sequence
identity [14,28–31]. Therefore, like PTN, the binding of
midkine to heparan sulfate and chondroitin sulfate
proteoglycans could be clearly demonstrated using
purified and soluble components [32,33]. Midkine
binds also ALK with a high affinity and this binding is
inhibited by PTN [34]. We have previously demonstra-
ted that midkine is a cytokine that binds the cell sur-
face expressed nucleolin as a low affinity receptor.
Synthetic and recombinant preparations of midkine
inhibited in a dose-dependent manner infection of cells
by various HIV-1 isolates; this inhibition is due to the
capacity of midkine to bind cells specifically and to
prevent the attachment of HIV particles to cells [5,35].
Nucleolin is a component of the cell surface which
could act as a receptor for various ligands. Indeed, on
the cell surface nucleolin interacts with several mole-
cules such as lipoproteins, J factor [50], and the alpha-1
chain of laminin [51]. Cell surface-expressed nucleolin
could also act as a receptor of viruses such as
Coxsackie B [52] and Human Parainfluenza Virus type
3 [53]. Indeed, while nucleolin does not have a hydro-
phobic domain [54], it is expressed on the cell surface,
and it represents 20% of the cytoplasmic portion of
nucleolin [55]. Our previous results showed that cyto-
plasmic nucleolin is found in small vesicles that appear
to translocate nucleolin to the cell surface. Transloca-
tion of nucleolin is markedly reduced at low tempera-
ture or in serum-free medium, whereas conventional
inhibitors of intracellular glycoprotein transport have
no effect. Thus, translocation of nucleolin is the conse-
quence of an active transport by a pathway which is
independent of the endoplasmic reticulum ⁄Golgi com-
plex [55].
Here, we show that PTN inhibits HIV infection by
binding the cell-surface expressed nucleolin leading to
the inhibition of HIV-attachment to the cell surface.
PTN binding to cells involves high and low affinity-
binding sites even in cells that are deficient for the
expression of both heparan and chondroitin sulfate
proteoglycans. The synthetic ligand of nucleolin, the
HB-19 pseudopeptide, prevents the binding of PTN
to the low affinity receptor, thus suggesting that such
a receptor is the cell surface-expressed nucleolin.
Accordingly, by confocal immunofluorescence laser
microscopy, we show that cell-surface bound PTN
colocalizes with that of surface-expressed nucleolin.
Results
Inhibition of HIV-infection by PTN
We investigated the inhibitory effect of PTN on HIV
infection by using the experimental model of HeLa P4
or HeLa P4C5 cells. HIV entry and replication in
these cells result in Tat-mediated transactivation of
HIV LTR, leading to the expression of the LacZ
gene. Consequently, the b-galactosidase activity could
be measured in cell extracts to monitor HIV entry.
The b-galactosidase expression in noninfected cells is
considered as the background value in this experi-
ment. As we had shown previously, midkine inhibited
HeLa P4 cells infection by HIV-1 LAI isolate with
more than 90% inhibition at 1 lmof midkine [8,35].
E. A. Said et al.Nucleolin is a low affinity receptor of pleiotrophin
FEBS Journal 272 (2005) 4646–4659 ª2005 FEBS 4647
In this model, PTN inhibited the entry of the X4
HIV-1 LAI isolate in a dose-dependent manner with
IC
50
and IC
90
values of 60 and 250 nm, respectively
(Fig. 1A, HIV-1 LAI). PTN also inhibited infection of
HeLa P4C5 cells by the R5 HIV-1 Ba-L isolate in a
dose-dependent manner with IC
50
and IC
90
values of
60 nmand 500 nm, respectively (Fig. 1A, HIV-1
Ba-L). A similar inhibition profile was obtained with
the infection of MT4 cells (data not shown). HeLa P4
cells preincubated with PTN at 20 C for 45 min and
washed with medium to remove unbound PTN, resis-
ted HIV-1 LAI infection. However, incubation of
HIV-1 LAI with PTN and centrifugation at 100 000 g
to pellet the virus gave an HIV pellet that was still
infectious (data not shown). These data indicate that
the inhibitory effect of PTN is mediated through its
action on target cells rather through a direct effect on
virus particles.
The effect of PTN on the HIV attachment was moni-
tored by measuring the concentration of the HIV major
core protein p24 in the lysate of HeLa P4 cells
that were incubated with HIV-1 LAI at room tempera-
ture in the presence of different concentrations of
PTN. PTN inhibited HIV-attachment in a dose-
dependent manner with more 50% and 90% inhibition
at 50 and 250 nm, respectively (Fig. 1B). These results
demonstrate that the inhibition of HIV infection by
PTN is due to its inhibitory effect on the attachment of
HIV particles.
The inhibiting action of PTN on the HIV-1
infection is mediated through the b-sheet
domains of PTN
PTN consists of two b-sheet domains located between
N- and C-terminal tails rich in lysine residues [44]. In
order to locate the domain of PTN responsible of the
inhibitory effect on HIV infection, we tested the capa-
city of deletion constructs corresponding to various
domains of PTN to inhibit infection of HeLa P4 cells
HIV-1 LAI HIV-1 Ba-L
Control
A
B
Control
No HIV
25
50
100
200
250
AZT
Control
AZT
30
60
125
250
500
MK 1 µM
60
125
250
500
1000
0
0 50 100
0.5 1 1.5
β-Galactosidase Activity (OD)
2.52 0 0.5 1 1.5
β-Galactosidase Activity (OD)
2
PTN
[nM]
PTN
[nM]
PTN
[nM] Fig. 1. Inhibition of HIV infection by PTN.
(A) HeLa P4 cells were treated (30 min,
37 C) with midkine (MK) (1 lM) or PTN (60,
125, 250, 500, 1000 nM). HeLa P4C5 cells
were treated (30 min, 37 C) with PTN (30,
60, 125, 250, 500 nM). HeLa P4 and HeLa
P4C5 cells were then infected with the HIV-
1 LAI or HIV-1 Ba-L isolate, respectively. At
48 h postinfection, the b-galactosidase activ-
ity was measured in cell extracts directly to
monitor HIV entry (the abscissa; OD, optical
density). The histogram AZT represents the
background b-galactosidase activity when
HIV retrotranscription is inhibited. (B) HeLa
P4 cells were incubated (45 min, 20 C)
with PTN (25, 50, 100, 200, 250 nM) and the
HIV-1 LAI isolate. HIV attachment was
monitored by measuring the concentration
of the HIV major core protein p24 in cells
extracts. The histogram No HIV represents
the background of p24 concentration in
the absence of virus attachment. The
mean ± SD of triplicate samples is shown.
Nucleolin is a low affinity receptor of pleiotrophin E. A. Said et al.
4648 FEBS Journal 272 (2005) 4646–4659 ª2005 FEBS
by HIV-1 LAI. The peptide PTN Nt-tail corresponds
to the N-terminal tail of PTN (residues 1–8), the pep-
tide PTN Ct-tail corresponds to the C-terminal tail of
PTN (residues 110–136), PTN (residues 9–110) corres-
ponds to the b-sheet domains, PTN (residues 1–110)
corresponds to the N-terminal tail and the two
b-sheet domains, PTN (residues 9–136) corresponds
to the C-terminal tail and the two b-sheet domains,
PTN-Nf corresponds to the b-sheets on the N-ter-
minal side (residues 9–59), and PTN-Cf corresponds
to the b-sheets on the C-terminal side (residues
60–110).
Whereas lysine-rich peptides corresponding to the N
and C-terminal tails of PTN have no effect on HIV-1
infection, peptides containing the b-sheet domains
[PTN (1–110) and PTN (9–136)], or peptides contain-
ing the b-sheets alone [PTN (9–110)] inhibit HIV infec-
tion by a dose-dependent manner, at an IC
50
value of
200 and 250 nmfor PTN (1–110), and PTN (9–136),
respectively (Fig. 2). The most potent inhibitory effect
is observed with the peptide PTN (9–110) that inhibits
HIV-1 LAI infection with an IC
50
value of 30 nm.
Finally, PTN-Nf does not have an effect on HIV infec-
tion, whereas PTN-Cf inhibits the infection with an
IC
50
of 200 nm(Fig. 2). These results suggest that the
domains containing the b-sheets are the regions
responsible for the inhibitory effect of PTN on HIV
infection. The presence of N or C-terminal tails with
the two b-sheet domains (at residues 9–110) decreased
the inhibitory effect of the two b-sheet domains with-
out the respective tail (Fig. 2, compare the results
obtained with PTN 1–110 and PTN 9–136 with PTN
9–110). The presence of either one of the tails alone
might affect the proper folding of such truncated PTN
constructs and consequently affect the inhibitory effect
on HIV infection.
Inhibition of HIV particles attachment by PTN
requires a cell surface component other than
heparan and chondroitin sulfate proteoglycans
Described as a HB-GAM, PTN interacts with glyco-
aminoglycans such as heparan sulfate proteoglycans
[25,26], which are also implicated in HIV attachment
to the cell surface [7]. In order to investigate whether
the inhibitory effect of PTN on HIV attachment is
due to its interaction with heparan or chondroitin sul-
fate proteoglycans, we used Chinese hamster ovary
(CHO) wild-type cells (CHO K1) and mutant cells
lines that are deficient in the expression of heparan
sulfate (CHO 677), or both heparan ⁄chondroitin sul-
fate proteoglycans (CHO 618) [38,39]. Despite lacking
proteoglycan expression, these mutant cell lines
express similar levels of the cell-surface nucleolin [8].
In these HIV attachment experiments, culture super-
natants were removed from CHO K1, 677 and 618 cells
pretreated with PTN or the nucleolin-binding HB-19
pseudopeptide, before adding the virus preparation on
cells. The fact that CHO K1 cells do not express the
HIV receptor CD4 or the coreceptors CCR5 and
CXCR4, demonstrates that HIV attachment should
mainly be mediated via the heparan ⁄chondroitin sul-
fate proteoglycans and cell-surface expressed nucleolin
Control
PTN 200 nM
PTN (1-110)
PTN (9-136)
PTN (9-110)
PTN Nt-tail
PTN Ct-tail
PTN-Nf
PTN-Cf
100 nM
200 nM
500 nM
100 nM
200 nM
500 nM
100 nM
200 nM
500 nM
100 nM
200 nM
500 nM
100 nM
200 nM
500 nM
100 nM
200 nM
500 nM
100 nM
200 nM
500 nM
0 50 100 150
% of HIV infection
Fig. 2. The inhibitory of various PTN domains on HIV infection.
HeLa P4 cells were preincubated or not (30 min, 37 C) with
200 nMof PTN, PTN (1–110), PTN (9–136), PTN (9–110), PTN Nt-tail
(1–9), PTN Ct-tail (110–136), PTN-Nf, or PTN-Cf at 100, 200 and
500 nMconcentrations. Cells were then infected with HIV-1 LAI
(90 min, 37 C). The b-galactosidase activity was measured at 48 h
postinfection. The percentage HIV infection (abscissa) gives the
proportion of b-galactosidase activity compared to infected cells
without PTN (histogram Control).
E. A. Said et al.Nucleolin is a low affinity receptor of pleiotrophin
FEBS Journal 272 (2005) 4646–4659 ª2005 FEBS 4649
[4]. HB-19 was included in these binding experiments
in order to estimate the contribution of nucleolin in
the HIV attachment process. Accordingly, HB-19
markedly inhibited HIV attachment at a concentration
of 1 lmin all CHO wild-type and mutant cell lines,
thus indicating the capacity of surface expressed
nucleolin to serve as a receptor for HIV binding inde-
pendent of heparan and chondroitin sulfate proteogly-
cans (Fig. 3). Interestingly, PTN at a concentration of
500 nminhibited HIV attachment by about 70% to
the surface of all CHO cell lines used in this assay
(Fig. 3). In similar HIV attachment assays, when PTN
was not removed before addition of HIV, then more
than 90% inhibition of HIV attachment was observed
(data not shown). These results do not rule out a
potential role of heparan ⁄chondroitin sulfate proteo-
glycans in the inhibitory activity of PTN, but suggest
the implication of other cell surface component(s). It
should be noted that HIV attachment in control CHO
cell lines is decreased by 50 and 80% in CHO 677
and CHO 618 cells, respectively (Fig. 3). This decrease
is probably due to the lack of heparan sulfate and
heparan ⁄chondroitin sulfate proteoglycan expression,
and illustrates the implication of such proteoglycans
in the HIV attachment process. The capacity of
HB-19 to inhibit HIV attachment in the CHO wild-
type cells is in accord with our previous results using
various CD4 positive and HIV permissive cell lines;
such results confirm once again that HIV attachment
is coordinated by both proteoglycans and nucleolin
[4].
Role of heparan and chondroitin sulfate proteo-
glycans in PTN binding on the cell surface
To evaluate the potential implication of the heparan
and chondroitin sulfate proteoglycans in PTN binding,
we used the three CHO cell lines previously described
[8]. These binding experiments were carried out at
20 C to prevent PTN entry into cells (see Experimen-
tal procedures). We first investigated the specific and
nonspecific binding of
125
I-labeled PTN to the wild-
type CHO K1 cells, which express heparan and chon-
droitin sulfate proteoglycans by washing cells at 300
and 150 mmNaCl, respectively. In cells washed with
300 mmNaCl,
125
I-labeled PTN specific binding occurs
in a dose-dependent manner and reaches saturation at
3lmof
125
I-labeled PTN (Fig. 4A), whereas total
binding does not reach a saturation limit. Because of
the considerable amount of nonspecific binding, cells
were routinely washed at 300 mmNaCl in all the fol-
lowing experiments. It is important to note that PTN
specific binding resists drastic wash conditions such as
normal or acidic culture medium (pH ¼4) containing
(or not) 2 mNaCl (not shown). Interestingly, the
125
I-labeled PTN binding profile (binding curve and
saturation point) to heparan sulfate-deficient CHO 677
cells (not shown) and to both heparan and chondroitin
sulfate proteoglycan-deficient CHO 618 cells was
similar to that observed for the wild-type CHO K1
cells (Fig. 4B). However, the levels of
125
I-labeled PTN
binding (amount of
125
I-labeled PTN bound to cells)
to the CHO 618 and 677 cells was lower than that to
the wild-type CHO K1 cells. These results indicate that
under our experimental conditions, heparan and chon-
droitin sulfate proteoglycans may play a role in the
overall binding of PTN to cells, even if the specific
binding reaches saturation independently of their pres-
ence. As high affinity binding sites for PTN have been
reported in the literature [27], we investigated the pres-
ence of such sites on CHO cells. Indeed, results show
that in CHO K1 cell lines the specific binding of the
125
I-labeled PTN reaches saturation at 2 nm(Fig. 4C).
A similar saturation curve was obtained in CHO 618
and HeLa cells (not shown). Taken together, these
results suggest the presence of low affinity and high
affinity binding sites of PTN on the cell surface.
Scatchard analysis of the
125
I-labeled PTN binding
using high and low concentrations confirmed the pres-
ence of low affinity and high affinity binding sites. The
estimated K
d
value for PTN binding to the low affinity
binding site on CHO K1 and 618 cells was 1.3 ·10
)6
m
(3.6 ·10
7
sites per cell) and 1.4 ·10
)6
m(1.9 ·10
7
sites
per cell), respectively (Table 1). The estimated K
d
value
for the binding of PTN to the high affinity binding site
0
Control
HB-19 1 µM
PTN 0.5 µM
Control
HB-19 1 µM
PTN 0.5 µM
Control
HB-19 1 µM
PTN 0.5 µM
CHO K1
CHO 677
CHO 618
100 200 300 400 500
p24 [pg/ml]
Fig. 3. Attachment of HIV particles to CD4
–
CHO cell lines, expres-
sing or not expressing heparan ⁄chondroitin sulfate proteoglycans is
inhibited by PTN and the nucleolin-binding HB-19 pseudopeptide.
CHO K1 cells (wild type), CHO 677 cells (deficient in heparan sul-
fate proteoglycan) or CHO 618 cells (deficient heparan ⁄chondroitin
sulfate proteoglycans) were treated (30 min, 20 C) with HB-19
(1 lM) or PTN (500 nM). Both reagents were then removed from
the culture before incubation of cells with the HIV-1 LAI isolate
(45 min, 20 C). HIV attachment was monitored by measuring the
concentration of the HIV major core protein p24 in the cells lysate.
The mean ± SD of triplicate samples is shown.
Nucleolin is a low affinity receptor of pleiotrophin E. A. Said et al.
4650 FEBS Journal 272 (2005) 4646–4659 ª2005 FEBS