The Journal of Immunology
Processing of HEBP1 by Cathepsin D Gives Rise to F2L, the
Agonist of Formyl Peptide Receptor 3
Thalie Devosse,* Raphae ¨l Dutoit,†Isabelle Migeotte,* Patricia De Nadai,*
Virginie Imbault,* David Communi,* Isabelle Salmon,‡and Marc Parmentier*
is an acetylated 21-aa peptide corresponding with the N terminus of the intracellular heme-binding protein 1 (HEBP1). In the
current work, we have investigated which proteases were able to generate the F2L peptide from its precursor HEBP1. Structure–
function analysis of F2L identified three amino acids, G3, N7, and S8, as the most important for interaction of the peptide with
FPR3. We expressed a C-terminally His-tagged form of human HEBP1 in yeast and purified it to homogeneity. The purified
protein was used as substrate to identify proteases generating bioactive peptides for FPR3-expressing cells. A conditioned medium
from human monocyte-derived macrophages was able to generate bioactivity from HEBP1, and this activity was inhibited by
pepstatin A. Cathepsin D was characterized as the protease responsible for HEBP1 processing, and the bioactive product was
identified as F2L. We have therefore determined how F2L, the specific agonist of FPR3, is generated from the intracellular protein
HEBP1, although it is unknown in which compartment the processing by cathepsin D occurs in vivo.
2011, 187: 1475–1485.
erosclerosis (1), arthritis (2), and cancer (3). Therefore, inflam-
matory processes need to be well coordinated and regulated. An
optimal immune response depends in particular on specific leu-
kocyte subsets recruited to the sites of inflammation. Neutrophils,
macrophages, and dendritic cells (DCs) are important cellular
mediators of innate immune defenses (4). Numerous molecules
are able to elicit chemotactic responses, and they can be classified
into different chemical and structural families, among which are
chemokines (5), complement factors C3a and C5a (6), leuko-
trienes, and formylated peptides. These latest factors, such as the
prototypic fMLF, are among the first identified and most potent
chemoattractants for phagocytic leukocytes (7–9).
The Journal of Immunology,
nflammation is a critical response of the organism to
pathogens, damaged cells, or inflammatory stimuli. However,
uncontrolled inflammation can lead to diseases such as ath-
These various families of chemotactic factors activate G protein-
coupledseven transmembrane domains receptors expressed not only
by leukocytes but also by a variety of other cell types. A receptor for
formylated peptides was first described in 1976 (10), and the cDNA
encoding the human formyl peptide receptor (FPR, now called
FPR1) was cloned in 1990 (11, 12). Two closely related human re-
ceptors, FPR-like 1 (FPRL1/FPR2) and FPR-like 2 (FPRL2/FPR3),
were subsequently cloned by low-stringency hybridization, using
the FPR1 cDNA as a probe (13–15). The three genes are clustered
on human chromosome 19q13.3 and encode receptors sharing a
high level of sequence identity (13, 14). N-formylated peptides are
derived either from bacterial proteins or from endogenous mito-
chondrial proteins. FPR1, and presumably the two other members
of the family, are therefore believed to be involved in host defense
mechanisms against invading pathogens and also in the sensing
of internal danger signals resulting from cellular dysfunction. The
prototypic fMLF displays high affinity for FPR1 (10) and low af-
finity for FPR2 (15). In addition to formylated peptides, a number
of structurally diverse agonists of FPR1 and/or FPR2 have been de-
scribed during recent years (16). In contrast, FPR3 does not respond
to fMLF (17) and presents a distinctive expression pattern among
leukocyte populations. FPR3 is expressed in monocytes, myeloid
and plasmacytoid DCs, some tissue-specific macrophage subpop-
ulations (particularly in lung, skin, and colon) and eosinophils, but
not in neutrophils (18).
In contrast to FPR1 and FPR2, a very few ligands have been
identified for FPR3. Humanin, a neuroprotective peptide in models
of Alzheimer’s disease, was shown to bind FPR3 and FPR2 with
high affinity (19, 20). The most specific ligand described for FPR3
is, however, the endogenous peptide F2L (21), an acetylated 21-aa
peptide derived from heme-binding protein 1 (HEBP1). F2L was
isolated from porcine spleen extracts on the basis of its bioactivity.
This highly conserved peptide activates FPR3 in the low nano-
molar range, is poorly active on FPR2, and is inactive on FPR1.
On FPR3-expressing cells, F2L triggers intracellular calcium re-
lease, inhibition of cAMP accumulation, and phosphorylation of
ERK1/2 MAPKs through the Giclass of G proteins. When tested
*Institut de Recherche Interdisciplinaire en Biologie Humaine et Mole ´culaire, Uni-
versite ´ Libre de Bruxelles, Campus Erasme, B-1070 Brussels, Belgium;†Institut de
Recherche en Microbiologie Jean-Marie Wiame, B-1070 Brussels, Belgium; and
‡Service d’Anatomie Pathologique, Universite ´ Libre de Bruxelles, Campus Erasme,
B-1070 Brussels, Belgium
Received for publication October 25, 2010. Accepted for publication May 24, 2011.
This work was supported by the Actions de Recherche Concerte ´es of the Commu-
naute ´ Franc ¸aise de Belgique, the Interuniversity Attraction Poles Programme (P6-14)
Belgian State Belgian Science Policy, the Walloon Region (Programme d’excellence
“CIBLES”), the European Union (Grant LSHB-CT-2005-518167/INNOCHEM), the
Fonds de la Recherche Scientifique Me ´dicale of Belgium, and the Fondation Me ´d-
icale Reine Elisabeth. T.D. was an aspirant of the Belgian Fonds National de la
Recherche Scientifique and was also supported by Te ´le ´vie and the Alice and David
Van Buuren Foundation.
Address correspondence and reprint requests to Prof. Marc Parmentier, Institut de
Recherche Interdisciplinaire en Biologie Humaine et Mole ´culaire, Universite ´ Libre
de Bruxelles, Campus Erasme, 808 Route de Lennik, B-1070 Brussels, Belgium.
E-mail address: firstname.lastname@example.org
The online version of this article contains supplemental material.
Abbreviations used in this article: CTS D, cathepsin D; DC, dendritic cell; FPR,
formyl peptide receptor; FPRL1/FPR2, formyl peptide receptor-like 1; FPRL2/FPR3,
formyl peptide receptor-like 2; HEBP1, heme-binding protein 1; siRNA, small in-
on FPR3-expressing leukocytes, F2L promotes calcium mobili-
zation and chemotaxis (18, 21).
How and in which circumstances F2L is generated in the or-
ganism is, however, still unknown. HEBP1 is an intracellular
tetrapyrrole-binding protein of 22 kDa. Initially purified from
mouse liver, HEBP1 is expressed in many tissues. Knockdown of
the cell heme content, suggesting that HEBP1 may be involved in
heme regulation, biosynthesis, or transport (22, 23). However, no
additional data have reinforced this hypothesis, and the biological
function of HEBP1 remains therefore poorly defined. The three-
dimensional structure of murine p22HBP was determined by nu-
clear magnetic resonance and consists of a 9-stranded distorted
b-barrel flanked by two long a-helices (24, 25). Located outside
the globular structure of HEBP1, residues 1–17 are disordered,
whereas residues 18–23 form a b-strand. This part of the protein is
not found in bacterial homologues of the SOUL/HEBP family. It
is therefore conceivable that cleavage of HEBP1 after the leucine
21 would release the F2L ligand while keeping the heme-binding
domain functional (24).
In the current study, we investigated the potential pathways
leading to the generation of F2L in the organism. According to the
role of macrophages in cleaning up cellular debris and the reso-
lution ofinflammatory processes, we searched for the generation of
F2L when recombinant HEBP1 was submitted to the action of
macrophage proteases or conditioned media. The use of specific
protease inhibitors and purified proteases demonstrated that the
lysosomal aspartyl endopeptidase cathepsin D (CTS D) is able
to process HEBP1 and generate F2L. This observation suggests
that F2L might be generated after tissue damage and macrophage
recruitment, favoring the recruitment of additional monocyte/
macrophages and DCs, which would contribute to tissue repair
and the control of the inflammatory process.
Materials and Methods
Plasmid construction and yeast transformation
The cDNA corresponding to human HEBP1 was amplified by PCR from
pCDNA3–HEBP1 using the primer pair oecj301 and oecj302 and was
introduced in the pCSC2 Saccharomyces cerevisiae expression vector (26)
by homologous recombination. The resulting plasmid, pCSC2–HEBP1
(pCSC294), allows the production of recombinant HEBP1 displaying a
C-terminal tag of six histidines. The S. cerevisiae BY4709 strain (MATa
ura3D0) was transformed with the vector using the lithium acetate pro-
cedure (27). Transformants were selected on YNB plates containing 20
HEBP1 production and purification
Strain BY4709 producing HEBP1 was culturedon YNB containing 20 mg/ml
glucose in a 13 l-batch bioreactor (Biolafitte) to an OD660nmof 1.69. Cells
were harvested by centrifugation and the pellet washed twice in water. Cells
were resuspended in buffer A (300 mM NaCl, 50 mM NaH2PO4) with
EDTA-free Complete Protease Inhibitor Cocktail (Roche) and lysed with
a French press. The lysate was centrifuged at 12,000 rpm (Sorvall RC5B, SS-
34 rotor) for 30 min at 4˚C. The supernatant was purified on an Ni-NTA
Sepharose column (Qiagen) eluted by a step gradient of 0, 90, and 150 mM
imidazole in buffer A. The serine protease inhibitor PMSF (1 mM; Sigma)
was added to the collected fractions (4 ml/fraction). A sample of the fractions
was loaded on a 12% polyacrylamide gel, and HEBP1 was identified by
Western blotting using a rabbit polyclonal Ab (Phoenix). Fractions of interest
were finally pooled, concentrated to 1 ml, and purified on a Superdex 75
column (GE Healthcare) run with buffer A at 1 ml/min. Fractions were an-
alyzed on a 12% polyacrylamide gel, followed by HEBP1 immunodetection
on Western blots, and purity was checked by Coomassie blue staining.
Truncated synthetic peptides and alanine scanning
Acetylated F2L (Ac-MLGMIKNSLFGSVETWPWQVL) and alanine
variants were synthesized locally by using the solid-phase Fmoc strategy.
Monoisotopic masses and sequences of all peptides were verified by mass
spectrometry. Because of their hydrophobicity, peptides were dissolved in
DMSO at 1 mM, and 25-fold intermediate dilutions were made in 50%
CH3CN, followed by further dilution in assay buffer to working concen-
trations. All peptides were assayed from 0.1 to 3000 nM (2 points per log)
in the aequorin-based assay on FPR3-expressing cells, and the EC50was
determined. The results are presented as the ratio between the EC50of the
peptide and the EC50of native F2L.
Aequorin-based luminescence assay of intracellular calcium
Calcium release was measured by an aequorin-based bioluminescence
assay, as previously described (28, 29). In brief, CHO-K1 cells coex-
pressing apoaequorin, Ga16, and FPR3 or control GPCRs were collected
from culture dishes, pelleted by centrifugation, and resuspended at 5 3 106
cells/ml in DMEM/Ham’s F12 containing 0.1% BSA (aequorin buffer).
The cell suspension was supplemented with 5 mM coelenterazine h
(Promega, Madison, WI) and incubated under shaking for 3 h 30 min at
room temperature in the dark, then diluted 5-fold in aequorin buffer. Fifty
microliters of cell suspension was injected onto 50 ml of agonist-
containing medium in 96-well plates, and light emission was recorded
for 40 s in a Centro LB 960 luminometer (Berthold Technologies). ATP
(20 mM; Sigma) was used as standard to normalize the data.
For the analysis of truncated F2L peptides, a F2L variant containing an
additional C-terminal tyrosine was used as tracer after labeling with125I
using the Iodogen method (the sp. act. was 900 Ci/mmol), as described
previously (16). Afterwards, we obtained a custom-made fluorescent F2L
derivative, containing 5(6)-carboxyfluorescein linked to an additional
C-terminal lysine (F2L–FAM), from JPT Peptide Technologies (Berlin,
Germany). Two hundred thousand FPR3-expressing CHO-K1 cells or par-
ental CHO-K1 cells in 100 ml binding buffer (DMEM-F12, containing
0.5% BSA and 0.1% NaN3; Life Technologies) in duplicate samples were
incubated in siliconized 1.5-ml microcentrifuge tubes (Sigma) with in-
creasing concentrations of F2L–FAM for 1 h at room temperature in the
dark. The cells were then washed with 2 volumes of binding buffer, pel-
leted, resuspended in 250 ml binding buffer, and analyzed by FACS
(FACScan; Becton Dickinson). FACS data were analyzed with the WinMDI
software. Nonspecific binding was determined in the presence of 10 mM
unlabeled F2L. For competition binding assays, cells were incubated with
10 nM F2L–FAM and increasing concentrations of HEBP1, unlabeled F2L,
or variants thereof. The binding data were analyzed with the GraphPad
Proteolytic processing of HEBP1 in conditioned medium from
Monocytes were isolated from venous blood of healthy donors by im-
munomagnetic bead cell sorting (MACS) according to the manufac-
turer’s specifications. These procedures received authorization from the
Ethics Committee of the Free University of Brussels Medical Faculty.
After Ficoll density gradient, monocytes were purified by positive selec-
tion using CD14 microbeads (Miltenyi Biotec). Macrophages were dif-
ferentiated from monocytes in the presence of 50 ng/ml recombinant
human M-CSF (R&D Systems) for 6 to 8 d. The purity of the cell prepa-
rations was evaluated to 95% or more by flow cytometry (CD206+CD14+,
or CD68+for permeabilized cells). Macrophage monolayers were in-
cubated in a proteolysis buffer (25 mM sodium acetate pH 3.6, 100 mM
NaCl) at 37˚C in a humidified atmosphere of 5% CO2during 24 h. The
medium was collected and incubated with HEBP1 with or without 10 mg/
ml pepstatin A (Sigma) for different periods of time at 37˚C. The medium
was then adjusted to pH 7 by 10 mM sodium bicarbonate pH 8, and
samples were engaged in the aequorin-based assay.
Inhibition of CTS D expression
CTS D expression was inhibited in human monocyte-derived macro-
phages by transfection of specific small interfering RNAs (siRNAs; A,
HSS102578; B, HSS175648; C, HSS175649; Invitrogen), using their re-
spective scrambled oligonucleotides as controls. Oligonucleotides (0.1, 1,
and 10 nM) were transfected using the INTERFERin reagent (Polyplus-
transfection) at days 6 and 7 of macrophage differentiation from mono-
cytes, and the purity of fully differentiated macrophages at day 7
was evaluated by flow cytometry (CD68+CD206+cells). Transfection
efficiency was monitored with 100 nM FITC-labeled oligonucleotide
(BLOCK-iT Fluorescent Oligo; Invitrogen). CTS D levels were evaluated
48 h after the second transfection by Western blotting using a mouse anti-
1476 HEBP1 PROCESSING
CTS D mAb (clone BC011, IgG2a, 0.1 mg/ml; Calbiochem), a rabbit
polyclonal Ab against GAPDH as control (1/2000; Cell Signaling), and an
Odyssey infrared imaging system analysis.
Proteolysis of HEBP1 by CTS D
Native CTS D from human spleen was purchased from Calbiochem. HEBP1
was incubated in proteolysis buffer with CTS D at 37˚C with or without
pepstatin A. The medium was adjusted to pH 7 by sodium bicarbonate pH 8,
and samples were tested for activity in the aequorin-based assay. For mass
spectrometry analysis, the samples were vacuum dried and resuspended in 1
ml 5% CH3CN/0.1% trifluoroacetic acid mixed with matrix mix (2 mg/ml
2,5-dihydroxybenzoic acid, 10 mg/ml a-cyano-4-hydroxycinnamic acid, 2
mM fucose). Mass spectrometry analysis was performed on a Q-TOF Ultima
Global mass spectrometer equipped with a MALDI source. In parallel,
samples were vacuum-dried, boiled for 10 min in 13 Novex loading buffer
(Invitrogen), and loaded on 10–20% polyacrylamide gradient gels in Tricine
buffer (Invitrogen). Gels were blotted onto 0.2-mm pore-size polyvinyl
difluoride membranes (Millipore), and HEBP1 fragments were immuno-
detected with a rabbit polyclonal Ab directed at the N terminus of the protein
(Phoenix Pharmaceutical) or stained using an improved Blum’s silver
staining protocol modified for high-sensitivity protein identification (30).
Preparation of mouse neutrophils
Fresh blood samples were collected from posterior vena cava of 8-wk-old
C57BL/6 mice. Neutrophils were recovered on a Ficoll density gradient
(Lymphoprep; Axis-Schield), and erythrocytes were lysed by ammonium
chloride. The cell population was consistently composed of .90% neu-
trophils, as determined by flow cytometry analysis (CD11b+Gr1+pop-
Leukocytes were isolated as described earlier and incubated for 30 min in
RPMI 1640 containing 5% decomplemented FBS before running the che-
motaxis assays. HEBP1 (100 nM) was incubated for 30 min at 37˚C with 60
nM CTS D in saline buffer (25 mM acetate buffer, 150 mM NaCl, pH 3.6).
The medium was neutralized with 1 M sodium bicarbonate pH 8 before
testing. Chemotaxis was assayed by a modification of the Boyden micropore
filter technique, as previously described (31). Briefly, the assay was con-
ducted in 48-well microchemotaxis chambers (Neuroprobe), using 5-mm
(human monocytes and monocyte-derived macrophages) or 3-mm (mouse
neutrophils) pore-size Nuclepore track-etched polycarbonate membranes
(Whatman). The cell suspension (10,000 macrophages/well and 50,000
monocytes or neutrophils/well) was loaded in the upper chamber and the
chemoattractant solution or vehicle in the lower chamber. Positive control
of cell migration was 10 nM synthetic F2L. Migration was conducted for an
adequate period of time (1 h for monocytes, 1 h 30 min for macrophages,
and 30 min for neutrophils) at 37˚C in humidified air containing 5% CO2.
After membrane removal, nonmigrated cells were scraped off its upper side.
Then, the filters were fixed in methanol and stained with Hoechst for 2 min.
Micrographs of the lower surface of the filters were taken, and the number
of cells was counted with ImageJ software (version 1.36b). Results are
expressed as a chemotactic index (ratio of migrated cells in the presence
versus the absence of chemoattractant). Chemotaxis was distinguished from
chemokinesis by a checkerboard test, in which the chemoattractant was
added to the upper chamber with the cells or to both chambers.
Immunohistochemistry and tissue microarrays
Normal tissues obtained from 30 different human organs in the Service
d’Anatomie Pathologique of the Erasme Hospital were selected using
H&E-stained slides. Corresponding paraffin-embedded blocks were then
precisely aligned with the marked slides. Three tissue cores (0.6 mm in
diameter) for each sample were punched using a precision instrument
(Beecher) and arrayed into a recipient block. Each organ was represented
by samples from three to five different patients. These procedures received
authorization from the Ethics Committee of the Free University of Brussels
Medical Faculty. Sections (5 mm) of formalin-fixed and paraffin-embedded
tissue samples or tissue microarrays were subjected to standard immuno-
histochemical procedures. Primary Abs were a mouse monoclonal anti-
HEBP1 [aa 1–21] (clone 729, IgG1b; Euroscreen) and a mouse monoclonal
anti-CTS D (clone C5, IgG2b; AbD Serotec). Microwave Ag retrieval was
performed twice in citrate buffer, pH 6, for 5 min. Detection was made
by the Vectastain Elite ABC kit (Vector Laboratories) using 3,39-dia-
minobenzidine tetrahydrochloride (Dako) as the peroxidase substrate. The
slides were counterstained with hematoxylin, dehydrated, and mounted in
DPX mounting medium (Sigma-Aldrich). Negative controls were con-
ducted by replacing primary Abs with corresponding isotype controls
(Vector Laboratories). HEBP1 and CTS D expression were evaluated by
two independent observers and assessed by scoring each spot: 0 for no
expression; 1 for low expression; and 2 for high expression. Results are
presented as the score means for different donors.
For the results of aequorin-based and chemotaxis assays, statistical sig-
nificance was determined using the one-way ANOVA test, and p values
,0.05 were considered significant.
HEBP1 production and purification
F2L is to date the only specific ligand of FPR3. However, no in-
formation is available yet regarding the situations and mechanisms
leading to the generation of F2L in the organism, including the
proteolytic enzymes involved in the cleavage of HEBP1. To gain
insight into this process, we first expressed and purified full-length
recombinant human HEBP1.
The HEBP1 sequence starts with an acetylated methionine and
contains no apparent signal peptide. N-terminal acetylation occurs
in cytosolic mammalian proteins and yeast proteins but rarely in
prokaryotic or archeal proteins (32, 33). In a subset of these pro-
teins, acetylation is required for some of their biological properties,
such as enzymatic activity, stability, DNA binding, protein–protein
interaction, or peptide–receptor recognition. Other proteins are
acetylated without known functional consequences. In the case of
F2L, acetylation is not required for binding to FPR3 or activation of
the receptor (21). It might, however, be necessary to stabilize F2L
and/or HEBP1 in vivo, for the (essentially unknown) functions of
HEBP1, or for the recognition of the protein by the putative pro-
tease(s) generating F2L (34). For all these reasons, we produced
recombinant human HEBP1 in the eukaryotic cell S. cerevisiae, so
that it would be correctly folded and acetylated. As F2L is the N-
terminal peptide of HEBP1, we placed a poly-histidine tag for
purification purposes at the C-terminal end of the protein.
After selection of a yeast clone, the amount of HEBP1 evaluated
by Western blotting represented close to 1% of the total protein
extract (data not shown). The recombinant protein was purified by
immobilized metal ion affinity chromatography and gel filtration
(Fig. 1A–C). Size-exclusion chromatography showed that HEBP1
is monomeric. The purity of the purified protein was evaluated on
12% polyacrylamide gels after Coomassie blue staining (Fig. 1C).
Using this procedure, we produced and purified around 300 mg
HEBP1 per liter of yeast culture. The integrity of purified HEBP1
was evaluated by electrospray mass spectrometry. The major peak
(21,943.9 Da) corresponded with acetylated HEBP1 and a smaller
peak (21,901.9 Da) with nonacetylated HEBP1 (Fig. 1D, inset),
suggesting that most of the purified protein is indeed acetylated.
Characterization of the binding and functional activities of
HEBP1 on FPR3
The functional activity of purified full-length HEBP1 was evalu-
ated on FPR3-expressing CHO-K1 cells by using an aequorin-
based calcium mobilization assay. The activity was expressed as
a percentage of the response induced by 10 mM ATP, acting on
endogenous P2Y receptors. No significant activity was detected at
low nanomolar concentrations of HEBP1 compared with the ac-
tivity of F2L used as a positive control. A weak activity was,
however, identified for 1 mM HEBP1 (Fig. 2A).
We then evaluated the capacity of full-length HEBP1 to bind
FPR3-expressing CHO-K1 cells in competition binding assays
using a fluorescent F2L analogue as tracer. F2L labeled by car-
boxyfluorescein at its C terminus (F2L–FAM) was obtained and
characterized. A saturation binding curve was obtained on FPR3-
expressing cells, with a KDof 20 nM (n = 3), whereas no specific
The Journal of Immunology1477
binding was detected on parental CHO-K1 cells (Fig. 2B). In
competition binding experiments, unlabeled F2L competed for
F2L–FAM binding with an IC50of 17 nM, whereas HEBP1 did
not inhibit F2L–FAM binding (Fig. 2C).
We also investigated the potential antagonist activity of HEBP1
in a calcium mobilization assay. FPR3-expressing CHO-K1 cells
were stimulated with 50 nM F2L in the presence of various
concentrations of HEBP1. Full-length HEBP1 was unable to an-
tagonize the activity of F2L (Fig. 2D), and the weak agonist effect
of the protein was observed at high concentration (1 mM). In
agreement with this agonist activity, preincubation of cells with
high concentrations of HEBP1 for 20 min desensitized the cells,
which displayed decreased calcium release in response to 90 nM
F2L (Fig. 2E).
Identification of amino acids involved in F2L binding to FPR3
We investigated the structure–function relationship of F2L using
modified peptides derived from the F2L sequence. First, an alanine-
scanning experiment was devised, in which each amino acid of F2L
was replaced by an alanine (Fig. 2F). In a functional assay, we
identified three residues (G3, N7, and S8) as important for FPR3
activation. Indeed, replacement of these residues by alanine in-
creased significantly the EC50 of the peptides by 2- to 9-fold.
Concurrent replacement of K6, N7, and S8by alanines increased the
EC50by almost two logs, confirming the important role of these
amino acids (Fig. 2F, 2G). Some individual substitutions (such as
L2, I5, and F10) appeared to improve slightly the potency of the
peptide, but the combination of these substitutions did not confirm
this tendency (Fig. 2G). The apparent importance of glycine at
position 3 suggests that flexibility is required in the N-terminal
part of the peptide for its interaction with FPR3. Finally, none of
the residues in the C-terminal half of F2L, including some large
hydrophobic amino acids, appeared to play a critical role.
We further investigated the importance of specific residues of
F2L by testing N- and C-terminally truncated F2L variants in
binding and functional assays. At the N terminus, removal of the
(Fig. 2J). In contrast, a clear loss of potency was observed after
deletion of M4, very poor activation was obtained when deleting I5,
and the peptide starting at K7was totally inactive. In the binding
assay, competition for radioiodinated F2L was significantly im-
paired by deletion of the first three amino acids, and no compe-
tition was detected for shorter peptides (Fig. 2K).These data are
consistent with the important role of the structure surrounding G3
in the N terminus of the peptide.
At the C terminus, no differences in activation and binding
parameters were observed after deletion of the last three residues.
Peptides 1–17 and 1–16 displayed an EC50twice as high as that of
F2L. EC50values could still be determined for peptides 1–15 and
1–12, but their efficacy was decreased. The activity of peptides
1–10 and 1–9 was still detectable at high concentrations, although
EC50values were not measurable. Peptide 1–6 was totally inactive
(Fig. 2H, 2I). As for N-terminal truncations, the binding param-
eters were affected more rapidly with successive truncations.
None of the peptides tested behaved as an antagonist, as binding
and activation parameters decreased consistently in parallel. These
results, summarized in Supplemental Table I, fit with the presence
of a core of residues, necessary for binding and activation of
FPR3, in the N-terminal part of the F2L peptide.
Proteolysis of HEBP1 in conditioned medium from
To identify conditions allowing proteolysis of HEBP1 and F2L
release, we incubated purified HEBP1 in different media poten-
tially containing required proteases. Conditioned media were
fication of full-length HEBP1. A,
A lysate of yeast S. cerevisiae ex-
pressing HEBP1 C-terminally tag-
ged with six histidines was loaded
on a nickel column and eluted by
a step gradient of imidazole (dashed
line). The black box indicates the
fractions containing HEBP1. B, The
three fractions were concentrated and
separated on a Superdex 75 column
(1 ml/min). The third peak corre-
sponds with HEBP1 (black box). C,
Samples from different purification
steps were loaded on a 12% poly-
acrylamide gel stained with Coo-
massie blue. D, Electrospray mass
spectrometry profile of purified
Production and puri-
1478 HEBP1 PROCESSING
prepared from human monocyte-derived macrophages, human
neutrophils purified from peripheral blood, and purified mouse
spleen cells. Incubation of HEBP1 for 30 min at 37˚C in macro-
phage conditioned medium resulted in the generation of a bi-
ological activity on FPR3-expressing CHO-K1 cells using the
aequorin-based assay (Fig. 3A). Neither the conditioned medium
without HEBP1 nor HEBP1 in the unconditioned medium induced
calcium release. Weaker activities were recorded after incubation
of HEBP1 in conditioned media from human neutrophils or mouse
spleen cells (Fig. 3B).
To identify the class of proteolytic enzymes responsible for
HEBP1 processing, we investigated the effect of specific protease
inhibitors on the generation of FPR3 agonists in macrophage
conditioned medium. The activity was inhibited very efficiently
by pepstatin A (1 mg/ml), a potent inhibitor of aspartyl proteases.
The vehicle alone (10% ethanol, 1% acetic acid) and other
protease inhibitors (pMSF and leupeptin) had no effect (Fig. 3C
and data not shown). These data suggest that macrophages re-
lease an aspartyl protease able to cleave HEBP1 under acidic
alysis of F2L. A, Concentration–
action curves for the synthetic
F2L peptide and purified HEBP1 on
FPR3-expressing CHO-K1 cells us-
ing the aequorin-based calcium mo-
bilization assay. B, Saturation binding
assay using fluorescent F2L–FAM as
tracer on FPR3-expressing CHO-K1
or parental wild-type cells. Bound
tracer was determined on individual
cells by FACS and expressed as mean
fluorescence. Nonspecific binding
was determined in the presence of
10 mM F2L. C, Competition binding
assay on FPR3-expressing CHO-K1
cells using F2L–FAM as tracer and
unlabeled F2L or HEBP1 as com-
petitors. D, Intracellular calcium re-
lease was evaluated by the aequorin-
based assay on FPR3-expressing cells
in response to a constant concen-
tration of F2L (50 nM) and in-
creasing concentrations of HEBP1.
E, FPR3-expressing CHO-K1 cells
were preincubated for 20 min with
increasing concentrations of HEBP1
before testing their response to 90
nM F2L in the aequorin-based assay.
F, Each peptide of an alanine scan of
F2L was tested on FPR3-expressing
cells in the aequorin-based assay.
The results are presented as the ratio
between the EC50 of the mutant
peptide and that of F2L, as derived
from the dose-response curves. G,
Functional response of variant F2L
peptides with three alanine sub-
stitutions (as in F). H and J, Con-
centration–action curves on FPR3-
expressing CHO-K1 cells using the
aequorin-based assay. I and K,
Competition binding assay on FPR3-
expressing CHO-K1 cells using
[125I]F2L as tracer and C-terminally
(I) and N-terminally (K) truncated
variants of F2L as competitors.
Mean 6 SEM, n = 3–5. *p , 0.05,
**p , 0.01.
The Journal of Immunology1479
Proteolysis of HEBP1 by purified CTS D
CTS D, a lysosomal aspartyl protease, was considered as a good
candidate for the processing of HEBP1. Indeed, macrophages
contain high amounts of CTS D, and the activity of this enzyme is
strongly inhibited by pepstatin A. The concentration of CTS D in
the medium of macrophages cultured for 24 h in acetate buffer was
evaluated to 63.9 ng/ml (n = 3). Incubation of human HEBP1 (250
nM) with purified human CTS D (10 ng/ml) for 30 min at 37˚C
generated a strong biological activity for FPR3-expressing CHO-
K1 cells (Fig. 4A). The specificity of this activity was demon-
strated by the absence of response of parental CHO-K1 cells and
cell lines expressing FPR2, ChemR23, or GPR88 as control
receptors (Fig. 4B).
We then evaluated the biological activity resulting from the
incubation of HEBP1 (250 nM) with increasing concentrations of
CTS D and showed, in the aequorin-based assay, a bell-shaped
curve with a maximum corresponding with 60 nM CTS D (Fig.
4C). In other experiments, a 15-min incubation of 60 nM CTS D
with increasing concentrations of HEBP1 (#2 mM) resulted in
a saturation curve with typical Michaelis–Menten kinetics (Km:
148 nM) (Fig. 4D). Finally, we performed a 24-h time course of
the bioactivity generated from HEBP1 (250 nM) by CTS D (60
nM). Activity was observed at early time points (5 min) of hy-
drolysis with a peak at 30 min, after which it slowly decreased
with time. The same samples were loaded on 12% polyacrylamide
gels, showing that most of the substrate was cleaved after 5 min,
whereas slower degradation was observed afterward (Fig. 4E, 4F).
Altogether, these results demonstrate a fast cleavage of HEBP1 by
CTS D, generating an agonist of FPR3, which can be further
degraded by the enzyme, although with slower kinetics.
(250 nM) was incubated with CTS D (60 nM) and
tested on FPR3-expressing CHO-K1 cells using the
aequorin-based assay. Acetate buffer, CTS D, or HEBP1
alone did not activate FPR3. B, CTS D-treated HEBP1
was also tested on parental CHO-K1 cells or CHO-K1
cell lines expressing other receptors, ChemR23 and
GPR88. C, HEBP1 (250 nM) was treated with in-
creasing concentrations of CTS D for 15 min and tested
on FPR3-expressing cells. The functional response
peaked at 60 nM CTS D. D, Increasing concentrations
of HEBP1 were incubated with 60 nM CTS D for 15
min and tested in the aequorin-based assay. The activity
follows a typical Michaelis–Menten saturation curve. E
and F, HEBP1 (250 nM) was incubated with 60 nM
CTS D for a time course of 24 h. The FPR3-stimulatory
activity was evaluated in the aequorin based-assay (E)
and HEBP1 immunoreactivity by Western blotting (F).
All incubations were performed at 37˚C in a heating
bath. The data displayed (mean 6 SEM) are repre-
sentative of at least three independent experiments.
***p , 0.001.
HEBP1 processing by CTS D. A, HEBP1
macrophages. After incubation, the activity was diluted four times and measured in the aequorin-based assay using FPR3-expressing CHO-K1 cells. No
significant activity was observed for the medium alone (acetate buffer pH 3.6), HEBP1 in medium, or macrophage conditioned medium in the absence of
HEBP1. B, HEBP1 (1 mM) was incubated for 30 min in conditioned media from macrophages, neutrophils, and mouse spleen cells, and the activity was
evaluated in the aequorin-based assay on FPR3-expressing CHO-K1 cells. The results are presented as the ratio between the activity observed in the
presence of HEBP1 and the activity observed in the absence of HEBP1. C, HEBP1 (1 mM) was incubated for 30 min in conditioned medium from
macrophages, with or without pepstatin A (10 mg/ml). The vehicle of pepstatin A (0.01% acetic acid, 0.09% ethanol) did not modify the response. Mean 6
SEM, n = 3. ***p , 0.001.
Proteolytic processing of HEBP1 in conditioned medium. A, HEBP1 (1 mM) was incubated for 30 min with conditioned medium from
1480 HEBP1 PROCESSING
to HEBP1 processing in macrophage conditioned medium, we
inhibited CTS D expression by transfection of three specific
siRNAs during the monocyte-to-macrophage differentiation. The
procedure did not affect differentiation, as macrophages appeared
fully differentiated 7 d after monocyte selection (Fig. 5A), and
transfection efficiency was shown to be near 100% using a fluo-
rescent oligonucleotide as control (Fig. 5B, fluorescence micros-
copy, 5C, flow cytometry). The three different siRNAs (A, B, and
C) significantly reduced the levels of immunoreactive CTS D in
monocyte-derived macrophages, as determined by Western blot-
ting (Fig. 5D). In contrast, the corresponding control oligonu-
cleotides (scrambled sequences) did not affect CTS D immuno-
reactivity. Pretreatment of monocyte-derived macrophages with
the three siRNAs prevented the hydrolysis of HEBP1 and the
generation of FPR3 agonist in conditioned media prepared from
these cells, as evaluated with the aequorin-based assay on FPR3-
expressing CHO-K1 cells (Fig. 5E).
Characterization of the peptides resulting from the processing
of HEBP1 by CTS D
We analyzed by mass spectrometry the nature of the peptides
resulting from the processing of HEBP1 by CTS D to determine
which proteolytic product(s) activate FPR3. In the absence of
tryptic cleavage, we identified in the HEBP1 hydrolysate a pep-
tide corresponding with acetylated F2L. Silver staining of poly-
acrylamide gels showed that several fragments were generated
from HEBP1 after 30 min of digestion by CTS D. The molecular
weights of these fragments were evaluated to 2.5, 5, 6, 10, and 22
kDa (Fig. 6A). Each band was excised from the gel, trypsinized,
and peptides were identified by mass spectrometry. The 2.5- and
6-kDa peptides contained the N terminus of HEBP1, whereas all
other fragments did not. Analysis of the peptides identified two
main cleavage sites in the protein: one after Leu21and the other
somewhere between residues 60 and 90 (Fig. 6C). We did not
recover enough material to perform sequencing and identify more
precisely this second cleavage site. A band around 12 kDa was
found for some preparations of HEBP1 in the absence of CTS D.
This band did not correspond with a fragment of HEBP1, and we
could not identify this yeast contaminant by mass spectrometry.
The same samples were loaded on a 10–20% polyacrylamide gel
and blotted onto 0.2-mm pore-size polyvinyl difluoride mem-
branes. Immunodetection with an anti-F2L Ab (Euroscreen) iden-
tified, among others, a diffuse band at 2.5 kDa (Fig. 6B). Cleavage
leading to F2L generation seems to be among the first events
during HEBP1 hydrolysis, as short incubations lead to F2L gen-
eration (Fig. 6B, left panel) whereas longer incubations generate
also larger peptides (Fig. 6B, right panel).
Chemotaxis of monocytes toward HEBP1-derived peptides
F2L is a chemoattractant agent for FPR3-expressing cells, such as
monocytes, DCs, and macrophages (18, 21, 29). We evaluated
therefore the ability of CTS D-treated HEBP1 to recruit FPR3-
expressing leukocytes. In microchemotaxis Boyden’s chambers,
processed HEBP1 promoted recruitment of human monocyte-
derived macrophages and human monocytes (Fig. 7B, 7C). The
chemotaxis index was comparable with that obtained for synthetic
F2L (10 nM) used as a positive control. No chemotaxis was ob-
served toward CTS D alone or unprocessed HEBP1. Checkerboard
analysis on monocyte-derived macrophages (Fig. 7A) excluded
showing complete differentiation (CD68+CD206+) at day 7 after purification of monocytes. B and C, Human monocyte-derived macrophages were
transfected with a fluorescent control oligonucleotide, and transfection efficacy was evaluated by fluorescence microscopy (B) and flow cytometry (C). D,
Three siRNAs (A, B, and C) and the respective scrambled oligonucleotides (S, 10 nM) as controls were transfected at the indicated concentrations at days 6
and 7 of human monocyte-to-macrophage differentiation. Detection of CTS D and GAPDH (as control) was made on Western blots, showing a decrease in
mature CTS D and its precursor after siRNAs, but not scrambled oligonucleotides, transfection. Untransfected cells were used as positive controls of CTS D
immunoreactivity (2). E, Conditioned media (CM) from siRNA-treated monocyte-derived macrophages were tested for their ability to generate bioactivity
from purified human HEBP1 (1 mM, 30-min incubation). After incubation, the media were diluted four times and the activity measured in the aequorin-
based assay using FPR3-expressing CHO-K1 cells. A complete loss of HEBP1 processing was observed for conditioned media resulting from siRNA-
treated macrophages. Scrambled oligonucleotides had no effect on HEBP1 hydrolysis.
Knockdown of CTS D by siRNAs prevents HEBP1 hydrolysis. A, Flow cytometry analysis of human monocyte-derived macrophages
The Journal of Immunology1481
chemokinesis in favor of chemotaxis, as previously demonstrated
for the F2L peptide (21). In mouse, the F2L peptide can also
promote neutrophil chemotaxis, although with a lower efficacy
than for human leukocytes (18, 35). We tested therefore high
concentrations of CTS D-treated HEBP1 on mouse neutrophils
and demonstrated chemotaxis of these cells. This chemotactic
activity demonstrated further the production of functional F2L
peptide and suggests a physiological relevance of the HEBP1
processing by CTS D.
Distribution of HEBP1 and CTS D in human organs
Expression of the mouse ortholog of HEBP1, p22HBP, was pre-
viously described invarious tissues by Northern blot analysis. High
expression was found in liver, kidney, and spleen. To evaluate more
precisely the expression of human HEBP1, we characterized a new
Ab directed against the N terminus of HEBP1 and used this Ab on
tissue microarrays. For each tissue spot, we assessed an immu-
nohistological score from 0 (no detectable expression) to 2 (high
expression) and computed the mean of all scores for each organ
(Fig. 8). Results are presented in Supplemental Table II. High
levels of HEBP1 immunoreactivity were observed in liver, spleen,
and lymph nodes, but also in pancreatic islets and proximal kid-
ney tubules. Lower expression was seen in seminal vesicles and
Leydig cells of the testis. No expression of HEBP1 was observed
in breast, prostate, intestinal segments, ovary, uterus, skin, esoph-
agus, and thymus. Using the same approach, we evaluated the
distribution of CTS D. In agreement with published data, CTS D
was highly expressed in most tissues, particularly in specialized
cells such as macrophages. We observed coexpression of high
levels of CTS D and HEBP1 in liver, kidney, and spleen, con-
sistent with a potential role of CTS D in HEBP1 processing
Proteolytic regulation of the biological activity of peptides and
proteins acting as ligands of GPCRs is not an uncommon process.
Many peptidic GPCR ligands are processed by proteases, resulting
intheir partial orfullinactivation. Almost asfrequently, proteolysis
is required for the generation of GPCR ligands or results in
a significant increase of their affinity and potency for the receptor.
specific endogenous ligand for FPR3. F2L corresponds with the N
terminus of the heme-binding protein HEBP1, but the mechanisms
involved in its production from the precursor protein are unknown.
In this work, we investigated how F2L is generated from HEBP1.
Human full-length HEBP1 displays low agonist activity toward
FPR3 and does not act as an antagonist. The recent characterization
of HEBP1 tridimensional structure has shown that the F2L peptide
is located outside the structured domain of the protein. Its pro-
teolytic release might therefore keep intact the function of HEBP1
(24, 25, 36). We determined the structure–function relations of
F2L using point mutants and truncated forms of the peptide. It
appears that the important part of the peptide lies near its N ter-
minus, the most crucial positions being G3, N7, and S8. However,
other low-affinity interaction sites are likely spread over the whole
length of F2L, as it was not possible to design much shorter
peptides while keeping a reasonable affinity for FPR3. We ob-
served indeed a shift of one log in the EC50between peptide 1–18
(almost fully active) and peptide 1–15, suggesting a contribution
of the WPW motif. F2L is highly conserved among mammalian
species, and the human peptide differs from mouse F2L by only
one aa in position 6, where lysine is replaced by arginine. Steric
hindrance probably prevents the access of full-length HEBP1 to
the binding pocket of F2L, limiting greatly its ability to activate
FPR3. Additional work will be required to identify peptides with
enhanced biological activities or antagonist properties.
To determine which proteases or protease families are able to
incubated with a panel of conditioned media. Indeed, no consensus
sequence for candidate proteases could be identified in HEBP1
using bioinformatics programs. We selected as a starting point cells
known to contain high amounts of proteases, including macro-
phages. Macrophages are phagocytic cells, which differentiate from
circulating monocytes, and are located at strategic sites in the
organism. They are involved in innate immune defense and the
initiation of adaptive immunity. Upon activation, macrophages
release a large set of proteases, including plasminogen activator,
collagenases, elastase-like enzymes, and other hydrolases, such as
lysozyme and b-glucuronidase. They also secrete high amounts of
lysosomal enzymes independently of external stimuli. The activity
of these hydrolases is generally low at the neutral pH of extra-
cellular fluids, but macrophages are able to acidify the pericellular
space by the activity of their plasma membrane proton pumps
(V-ATPase) and secretion of lactic acid. This acidification is pos-
sibly a prerequisite for degradative processes, and the concentration
HEBP1 (6 mg) was processed by 200 ng CTS D for 5 or 30 min at 37˚C
and loaded on a 10–20% polyacrylamide gradient gel before silver staining
(A) or immunoblot detection with an anti-F2L Ab (B). A diffuse band was
detected at 2.5 kDa, the size of F2L. This band was more intense after
a short incubation time. C, The silver-stained bands were analyzed by mass
spectrometry after trypsinization. The mass of the tryptic fragments
identified are presented in the middle column and the corresponding region
of HEBP1 protein sequence in the left column. D, Schematic representa-
tion of the cleavage sites in HEBP1 by CTS D (triangles) and of the
resulting fragments detected.
HEBP1 processing by CTS D generates F2L. A and B,
1482 HEBP1 PROCESSING
of released enzymes is often sufficient for inducing lytic damage
to other cells (37). Therefore, in the microenvironment of mac-
rophages, the pH may be lower than in serum or extracellular
fluids, resulting in the activation of released lysosomal enzymes.
The testing of conditioned media allowed us to identify CTS D
as a candidate for HEBP1 processing and F2L generation, as the
activity on FPR3 was totally inhibited by pepstatin A, an aspartyl
protease inhibitor, as well as by transfection of siRNAs targeting
CTS D transcripts. Moreover, CTS D is described as an endo-
peptidase cleaving mainly after hydrophobic amino acids, partic-
ularlyleucine andphenylalanine.Thisisconsistentwiththe VL/SK
motif as the cleavage site for F2L generation. CTS D is a lysosomal
aspartic endopeptidase that plays an essential role in cell ho-
meostasis by degrading aging proteins in the lysosomal com-
partment and recycling their components. This enzyme is also
in the regulation of programmed cell death (reviewed in Refs. 38–
40). Deregulated CTS D activity (altered expression, increased
secretion, or unbalance between protease and endogenous inhib-
itors) is a factor contributing to various diseases characterized by
a chronic inflammatory state, such as cancer, asthma, atheroscle-
rosis, Alzheimer’s disease, periodontitis, rheumatoid arthritis, in-
flammatory bowel disease, osteoarthritis, and pulmonary fibrosis
(reviewed in Ref. 41).
In our hands, human purified CTS D cleaved, in less than 5 min,
HEBP1 at two main sites, leading to the generation of N-terminal
peptides. One of these peptides was identified by mass spec-
trometry as F2L, the endogenous ligand of FPR3. At this point, we
cannot exclude the possibility that other proteases might generate
F2L or other HEBP1-derived peptides active on FPR3. A peptide
corresponding with the first 50 aa of HEBP1 was indeed identified
in extracts from porcine spleen (21), although displaying lower
potency than F2L on FPR3. F2L and other related peptides might,
however, act together in leukocyte recruitment.
CTS D is produced by almost all cells but is expressed at
particularly high levels by macrophages in some tissues, such as
liver, spleen, and kidney. In contrast, our immunohistochemical
screening of tissues has shown that HEBP1 is not expressed
ubiquitously butrather specifically in liver, spleen, lymph nodes, as
HEBP1 processed by CTS D. A, A checkerboard
chemotaxis assay was conducted on human mono-
cyte-derived macrophages. The product of HEBP1
hydrolysis by CTS D was added to the upper and/or
lower chambers. Cells were added to the upper
chamber and allowed to migrate for 90 min at 37˚C.
The number of cells having migrated completely
through the filter is indicated. B–D, Chemotaxis of
human monocytes (B), monocyte-derived macro-
phages (C), and mouse neutrophils (D) in response
to 1 mM (human cells) or 10 mM (mouse cells)
HEBP1 processed by 60 nM CTS D for 30 min at
37˚C. Synthetic F2L at 10 nM (human cells) or 400
nM (mouse cells) was used as positive control of
migration. The displayed responses are representa-
tive of three experiments, and results are expressed
as the ratio of migrated cells in the presence ver-
sus the absence of chemoattractant. **p , 0.01,
***p , 0.001.
7. Leukocyte chemotaxistoward
croarrays. A, Overview of a tissue microarray section.
A rectangle of 15 cores corresponds with one organ;
each column corresponds with one donor, and the three
lines correspond with core replicates. B, Immunohis-
tochemical staining was scored 0 for no expression (left
panel), 1 for low expression or dispersed cells (middle
panel, arrows), and 2 for high expression of HEBP1
(right panel). Original magnification 3200. Arrows
indicate stained cells.
HEBP1 immunodetection on tissue mi-
The Journal of Immunology 1483
well as in kidney and pancreas, where CTS D is also highly
expressed. The correlated expression of HEBP1 and CTS D in
several tissues supports the physiological relevance of HEBP1
processing by CTS D.
HEBP1 appears as a good substrate for CTS D, but we cannot
FPR3 ligands, with the contribution of other proteases acting in
parallel or in cascade. A similar situation prevails for other bioactive
peptides such as hemorphins (42), a family of small oligopeptides
(4-mers to 10-mers) displaying high affinity for atypical m opioid
receptors. The mechanism of their generation is still not well un-
derstood, but the first step is likely the hydrolysis of hemoglobin
b-chain by CTS D. Then, other hemorphin variants are produced
by other proteases, including DPP-IV (43) and ACE (44). In this
system, macrophages cultured in an acidic environment are able to
proteolyze hemoglobin and generate hemorphins (45).
In this study, we showed that the F2L peptide, derived from the
processing of nanomolar concentrations of HEBP1, is active on
FPR3 and able to induce the recruitment of FPR3-expressing
monocytes, monocyte-derived macrophages, as well as mouse
neutrophils. It is, however, still unclear how cytosolic HEBP1 can
give rise to an extracellular chemoattractant factor. HEBP1 is
indeed clearly devoid of signal peptide, and all available data point
toward an intracellular localization. This raises the question of how
HEBP1 or F2L are secreted or otherwise released from cells.
First, HEBP1 might be processed inside the cell and generate
F2L, in a process such as apoptosis. Apoptosis is traditionally
characterized by the maintenance of organelle integrity and con-
densation and fragmentation of DNA, ending in the breaking up of
the cell into apoptotic bodies. It can occur through numerous
pathways, all converging to the activation of the caspase family of
proteases. Recent studies have described a significant role for CTS
D in the apoptosis of inflammatory cells such as neutrophils. The
clearance of activated neutrophils, essentially by macrophages, is
D was proposed as a key initiator of apoptosis in neutrophils,
through the activation of caspase-8 (39, 46). In this model, loss of
lysosomal integrity and release of CTS D in the cytosol were
described as primary events in the induction of apoptosis (47). The
generation of F2L might therefore take place in apoptotic cells,
but the peptide would still require to be secreted to recruit mac-
rophages, thereby contributing to the resolution of inflammation.
HEBP1 might also be released from the cell through a non-
conventional secretory pathway or as a result of necrosis. Sev-
eral distinct nonconventional secretory mechanisms have been
described (reviewed in Ref. 48). IL-1a is myristoylated, then
translocates to the cell membrane and is released by calpain-
induced proteolysis. IL-1b is generated from its precursor by
the IL-converting enzyme, a member of the caspase family. It then
enters the endosomal compartment via the ABCA1 transporter,
before being sorted and exported by secretory vesicles. In the case
of F2L, if full-length HEBP1 was released from cells, proteolysis
should occur in the extracellular compartment in the presence of
extracellular CTS D. A similar situation would prevail if HEBP1
is released when cells undergo necrosis. Necrotic death is asso-
ciated with the release of the cellular content in the extracellular
space. Dying cells and surrounding live cells release a set of in-
flammatory mediators, which recruit phagocytes (such as neu-
trophils and macrophages) that clear the cell debris. Under phy-
siological conditions, CTS D is sorted to the lysosomes and
found intracellularly. However, in some physiological and patho-
logical conditions, CTS D is secreted by various cell types. In-
deed, procathepsin D was found in human, bovine, and rat milk
(49), serum, sweat (50), and urine. It is also secreted at high levels
in tumors (reviewed in Refs. 51, 52) and atherosclerotic lesions
(53). In these pathologies, F2L could participate in the inflam-
matory response by recruiting leukocytes expressing FPR3, such
as macrophages, DCs, and eosinophils.
In conclusion, in this work we identified a protease able to
process HEBP1 and to generate biologically active F2L peptide.
F2L is the only endogenous ligand specific for FPR3 and che-
motactic for FPR3-expressing leukocyte populations. Further in-
vestigations will be required to determine in which physiologi-
cal and pathological conditions F2L is produced by CTS D. One
such condition is possibly apoptosis, in which F2L could con-
tribute to the resolution of inflammation. Alternatively, F2L might
be generated extracellularly by CTS D in pathological conditions
such as cancer or atherosclerosis. In this context, understanding
the mechanisms leading to HEBP1/F2L secretion or release from
the cells will be important.
We thank Dr. Guy Vandenbussche for mass spectrometry analyses of puri-
fied recombinant HEBP1 and Franc ¸oise Gre ´goire for the synthesis of F2L
The authors have no financial conflicts of interest.
1. Lundberg, A. M., and G. K. Hansson. 2010. Innate immune signals in athero-
sclerosis. Clin. Immunol. 134: 5–24.
2. Libby, P. 2008. Role of inflammation in atherosclerosis associated with rheu-
matoid arthritis. Am. J. Med. 121(10, Suppl 1): S21–S31.
3. Mantovani, A., P. Allavena, A. Sica, and F. Balkwill. 2008. Cancer-related in-
flammation. Nature 454: 436–444.
4. Serbina, N. V., T. Jia, T. M. Hohl, and E. G. Pamer. 2008. Monocyte-mediated
defense against microbial pathogens. Annu. Rev. Immunol. 26: 421–452.
5. Zlotnik, A. 2006. Chemokines and cancer. Int. J. Cancer 119: 2026–2029.
6. Rabiet, M. J., E. Huet, and F. Boulay. 2007. The N-formyl peptide receptors and
the anaphylatoxin C5a receptors: an overview. Biochimie 89: 1089–1106.
7. Marasco, W. A., S. H. Phan, H. Krutzsch, H. J. Showell, D. E. Feltner, R. Nairn,
E. L. Becker, and P. A. Ward. 1984. Purification and identification of formyl-
methionyl-leucyl-phenylalanine as the major peptide neutrophil chemotactic
factor produced by Escherichia coli. J. Biol. Chem. 259: 5430–5439.
8. Schiffmann, E., H. V. Showell, B. A. Corcoran, P. A. Ward, E. Smith, and
E. L. Becker. 1975. The isolation and partial characterization of neutrophil
chemotactic factors from Escherichia coli. J. Immunol. 114: 1831–1837.
9. Schiffmann, E., B. A. Corcoran, and S. M. Wahl. 1975. N-formylmethionyl
peptides as chemoattractants for leucocytes. Proc. Natl. Acad. Sci. USA 72:
10. Showell, H. J., R. J. Freer, S. H. Zigmond, E. Schiffmann, S. Aswanikumar,
B. Corcoran, and E. L. Becker. 1976. The structure-activity relations of synthetic
peptides as chemotactic factors and inducers of lysosomal secretion for neu-
trophils. J. Exp. Med. 143: 1154–1169.
11. Boulay, F., M. Tardif, L. Brouchon, and P. Vignais. 1990. The human N-
formylpeptide receptor. Characterization of two cDNA isolates and evidence for
a new subfamily of G-protein-coupled receptors. Biochemistry 29: 11123–11133.
12. Boulay, F., M. Tardif, L. Brouchon, and P. Vignais. 1990. Synthesis and use of
a novel N-formyl peptide derivative to isolate a human N-formyl peptide re-
ceptor cDNA. Biochem. Biophys. Res. Commun. 168: 1103–1109.
13. Bao, L., N. P. Gerard, R. L. Eddy, Jr., T. B. Shows, and C. Gerard. 1992.
Mapping of genes for the human C5a receptor (C5AR), human FMLP receptor
(FPR), and two FMLP receptor homologue orphan receptors (FPRH1, FPRH2)
to chromosome 19. Genomics 13: 437–440.
14. Murphy, P. M., T. Ozc ¸elik, R. T. Kenney, H. L. Tiffany, D. McDermott, and
U. Francke. 1992. A structural homologue of the N-formyl peptide receptor.
Characterization and chromosome mapping of a peptide chemoattractant re-
ceptor family. J. Biol. Chem. 267: 7637–7643.
15. Ye, R. D., S. L. Cavanagh, O. Quehenberger, E. R. Prossnitz, and C. G. Cochrane.
1992. Isolation of a cDNA that encodes a novel granulocyte N-formyl peptide
receptor. Biochem. Biophys. Res. Commun. 184: 582–589.
16. Migeotte, I., D. Communi, and M. Parmentier. 2006. Formyl peptide receptors:
a promiscuous subfamily of G protein-coupled receptors controlling immune
responses. Cytokine Growth Factor Rev. 17: 501–519.
17. Durstin, M., J. L. Gao, H. L. Tiffany, D. McDermott, and P. M. Murphy. 1994.
Differential expression of members of the N-formylpeptide receptor gene cluster
in human phagocytes. Biochem. Biophys. Res. Commun. 201: 174–179.
18. Devosse, T., A. Guillabert, N. D’Haene, A. Berton, P. De Nadai, S. Noel,
M. Brait, J. D. Franssen, S. Sozzani, I. Salmon, and M. Parmentier. 2009. Formyl
peptide receptor-like 2 is expressed and functional in plasmacytoid dendritic
1484 HEBP1 PROCESSING
cells, tissue-specific macrophage subpopulations, and eosinophils. J. Immunol.
19. Harada, M., Y. Habata, M. Hosoya, K. Nishi, R. Fujii, M. Kobayashi, and
S. Hinuma. 2004. N-Formylated humanin activates both formyl peptide receptor-
like 1 and 2. Biochem. Biophys. Res. Commun. 324: 255–261.
20. Ying, G., P. Iribarren, Y. Zhou, W. Gong, N. Zhang, Z. X. Yu, Y. Le, Y. Cui, and
J. M. Wang. 2004. Humanin, a newly identified neuroprotective factor, uses the
G protein-coupled formylpeptide receptor-like-1 as a functional receptor. J.
Immunol. 172: 7078–7085.
21. Migeotte, I., E. Riboldi, J. D. Franssen, F. Gre ´goire, C. Loison, V. Wittamer,
M. Detheux, P. Robberecht, S. Costagliola, G. Vassart, et al. 2005. Identification
and characterization of an endogenous chemotactic ligand specific for FPRL2. J.
Exp. Med. 201: 83–93.
22. Taketani, S., Y. Adachi, H. Kohno, S. Ikehara, R. Tokunaga, and T. Ishii. 1998.
differentiation of urine erythroleukemia cells. J. Biol. Chem. 273: 31388–31394.
23. Zylka, M. J., and S. M. Reppert. 1999. Discovery of a putative heme-binding
protein family (SOUL/HBP) by two-tissue suppression subtractive hybridization
and database searches. Brain Res. Mol. Brain Res. 74: 175–181.
24. Dias, J. S., A. L. Macedo, G. C. Ferreira, F. C. Peterson, B. F. Volkman, and
B. J. Goodfellow. 2006. The first structure from the SOUL/HBP family of heme-
binding proteins, murine P22HBP. J. Biol. Chem. 281: 31553–31561.
25. Gell, D. A., B. J. Westman, D. Gorman, C. K. Liew, J. J. Welch, M. J. Weiss, and
J. P. Mackay. 2006. A novel haem-binding interface in the 22 kDa haem-binding
protein p22HBP. J. Mol. Biol. 362: 287–297.
26. Dutoit, R., E. Dubois, and E. Jacobs. 2010. Selection systems based on
dominant-negative transcription factors for precise genetic engineering. Nucleic
Acids Res. 38: e183.
27. Ito, H., Y. Fukuda, K. Murata, and A. Kimura. 1983. Transformation of intact
yeast cells treated with alkali cations. J. Bacteriol. 153: 163–168.
28. Button, D., and M. Brownstein. 1993. Aequorin-expressing mammalian cell
lines used to report Ca2+ mobilization. Cell Calcium 14: 663–671.
29. Le Poul, E., S. Hisada, Y. Mizuguchi, V. J. Dupriez, E. Burgeon, and
M. Detheux. 2002. Adaptation of aequorin functional assay to high throughput
screening. J. Biomol. Screen. 7: 57–65.
30. Mortz, E., T. N. Krogh, H. Vorum, and A. Go ¨rg. 2001. Improved silver staining
protocols for high sensitivity protein identification using matrix-assisted laser
desorption/ionization-time of flight analysis. Proteomics 1: 1359–1363.
31. Boyden, S. 1962. The chemotactic effect of mixtures of antibody and antigen on
polymorphonuclear leucocytes. J. Exp. Med. 115: 453–466.
32. Polevoda, B., and F. Sherman. 2000. Nalpha -terminal acetylation of eukaryotic
proteins. J. Biol. Chem. 275: 36479–36482.
33. Polevoda, B., and F. Sherman. 2003. N-terminal acetyltransferases and sequence
requirements for N-terminal acetylation of eukaryotic proteins. J. Mol. Biol. 325:
34. Polevoda, B., and F. Sherman. 2002. The diversity of acetylated proteins. Ge-
nome Biol. 3: reviews0006.
35. Gao, J. L., A. Guillabert, J. Hu, Y. Le, E. Urizar, E. Seligman, K. J. Fang,
X. Yuan, V. Imbault, D. Communi, et al. 2007. F2L, a peptide derived from
heme-binding protein, chemoattracts mouse neutrophils by specifically activating
Fpr2, the low-affinity N-formylpeptide receptor. J. Immunol. 178: 1450–1456.
36. Dias, J. S., A. L. Macedo, G. C. Ferreira, N. Jeanty, S. Taketani,
B. J. Goodfellow, F. C. Peterson, and B. F. Volkman. 2005. 1H, 15N and 13C
resonance assignments of the heme-binding protein murine p22HBP. J. Biomol.
NMR 32: 338.
37. Tapper, H., and R. Sundler. 1990. Role of lysosomal and cytosolic pH in the
regulation of macrophage lysosomal enzyme secretion. Biochem. J. 272: 407–
38. Benes, P., V. Vetvicka, and M. Fusek. 2008. Cathepsin D—many functions of
one aspartic protease. Crit. Rev. Oncol. Hematol. 68: 12–28.
39. Conus, S., R. Perozzo, T. Reinheckel, C. Peters, L. Scapozza, S. Yousefi, and
H. U. Simon. 2008. Caspase-8 is activated by cathepsin D initiating neutrophil
apoptosis during the resolution of inflammation. J. Exp. Med. 205: 685–698.
40. Minarowska, A., M. Gacko, A. Karwowska, and L. Minarowski. 2008. Human
cathepsin D. Folia Histochem. Cytobiol. 46: 23–38.
41. Zavasnik-Bergant, T., and B. Turk. 2007. Cysteine proteases: destruction ability
versus immunomodulation capacity in immune cells. Biol. Chem. 388: 1141–
42. Fruitier, I., I. Garreau, and J. M. Piot. 1998. Cathepsin D is a good candidate for
the specific release of a stable hemorphin from hemoglobin in vivo: VV-hem-
orphin-7. Biochem. Biophys. Res. Commun. 246: 719–724.
43. Cohen, M., I. Fruitier-Arnaudin, and J. M. Piot. 2004. Hemorphins: substrates
and/or inhibitors of dipeptidyl peptidase IV. Hemorphins N-terminus sequence
influence on the interaction between hemorphins and DPPIV. Biochimie 86: 31–
44. Lantz, I., E. L. Gla ¨msta, L. Talba ¨ck, and F. Nyberg. 1991. Hemorphins derived
from hemoglobin have an inhibitory action on angiotensin converting enzyme
activity. FEBS Lett. 287: 39–41.
45. Dagouassat, N., I. Garreau, F. Sannier, Q. Zhao, and J. M. Piot. 1996. Generation
of VV-hemorphin-7 from globin by peritoneal macrophages. FEBS Lett. 382:
46. Deiss, L. P., H. Galinka, H. Berissi, O. Cohen, and A. Kimchi. 1996. Cathepsin D
protease mediates programmed cell death induced by interferon-gamma, Fas/
APO-1 and TNF-alpha. EMBO J. 15: 3861–3870.
47. Conus, S., and H. U. Simon. 2008. Cathepsins: key modulators of cell death and
inflammatory responses. Biochem. Pharmacol. 76: 1374–1382.
48. Nickel, W., and C. Rabouille. 2009. Mechanisms of regulated unconventional
protein secretion. Nat. Rev. Mol. Cell Biol. 10: 148–155.
49. Lkhider, M., R. Castino, E. Bouguyon, C. Isidoro, and M. Ollivier-Bousquet.
2004. Cathepsin D released by lactating rat mammary epithelial cells is involved
in prolactin cleavage under physiological conditions. J. Cell Sci. 117: 5155–
50. Baechle, D., T. Flad, A. Cansier, H. Steffen, B. Schittek, J. Tolson, T. Herrmann,
H. Dihazi, A. Beck, G. A. Mueller, et al. 2006. Cathepsin D is present in human
eccrine sweat and involved in the postsecretory processing of the antimicrobial
peptide DCD-1L. J. Biol. Chem. 281: 5406–5415.
51. Fusek, M., and V. Vetvicka. 2005. Dual role of cathepsin D: ligand and protease.
Biomed. Pap. Med. Fac. Univ. Palacky Olomouc Czech Repub. 149: 43–50.
52. Liaudet-Coopman, E., M. Beaujouin, D. Derocq, M. Garcia, M. Glondu-Lassis,
V. Laurent-Matha, C. Pre ´bois, H. Rochefort, and F. Vignon. 2006. Cathepsin D:
newly discovered functions of a long-standing aspartic protease in cancer and
apoptosis. Cancer Lett. 237: 167–179.
53. Hakala, J. K., R. Oksjoki, P. Laine, H. Du, G. A. Grabowski, P. T. Kovanen, and
M. O. Pentika ¨inen. 2003. Lysosomal enzymes are released from cultured human
macrophages, hydrolyze LDL in vitro, and are present extracellularly in human
atherosclerotic lesions. Arterioscler. Thromb. Vasc. Biol. 23: 1430–1436.
The Journal of Immunology1485