Báo cáo khoa học: A role for serglycin proteoglycan in granular retention and processing of mast cell secretory granule components ppt

A role for serglycin proteoglycan in granular retention
and processing of mast cell secretory granule components
Frida Henningsson*, Sonja Hergeth*, Robert Cortelius, Magnus A
˚
brink and Gunnar Pejler
Swedish University of Agricultural Sciences, Department of Molecular Biosciences, The Biomedical Center, Uppsala, Sweden
Mast cells (MCs) are characterized by their large con-
tent of electron-dense secretory granules, and these
granules are released following MC activation, a pro-
cess that can be accomplished by various mechanisms,
including antigen-mediated crosslinking of surface-
associated IgE and exposure to neuropeptides, anaphy-
latoxins or calcium ionophores [1,2]. The MC granules
contain a broad array of bioactive compounds, with
the exact composition being dependent on the partic-
ular species and subclass of MC [1,3]. Histamine is a
major constituent of all types of MC, and it is now
well recognized that MC granules can contain a num-
ber of different cytokines, such as tumor necrosis
factor-a, interleukin (IL)-4, IL-5, IL-13, transforming
growth factor-b and vascular endothelial growth factor
[2]. Moreover, the MC granules contain b-hexosamini-
dase and a number of MC-specific neutral proteases:
chymases, tryptases and carboxypeptidase A (CPA)
[4,5]. In addition, MC granules contain large amounts
of highly sulfated and thereby negatively charged pro-
teoglycans (PGs) of the serglycin (SG) type, and it is
these PGs that give the typical metachromatic staining
of MCs with cationic dyes [6]. In MCs, SG PGs can
accommodate either (or both) chondroitin sulfate or
heparin side chains, depending on MC subclass [7].

The processes involved in MC degranulation, in par-
ticular the signal transduction pathways, have been the
subject of intense investigation [8,9]. In contrast, strik-
ingly little is known regarding the actual formation of
MC secretory granules. For example, the factors that
Keywords
mast cells; proteases; proteoglycans;
serglycin; sorting
Correspondence
G. Pejler, Swedish University of Agricultural
Sciences, Department of Molecular
Biosciences, The Biomedical Center,
Box 575, 751 23 Uppsala, Sweden
Fax: +46 18 550762
Tel: +46 18 4714090
E-mail:
*These authors contributed equally to this
work
(Received 28 June 2006, revised 15 August
2006, accepted 4 September 2006)
doi:10.1111/j.1742-4658.2006.05489.x
In the absence of serglycin proteoglycans, connective tissue-type mast cells
fail to assemble mature metachromatic secretory granules, and this is
accompanied by a markedly reduced ability to store neutral proteases.
However, the mechanisms behind these phenomena are not known. In this
study, we addressed these issues by studying the functionality and morphol-
ogy of secretory granules as well as the fate of the secretory granule prote-
ases in bone marrow-derived mast cells from serglycin
+ ⁄ +
and serglycin

– ⁄ –
mice. We show that functional secretory vesicles are formed in both the
presence and absence of serglycin, but that dense core formation is defect-
ive in serglycin
– ⁄ –
mast cell granules. The low levels of mast cell proteases
present in serglycin
– ⁄ –
cells had a granular location, as judged by immu-
nohistochemistry, and were released following exposure to calcium iono-
phore, indicating that they were correctly targeted into secretory granules
even in the absence of serglycin. In the absence of serglycin, the fates of
the serglycin-dependent proteases differed, including preferential degrada-
tion, exocytosis or defective intracellular processing. In contrast, b-hexosa-
minidase storage and release was not dependent on serglycin. Together,
these findings indicate that the reduced amounts of neutral proteases in the
absence of serglycin is not caused by missorting into the constitutive path-
way of secretion, but rather that serglycin may be involved in the retention
of the proteases after their entry into secretory vesicles.
Abbreviations
BMMC, bone marrow-derived mast cell; CPA, carboxypeptidase A; MC, mast cell; mMCP, mouse mast cell protease; PG, proteoglycan;
SG, serglycin; TEM, transmission electron microscopy.
FEBS Journal 273 (2006) 4901–4912 ª 2006 The Authors Journal compilation ª 2006 FEBS 4901
determine the sorting of granular components are lar-
gely undefined, and the mechanisms that lead to the
assembly of the electron-dense, metachromatically
staining granules seen in mature MCs have been
poorly investigated. In a recent study, we generated a
mouse strain in which the SG gene was targeted [10].
We found that, in the absence of SG, mature metach-

romatically staining granules were not observed. Fur-
thermore, we noted that storage, but not mRNA
expression, of the various MC proteases was dramatic-
ally defective in SG
– ⁄ –
MCs. However, the underlying
mechanisms behind these observations were not
defined. The aim of this study was therefore to provide
insights into this issue by determining the fate of SG-
dependent proteases in cells with an inactivated SG
gene. Our results are compatible with a model of secre-
tory granule maturation in which SG PG is not
involved in the transport of compounds into secretory
vesicles, but is essential for retention of certain constit-
uents after their entry into the granules.
Results
This study was undertaken to resolve the mechanism
behind the previously observed severe granule defects in
SG
– ⁄ –
MCs [10]. Given the dramatic effects of the SG
knockout on granular staining properties and storage of
proteases, it was first important to determine whether
the lack of SG affected the actual assembly of granules
and whether secretory granules were also functional in
the absence of SG. To this end, bone marrow cells
were recovered from SG
+ ⁄ +
and SG
– ⁄ –

mice and were
in vitro differentiated into mature bone marrow-derived
MCs (BMMCs) by culturing in medium containing IL-3
[11,12]. Cells were recovered at various stages of cellular
differentiation, and their morphology was examined
after staining with May–Gru
¨
nwald ⁄ Giemsa. At day 0,
as expected, no cells with an MC-like appearance were
observed. Starting from day 5, cells containing ‘empty’
(May–Gru
¨
nwald ⁄ Giemsa-negative) vesicles were
observed. Such vesicles were seen both in SG
+ ⁄ +
and
SG
– ⁄ –
cells, indicating that their formation was not
dependent on SG. When the cells were cultured further,
the number of May–Gru
¨
nwald ⁄ Giemsa-negative vesi-
cles gradually decreased in both SG
+ ⁄ +
and SG
– ⁄ –
cells.
This was accompanied by the appearance, from about
day 12, of May–Gru

¨
nwald ⁄ Giemsa-positive granular
structures in SG
+ ⁄ +
cells. A gradual increase in
May–Gru
¨
nwald ⁄ Giemsa staining was seen over time in
SG
+ ⁄ +
cells. In contrast, May–Gru
¨
nwald ⁄ Giemsa-pos-
itive vesicles were not seen in SG
– ⁄ –
cells at any stage of
differentiation (not shown). These results are in agree-
ment with those of a previous study [10].
One potential explanation for the lack of May–
Gru
¨
nwald ⁄ Giemsa-negative vesicles in SG
– ⁄ –
cells at
later stages of differentiation could be that immature
granules are generated in the absence of SG, but that
the lack of SG causes their disruption. Alternatively,
secretory granules could be present at late stages of
maturation also in SG
– ⁄ –

cells, but not be visible by
conventional microscopy. To provide further insights
into this issue, we examined the cells at the ultrastruc-
tural level by transmission electron microscopy (TEM).
The TEM analysis indeed revealed the existence of
secretory granule-like organelles in SG
– ⁄ –
cells, and
these organelles were found in approximately equal
numbers as in SG
+ ⁄ +
cells (Fig. 1; upper panels).
However, the morphology of the granules was differ-
ent; whereas dense core formation was seen in SG
+ ⁄ +
granules, the contents of the SG
– ⁄ –
granule were of
more amorphous character, without defined electron-
dense cores (Fig. 1; lower panels).
To address whether the secretory granules were
functional, we measured the ability of the MCs to
release b-hexosaminidase, a granule component, upon
exposure to calcium ionophore A23187. As shown in
Fig. 2A, equal amounts of b-hexosaminidase were
released by SG
+ ⁄ +
and SG
– ⁄ –
cells after calcium iono-

phore stimulation, and the kinetics of release were sim-
ilar. Furthermore, the levels of b-hexosaminidase in
conditioned medium from nonstimulated cells were
similar in cultures of both genotypes (Fig. 2A), indica-
ting that the lack of SG PG did not result in increased
spontaneous release of b-hexosaminidase. These find-
ings indicate that the general ability of MCs to degran-
ulate is not dependent on SG. Experiments were also
undertaken to investigate whether the level of stored
b-hexosaminidase is influenced by SG. Although
b-hexosaminidase activity was already detected at day
0, the intracellular content of this enzyme increased
markedly after 6 days of culture, and reached a plat-
eau from about day 12 (Fig. 2B). Both the kinetics of
accumulation and the level of maximal storage were
virtually identical in SG
+ ⁄ +
and SG
– ⁄ –
cells, indica-
ting that the storage of b-hexosaminidase is independ-
ent of SG.
The results above indicate that functional MC secre-
tory granules are formed independently of SG PG.
Hence, the defective storage of MC proteases in the
absence of SG [10] is not due to defects in the forma-
tion or functionality of granular compartments. In
order to understand the mechanism by which SG PG
promotes storage of these compounds, the strategy in
the next set of experiments was to follow the fates

of the SG-dependent proteases when SG was absent.
To address these issues, we examined the expression,
Role of serglycin in secretory granule assembly F. Henningsson et al.
4902 FEBS Journal 273 (2006) 4901–4912 ª 2006 The Authors Journal compilation ª 2006 FEBS
cellular storage, processing and secretion of SG and of
various MC proteases at different stages of MC differ-
entiation, as described below.
SG core protein transcript was already clearly
detectable at day 0, but the level of transcript appeared
to increase after 5 days of culture (Fig. 3A). In con-
trast, no detectable mRNA for mouse MC protease 5
(mMCP-5; a chymase), the tryptase mMCP-6 or CPA
was detected at day 0. Starting from day 5, however,
mMCP-6 and CPA transcripts were clearly detected,
and they appeared to increase further after 12 days of
culture. The onset of mMCP-5 mRNA expression was
somewhat delayed, with clearly detectable transcript
being seen from day 12. mMCP-5, mMCP-6 and CPA
transcripts were detected in both SG
+ ⁄ +
and SG
– ⁄ –
cells, and the kinetics as regards onset of mRNA
expression were similar in both genotypes, indicating
that the knockout of SG does not affect cellular differ-
entiation into MCs, as judged by the transcription of
the MC protease genes.
Further experiments were carried out to examine
how the mRNA expression profiles of SG
+ ⁄ +

and
SG
– ⁄ –
cells were reflected at the protein level (Fig. 3B).
Immunoblot analysis of SG
+ ⁄ +
cell extracts showed
that neither of the MC proteases were present at day
0. mMCP-5 protein was detected starting from day 12,
i.e. at the same time as when gene transcription was
first seen. mMCP-5 protein accumulated over time,
with a plateau of maximal storage seen after 26 days
of culture. mMCP-6 storage showed similar kinetics as
for mMCP-5. In contrast, CPA protein was detected
as early as after 5 days of culture, and a maximal plat-
eau of storage was already seen at day 12. Both pro-
CPA and mature CPA were detected in SG
+ ⁄ +
cells.
A dramatically different pattern was seen in SG
– ⁄ –
cells. mMCP-5 protein was not detected at any time
point, and mMCP-6 protein, although being detect-
able, was present at markedly lower levels than in
SG
+ ⁄ +
cells. Notably, however, mMCP-6 accumula-
tion in SG
– ⁄ –
cells showed similar kinetics as in

SG
+ ⁄ +
cells. In contrast, the total amounts of CPA
antigen (pro-CPA + mature CPA) were approximately
equal in SG
– ⁄ –
and SG
+ ⁄ +
cells. An interesting
observation was that only the pro-form of CPA was
Fig. 1. Transmission electron micrographs. The upper panels show representative mature (5 weeks of culture) bone marrow-derived mast
cells (BMMCs) from serglycin (SG)
+ ⁄ +
and SG
– ⁄ –
mice (original magnification 5000·). The lower panels show enlarged granules (original
magnification · 40 000).
F. Henningsson et al. Role of serglycin in secretory granule assembly
FEBS Journal 273 (2006) 4901–4912 ª 2006 The Authors Journal compilation ª 2006 FEBS 4903
detected in SG
– ⁄ –
cells, indicating that pro-CPA
processing into mature protease is strongly dependent
on SG. A plausible explanation for this finding is that
the protease(s) that are responsible for the pro-CPA
processing is dependent on SG. In agreement with
such a notion, we showed recently that pro-CPA
processing was defective in cathepsin E
– ⁄ –
MCs and

that the storage of cathepsin E in MCs is dependent
on heparin PG [13]. Thus, a likely explanation for the
defective pro-CPA processing in SG
– ⁄ –
MCs is that the
lack of SG also results in defective cathepsin E storage
and that this, in turn, results in defective pro-CPA
processing, leading to an accumulation of pro-CPA.
Next, we investigated the possibility that the pro-
teases were constitutively secreted in the absence of
SG PG. Conditioned media were collected at different
stages of MC differentiation, and were analyzed for
the presence of secreted MC proteases. As shown in
Fig. 3C, mMCP-5 protein was present at low levels
in conditioned medium from SG
+ ⁄ +
cells after
prolonged culture, but was absent in medium from
SG
– ⁄ –
cells. In contrast, mMCP-6 protein was clearly
detected, starting from about day 14, in conditioned
medium from both SG
+ ⁄ +
and SG
– ⁄ –
cells. CPA, pre-
dominantly in its mature form, was secreted into the
medium by SG
+ ⁄ +

cells, starting at about day 5. In
contrast, only the pro-form of CPA was secreted by
SG
– ⁄ –
cells. The total level of secreted CPA (pro-CPA
and mature CPA) was somewhat higher in medium
from SG
– ⁄ –
than in that from SG
+ ⁄ +
MCs, in partic-
ular at early time points (Fig. 3C). Note that, at early
time points, pro-CPA dominated over mature protease,
both intracellularly (Fig. 3B; day 5) and in conditioned
medium from SG
+ ⁄ +
cells (Fig. 3C; days 6 and 12),
indicating that efficient processing of pro-CPA is
dependent on the degree of MC maturation. In accord-
ance with this notion, only the mature form of CPA is
detected in fully mature connective tissue-type MCs
recovered in vivo [13], and only the pro-form of CPA
is detected in poorly differentiated transformed cell
lines of MC origin (M. Grujic & G. Pejler, unpub-
lished results).
The results above indicate that mMCP-6 and pro-
CPA are secreted by SG
– ⁄ –
MCs. A possible explan-
ation for these findings would be that the absence of

SG causes missorting of mMCP-6 and pro-CPA into
the constitutive rather than into the regulated pathway
of secretion. If indeed this were the case, mMCP-6 and
pro-CPA would not be present in the secretory gran-
ules, and exposure of SG
– ⁄ –
cells to MC-degranulating
agents would not cause increased release of mMCP-6
and pro-CPA. If, on the other hand, mMCP-6 and
pro-CPA are in fact located in secretory granules also
in SG
– ⁄ –
cells, MC degranulation would be expected to
induce their release. In order to address these possibil-
ities, SG
+ ⁄ +
and SG
– ⁄ –
MCs were exposed to the cal-
cium ionophore A23187, a compound that is regularly
used as an MC secretagogue [14]. As shown in
Fig. 4A, exposure of SG
+ ⁄ +
cells to A23187 resulted
in clearly detectable mMCP-6 and CPA in conditioned
medium. Strikingly, calcium ionophore stimulation
also resulted in the release of mMCP-6 and pro-CPA
by SG
– ⁄ –
cells (Fig. 4A). The implication of these find-

ings is that mMCP-6 and pro-CPA are sorted into
releasable secretory vesicles despite the absence of SG.
To obtain further evidence for this, SG
– ⁄ –
cells were
stained for mMCP-6 antigen, before and after expo-
sure to calcium ionophore. In resting SG
– ⁄ –
cells,
0
12345
0
10
20
30
40
50
60
70
80
90
100
Hours
+/+ non-stim
+/+ A23187
-/- non-stim
-/- A23187
Hexosaminidase release
(% of total)
A

0
010203040
20
40
60
80
100
120
Days of culture
+/+
-/-
Hexosaminidase content
(% of maximal)
B
Fig. 2. b-Hexosaminidase content and release. (A) Conditioned
media from mature (5 weeks of culture) serglycin (SG)
+ ⁄ +
(filled
symbols) and SG
– ⁄ –
(open symbols) cells were analyzed for b-hex-
osaminidase activity, without stimulation (squares) or after addition
of A23187 (circles). (B) SG
+ ⁄ +
(filled squares) and SG
– ⁄ –
(open
squares) cells taken at different stages (days 0–33) of differenti-
ation in interleukin (IL)-3-containing medium were analyzed for total
intracellular b-hexosaminidase activity. Results are expressed as

percentages, where the b-hexosaminidase content in SG
+ ⁄ +
cells
at day 33 is set as 100%. Results are expressed as means of tripli-
cate determinations ± SD.
Role of serglycin in secretory granule assembly F. Henningsson et al.
4904 FEBS Journal 273 (2006) 4901–4912 ª 2006 The Authors Journal compilation ª 2006 FEBS
mMCP-6 was found in granule-like compartments
close to the plasma membrane (Fig. 4C), indeed sup-
porting the idea that mMCP-6 is transported into
secretory granules even in the absence of SG PG. Fur-
thermore, after exposure to A23187, SG
– ⁄ –
cells
showed signs of degranulation and it was also evident
that the released granules stained positively for
mMCP-6 (Fig. 4C). Preimmune serum gave only dif-
fuse overall staining of SG
– ⁄ –
cells and a total absence
of granular staining of either unstimulated or A23187-
stimulated cells (Fig. 4C).
Next, we investigated the possibility that the MC
proteases are subjected to degradation by lysosomal
proteases when SG PG is absent. For this purpose,
cells were incubated with NH
4
Cl in order to raise the
pH of acidic intracellular compartments, including
lysosomes and secretory granules, and thereby inacti-

vate lysosomal proteases. Incubation of SG
– ⁄ –
MCs
with NH
4
Cl did not affect the level of intracellular
mMCP-6, indicating that degradation by lysosomal
mechanisms is not a primary fate of mMCP-6 when
SG is absent (Fig. 5). In contrast, NH
4
Cl caused an
SG
mMCP-6
mMCP-5
CPA
HPRT
0512
19 26 33
0512
19 26 33
SG
-/-
SG
+/+
A
Days:
0512
19 26 33
05 12
19 26 33Days:

0614
20 26 33
0614
20 26 33Days:
mMCP-6
mMCP-5
pro-CPA
CPA
B
mMCP-6
mMCP-5
pro-CPA
CPA
C
Fig. 3. mRNA expression and protein analysis. (A) Total RNA was prepared from serglycin (SG)
+ ⁄ +
and SG
– ⁄ –
bone marrow-derived cells
after different durations (days 0–33) of culture in medium containing interleukin (IL)-3. The RNA was used for analysis of the expression of
mouse mast cell protease (mMCP)-5, mMCP-6, carboxypeptidase A (CPA) and SG by RT-PCR. Expression of hypoxanthine–guanine phopho-
ribosyltransferase was used as housekeeping control. (B) Cell extracts were prepared from cells taken at various stages (days 0–33) of differ-
entiation and were subjected to immunoblot analysis using antisera towards mMCP-5, mMCP-6 and CPA. (C) Secretion of MC proteases
from SG
+ ⁄ +
and SG
– ⁄ –
cells. Conditioned media were recovered from SG
+ ⁄ +
and SG

– ⁄ –
cells at various stages (days 0–33) of differentiation
in IL-3-containing medium. The media were concentrated and subjected to immunoblot analysis for mMCP-5, mMCP-6 and CPA. The results
shown are representative of three independent experiments.
F. Henningsson et al. Role of serglycin in secretory granule assembly
FEBS Journal 273 (2006) 4901–4912 ª 2006 The Authors Journal compilation ª 2006 FEBS 4905
accumulation of mMCP-5 protein in SG
– ⁄ –
cells,
whereas mMCP-5 levels were not affected in SG
+ ⁄ +
MCs (Fig. 5; upper panel). This indicates that mMCP-
5, in the absence of SG, is degraded by proteases with
low pH optima, possibly in lysosomal compartments.
Interestingly, several ‘lysosomal’ proteases, e.g. cathep-
sin C, cathepsin D and cathepsin E, have been found
not only in lysosomes but also in MC secretory gran-
ules [13,15,16]. Thus, the degradation of mMCP-5 does
not necessarily have to involve transport to lysosomes,
but could in fact occur within the secretory granule
compartment. NH
4
Cl did not cause any noticeable
accumulation of pro-CPA or mature CPA in either
SG
+ ⁄ +
or SG
– ⁄ –
MCs. However, the NH
4

Cl treat-
ment resulted in the accumulation of an intermediate
form of CPA, of somewhat lower molecular weight
than pro-CPA (Fig. 5; lower panel). Most likely, this
compound represents an intermediate in processing.
These findings indicate that the processing of pro-CPA
occurs in (at least) two steps, and that the processing
of the intermediate form of CPA to mature protease is
dependent on a (lysosomal?) protease with an acidic
pH optimum. Control experiments showed that NH
4
Cl
did not affect cellular viability (not shown).
Degradation by the proteasome pathway could con-
stitute an alternative degradative pathway in the
absence of SG. However, incubation of cells with lac-
tacystin, an inhibitor of proteasome function, did not
cause any accumulation of MC proteases in SG
– ⁄ –
MCs (not shown).
AB
C
Fig. 4. Protease release after mast cell (MC)
degranulation. Serglycin (SG)
+ ⁄ +
and SG
– ⁄ –
MCs (after 5 weeks of culture) were treated
with calcium ionophore A23187. (A) Medium
fractions from SG

+ ⁄ +
and SG
– ⁄ –
cells were
subjected to immunoblot analysis using anti-
sera towards carboxypeptidase A (CPA) and
mouse mast cell protease (mMCP)-6. Note
the increase in extracellular mMCP-6 and
CPA antigen, in both SG
+ ⁄ +
and SG
– ⁄ –
cul-
tures, after stimulation with A23187. (B) Cell
fractions from SG
+ ⁄ +
and SG
– ⁄ –
cells were
analyzed for mMCP-6 and CPA by immuno-
blotting. (C) Cytospin slides were prepared
from nontreated and A23187-treated SG
– ⁄ –
cells and were immunohistochemically
stained for the presence of mMCP-6 anti-
gen. Note the granular staining for mMCP-6,
both before and after A23187 stimulation.
The results shown are representative of
three independent experiments.
Role of serglycin in secretory granule assembly F. Henningsson et al.

4906 FEBS Journal 273 (2006) 4901–4912 ª 2006 The Authors Journal compilation ª 2006 FEBS
As shown in Fig. 3A, SG core protein mRNA was
already expressed at day 0. However, maximal MC
protease accumulation was not obtained until about
day 26 (Fig. 3B), a finding that may appear contradict-
ory, considering the strong dependence of the MC pro-
teases on SG for storage. This indicates that the levels
of stored proteases are not directly related to the
amount of SG core protein mRNA being expressed.
One potential explanation could be that the amount of
actual sulfated PGs is not directly correlated with the
level of SG mRNA, and it was therefore of interest to
also follow the levels of sulfated PGs during the course
of MC differentiation. To this end, SG
+ ⁄ +
MCs at
different stages of differentiation were biosynthetically
labeled with
35
SO
4
2–
.
35
S-labeled PGs were recovered
both from conditioned medium (secreted PGs) and
from cell extracts, and were quantified. As shown in
Fig. 6A, the levels of secreted PGs were similar at all
time points tested. In contrast, the levels of intracellu-
lar PGs increased markedly over time. Notably, the

level of intracellular PGs did not reach a plateau, as
observed for the proteases (Fig. 3B). Rather, the level
of cell-associated PGs showed a continuous increase
over time (Fig. 6A). Notably, the latter is in good
agreement with the relatively late appearance of May–
Gru
¨
nwald ⁄ Giemsa-positive granules as compared to
the onset of SG mRNA expression (see above).
Next, the possibility that the different abilities of
MCs to store proteases at different stages of differenti-
ation could be due to differences in the electrostatic
charge of the PGs was addressed. For this purpose,
MC PGs recovered at different time points during the
course of MC development were examined by anion
exchange chromatography. At early time points (day
10), there was a distinction between two separate PG
populations, one with low anionic charge density
[coeluting with standard chondroitin sulphate A
(CS-A)] and one population with a markedly higher
charge (coeluting with standard pig mucosal heparin)
(Fig. 6B). Similar elution profiles were seen for PGs
recovered from the cell layer and from conditioned
medium. In contrast, only highly charged PGs were
seen at day 23 (Fig. 6B) and day 34 (not shown), again
with similar charge densities being displayed by cell-
associated and extracellular PGs. Together, these results
indicate that the MC maturation process is associated
with both increased total synthesis of sulfated PGs and
increased charge density of the synthesized PGs.

Discussion
Although the knockout of both SG [10] and N-de-
acetylase ⁄ N-sulfotransferase-2 [17,18], the latter being
an enzyme involved in the sulfation of heparin chains
attached to the SG core protein, has provided strong
evidence for a crucial role of PGs in mediating the
storage of secretory granule compounds in MCs, the
mechanism behind these observations has not been
established. One potential mechanism would be that
SG is important for the formation of the secretory
granule. However, we show here that SG
– ⁄ –
MCs also
displayed clearly discernible secretory vesicle-like struc-
tures. By conventional microscopy, such vesicles were
May–Gru
¨
nwald ⁄ Giemsa-negative and, interestingly,
May–Gru
¨
nwald ⁄ Giemsa-negative vesicles were also
seen in SG
+ ⁄ +
cells at early stages of differentiation.
Most likely, these structures represent immature secre-
tory granules in which the PG content is too low to
stain with May–Gru
¨
nwald ⁄ Giemsa. In accordance with
this, it was observed that the intracellular content of

highly sulfated PGs was relatively low at the corres-
ponding (early) time point. At later stages of differenti-
ation, in contrast, SG
+ ⁄ +
MCs showed staining with
May–Gru
¨
nwald ⁄ Giemsa, and this correlated well with
a marked increase in the recovery of highly sulfated
intracellular PGs.
The presence of secretory vesicle-like structures in
SG
– ⁄ –
cells was also supported at the ultrastructural
level. TEM analysis showed the presence of highly elec-
tron-dense granules in SG
+ ⁄ +
cells, but also showed
Fig. 5. Inhibition of lysosomal proteases. Serglycin (SG)
+ ⁄ +
and
SG
– ⁄ –
cells (after 5 weeks of culture) were incubated for 6 h with
20 m
M NH
4
Cl, or for 20 h with 5 mM NH
4
Cl. Cell extracts were

subsequently subjected to immunoblot analysis using antisera
towards carboxypeptidase A (CPA) and mouse mast cell protease
(mMCP)-6. The arrow indicates a ‘semiprocessed’ form of CPA.
The results shown are representative of three independent experi-
ments.
F. Henningsson et al. Role of serglycin in secretory granule assembly
FEBS Journal 273 (2006) 4901–4912 ª 2006 The Authors Journal compilation ª 2006 FEBS 4907
an abundance of granule-like organelles in SG
– ⁄ –
cells.
Importantly, however, the granule matrix in SG
– ⁄ –
cells
was more amorphous than in SG
+ ⁄ +
cells, and showed
less defined dense core formation. Our results also pro-
vide evidence that SG
– ⁄ –
cells retain the full capability
to undergo stimulus-induced degranulation, as deter-
mined by the ability to release b-hexosaminidase in
response to calcium ionophore. Together, our data thus
indicate that SG PG is not necessary for the formation
of MC secretory granules, and nor is SG involved in
mechanisms of degranulation.
Another possible explanation for the storage defects
seen in SG
– ⁄ –
MCs would be that SG PG is needed for

correct intracellular sorting of the MC proteases into
the secretory granules, the alternative fate being secre-
tion by the constitutive pathway. If this was the case,
it would be expected that SG-binding compounds such
as the MC proteases would be preferentially released
into the extracellular space by SG
– ⁄ –
MCs. Such mis-
sorting would result in excessive accumulation of
granule compounds in conditioned medium from
SG
– ⁄ –
cells. We here provide evidence that CPA, in its
pro-form, is secreted at higher levels by SG
– ⁄ –
cells
than by their SG
+ ⁄ +
counterparts, indeed indicating
that the lack of SG PG causes increased constitutive
release. However, rerouting into the constitutive path-
way of secretion does not seem to be a general effect
on all MC proteases when SG PG is lacking, as shown
0.0
0.5
1.0
1.5
2.0
2.5
0.0

0.5
1.0
1.5
2.0
2.5
0.0
0.5
1.0
1.5
2.0
2.5
0.0
0.5
1.0
1.5
2.0
2.5
20 40 60
0
100
200
300
400
500
0
100
200
300
400
500

20 40 60
0
100
200
300
0
100
200
300
20 40 60 20 40 60
A
530
; LiCl concentration (M
-1
)
Fraction number Fraction number
Cell fraction Medium fraction
B
Day 23
Day 10
Day 23
Day 10
hep
CS
0
2000
4000
6000
8000
10000

Day 10 Day 23 Day 34
A

cell fractions
medium fractions
Fig. 6. Analysis of sulfated proteoglycans.
Serglycin (SG)
+ ⁄ +
bone marrow cells were
cultured for different times (10, 23 34 days)
in medium containing interleukin (IL)-3 and
were biosynthetically labeled with
35
SO
4
2–
(A) Total recovery of
35
S-labeled proteogly-
cans per 1 · 10
6
cells into cell (filled bars)
and medium (hatched bars) fractions. (B)
Charge density of sulfated proteoglycans.
35
S-labeled proteoglycans isolated from cell
and medium fractions, both derived from
SG
+ ⁄ +
cells, were mixed with internal stand-

ards of heparin (hep) and chondroitin sulfate
(CS) and were subjected to anion exchange
chromatography on a DEAE–Sephacel col-
umn. The column was eluted with a linear
gradient of LiCl. Fractions were analyzed for
radioactivity (filled symbols) and for uronic
acid in order to detect the internal standards
(open symbols; A
530
). The results shown are
representative of two independent experi-
ments.
Role of serglycin in secretory granule assembly F. Henningsson et al.
4908 FEBS Journal 273 (2006) 4901–4912 ª 2006 The Authors Journal compilation ª 2006 FEBS
by the fact that mMCP-6 and mMCP-5 were found at
similar or higher levels in medium from SG
+ ⁄ +
cells
than in medium from their SG
– ⁄ –
counterparts. More-
over, our data provide evidence that the SG-dependent
MC proteases are in fact present in functional secre-
tory granules even in the absence of SG PG. This indi-
cates that transport of MC proteases into the secretory
granule compartments can occur independently of
SG PG. Hence, SG does not appear to function as
general sorting vehicle for MC proteases.
From the data presented here and previously [10], it is
clear that knockout of the SG gene results in a major

reduction of mMCP-5, mMCP-6 and CPA storage.
However, the blockade is not complete, indicating that
the proteases can actually be stored to some extent in
MC granules even in the absence of SG PGs to which
they can bind. In turn, this may be explained by low lev-
els of granular storage in the absence of any partner PG.
An alternative explanation could be that there are low
levels of PGs other than SG in the MC granule, and that
such PG species can provide some compensation for the
lack of SG in terms of promoting MC protease storage.
However, the presence of non-SG PG species within
MC granules remains to be demonstrated.
So, how does the lack of SG cause such a dramatic
reduction of stored MC proteases? Although general
mechanisms involved in the intracellular sorting of
granule components are still relatively poorly defined,
two major hypotheses have emerged: ‘sorting by entry’
and ‘sorting by retention’ [19]. In the sorting by entry
hypothesis (e.g. the mannose 6-phosphate system [20]),
each secretory granule compound has a unique sorting
signal that interacts with a cognate receptor on the
luminal side of specific regions in the trans-Golgi
network, leading to budding from the trans-Golgi
network of vesicles containing only molecules with the
corresponding specific sorting signals. In the sorting by
retention hypothesis, certain compounds entering the
immature granules may carry sorting motifs that inter-
act with the limiting membrane, but luminal proteins
that are not associated with the trans-Golgi network
membrane may also be included in the budding vesicle.

According to this hypothesis, the contents of the imma-
ture granule are subsequently refined, both by conden-
sation of selected compounds and by removal of others
by vesicular transport, e.g. to lysosomes for destruction,
or to the extracellular space by ‘constitutive-like’ or
‘piecemeal’ exocytosis [19,21]. This process will eventu-
ally result in secretory granule maturation. Although
we cannot at this stage with certainty define the role of
SG PG in this process, our results are clearly compat-
ible with a model in which SG organizes secretory gran-
ule maturation according to the sorting by retention
hypothesis. In support of this, all of the MC proteases
that have been shown to be SG-dependent for storage,
i.e. mMCP-4, mMCP-5, mMCP-6 and CPA (this study
and [10]), display high affinity for sulfated glycosami-
noglycans [22–25]. It is therefore possible that mMCP-
4, mMCP-5, mMCP-6 and CPA are transported into
immature granules independently of SG, but that their
retention within the granules is dependent on their tight
electrostatic interaction with SG PG. However, our
data indicate that interaction with SG PG is not an
absolute prerequisite for retention of all granule com-
pounds within the granule, as shown by the lack of SG
dependence for the storage of b-hexosaminidase.
The sorting for retention model of granule matur-
ation implies that compounds not selected for retent-
ion are expelled from the maturing granule by
vesicular transport. In line with this model, our results
suggest that mMCP-5 is targeted to degradation if not
retained by SG. We also see a marked secretion of

pro-CPA by SG
– ⁄ –
cells, possibly as a consequence
of defective retention. However, there is also secretion
of CPA, albeit in its mature form, from SG
+ ⁄ +
cells.
mMCP-6 is also secreted by SG
– ⁄ –
cells, but in con-
trast to pro-CPA and mature CPA, mMCP-6 secretion
was somewhat higher in SG
+ ⁄ +
cells than in their
SG
– ⁄ –
counterparts. However, the level of mMCP-6
protein in the conditioned medium from SG
– ⁄ –
cells
was considerably higher than the intracellular level,
indicating that secretion rather than storage is the
dominating pathway for mMCP-6 in the absence of
SG. One possible explanation for these findings is that
there is continuous low-level release of secretory gran-
ule compounds in normal MCs, a process often
referred to as ‘piecemeal’ degranulation [21]. In the
absence of SG as a retention vehicle, this slow release
may constitute the dominating pathway.
In summary, this study has provided the first

insights into the mechanism by which SG PG regulates
MC secretory granule homeostasis.
Experimental procedures
Cell culture
Experiments were performed on SG
+ ⁄ +
and SG
– ⁄ –
mice,
back-crossed to C57BL ⁄ 6J for 10 generations. The experi-
ments were approved by the local ethical committee.
BMMCs were obtained by culturing femural and tibial
bone marrow cells in DMEM (SVA, Uppsala, Sweden),
supplemented with 10% heat-inactivated fetal bovine serum
(Biotech line AS), 50 lgÆmL
)1
gentamicin sulfate (SVA),
2mml-glutamine (SVA) and 50% WEHI-3B conditioned
medium (which contains IL-3) for 3 weeks. The cells were
F. Henningsson et al. Role of serglycin in secretory granule assembly
FEBS Journal 273 (2006) 4901–4912 ª 2006 The Authors Journal compilation ª 2006 FEBS 4909
kept at a concentration of 500 000 cellsÆmL
)1
, and the med-
ium was changed every third day.
Staining
Three hundred thousand cells were collected on cytospin
slides (700 r.p.m., 5 min) and stained with May–Gru
¨
nw-

ald ⁄ Giemsa. The slides were first fixed in methanol for
3 min, and then stained with May–Gru
¨
nwald (Merck, Sol-
lentuna, Sweden) for 15 min. After being washed with
water, the slides were stained with 5% Giemsa (Merck) in
water for 10 min.
TEM
Cells were fixed for 6 h in 2% glutaraldehyde in a 0.1 m
sodium cacodylate buffer supplemented with 0.1 m sucrose,
and this was followed by 1.5 h of postfixation in 1%
osmium tetroxide dissolved in the same cacodylate buffer.
After dehydration in ethanol, the cells were embedded in
the epoxy resin Agar 100 (Agar Scientific, Stansted, UK).
Ultrathin sections were placed on copper grids covered with
a film of polyvinyl formal plastic (Formvar; Agar Scientific)
and contrasted with uranyl acetate and lead citrate. Elec-
tron micrographs were taken with a Hitachi electron micro-
scope (Hitachi Ltd, Tokyo, Japan).
RT-PCR
Total RNA was isolated using the NucleoSpin RNA II kit
(Macherey Nagel, Du
¨
ren, Germany). Total RNA was used
for first-strand cDNA synthesis using SuperScriptII for
RT-PCR using primers specific for the MC proteases and
SG. Hypoxanthine–guanine phosphoribosyltransferase was
used as a positive control for the RT-PCR. The PCR prim-
ers used were as specified elsewhere [10].
Immunoblotting

At each studied time point, 2 · 10
6
cells were taken from the
main cultures, centrifuged at 300 g, (Megafuge 1.0R; Heraeus;
equipped with a swing out rotor), resuspended in 1 mL of
serum-free medium (as described above), and further cultured
overnight. The cells were thereafter pelleted by centrifugation
at 300 g (10 min, Megafuge 1.0R; Heraeus; equipped with a
swing out rotor), and both the pellet and the medium fraction
were recovered and stored at ) 20 °C until analysis. Before
analysis, the recovered media were concentrated 50 times
using Amicon Ultra-4 centrifugal filter device (Millipore,
Solna, Sweden) and then mixed with 20 lL of SDS ⁄ PAGE
sample buffer containing 5% b-mercaptoethanol. Cell pellets
(1 · 10
6
cells) were dissolved in 300 lL of SDS ⁄ PAGE sample
buffer containing 5% b-mercaptoethanol. Immunoblotting
was carried out as previously described [10].
Proteoglycan isolation and analysis
SG
+ ⁄ +
cells (20 · 10
6
) from days 9, 22 and 33 of culture
were biosynthetically labeled overnight with 0.32 mCi of
carrier-free
35
SO
4

2–
(GE Healthcare, Uppsala, Sweden).
Cells were pelleted by centrifugation for 10 min at 300 g
(Megafuge 1.0R; Heraeus; equipped with a swing out
rotor). Cells and conditioned media were stored at ) 20 °C
until further analysis. For isolation of cell fraction glycos-
aminoglycans, cell pellets were solubilized in 500 lLof
NaCl ⁄ P
i
⁄ 2 m NaCl ⁄ 0.5% Triton X-100 at 4 °C for 30 min.
Then, the solubilisates were diluted with 9.5 mL of
H
2
O ⁄ 0.5% Triton X-100 and applied to columns contain-
ing 0.4 mL of DEAE–Sephacel, equilibrated with 50 mm
Tris ⁄ 0.1 m NaCl ⁄ 0.1% Triton X-100 (pH 8.0). Conditioned
media were loaded directly onto the columns. After wash-
ing with 10 mL of 50 mm Tris ⁄ HCl (pH 8.0) ⁄ 0.1 m
NaCl ⁄ 0.1% Triton X-100 and 10 mL of 50 mm sodium
acetate ⁄ 0.15 m NaCl ⁄ 0.1% Triton X-100 (pH 4.0), samples
were eluted with 50 mm sodium acetate ⁄ 2 m NaCl
(pH 5.5). Four fractions of 1100 lL each were collected
and analyzed for radioactivity by scintillation counting.
Fractions containing radioactive material were pooled and
diluted with H
2
O to yield 0.05 m NaCl, and subjected to
anion exchange chromatography on a 5 mL column of
DEAE–Sephacel connected to an FPLC system. The col-
umn was eluted with a gradient of increasing concentra-

tions of LiCl, from 0.05 m to 2 m in 50 mm sodium acetate
(pH 4.0) at a flow rate of 0.5 mLÆmin
)1
. Fractions (0.5 mL)
were collected and analyzed for
35
S radioactivity. As an
internal standard, 200 lL of a mixture of unlabeled heparin
(4 mgÆmL
)1
) and CS-A (5 mgÆmL
)1
; Sigma, Stockholm,
Sweden) was added to each sample before anion exchange
chromatography analysis. Internal standards were detected
by the carbazole method: 25 lL of each fraction was mixed
with 300 lL of concentrated H
2
SO
4
⁄ 25 mm K
2
B
4
O
7
and
10 lL of carbazole reagent (0.125% carbazole in 96% eth-
anol). The samples were boiled for 10 min and cooled, and
the absorbance at 530 nm was measured.

Inhibition of proteasome and lysosome function
Mature BMMCs (1 · 10
6
) were cultured in 5 lm lactacystin
(Affiniti Research Products, Exeter, UK). After incubation
overnight, cells were pelleted, solubilized and subjected to
immunoblotting for mMCP-5, mMCP6 and CPA. Lyso-
somal function was inhibited by incubation of cells with 5
or 20 mm NH
4
Cl. After 6–20 h, cells were pelleted, solubi-
lized and subjected to immunoblotting.
Degranulation
To induce MC degranulation, 2 · 10
6
cells were incubated
for 120 min in the presence of 2 lm of the calcium iono-
Role of serglycin in secretory granule assembly F. Henningsson et al.
4910 FEBS Journal 273 (2006) 4901–4912 ª 2006 The Authors Journal compilation ª 2006 FEBS
phore A23187 (Sigma-Aldrich, Stockholm, Sweden). Cell
fractions and conditioned media were recovered and subjec-
ted to immunoblot analysis. In addition, cytospin slides
were prepared.
Immunohistochemistry
Cytospin slides were stained with rabbit anti-mMCP-6
serum (1 : 200 dilution in NaCl ⁄ P
i
). Biotinylated goat
anti-(rabbit IgG) serum (affinity purified; Vector Laborat-
ories, Burlingame, CA) was used as secondary antibody

(1 : 100 dilution in NaCl ⁄ P
i
). Staining was performed
with a standard protocol using the biotin–avidin-based
Vectastain Elite kit (Vector Laboratories) and diamino-
benzadine (DAB) for detection of the secondary anti-
body, according to the protocol supplied by the
manufacturer. As a negative control, preimmune serum
was used. After staining, slides were dehydrated in
increasing concentrations (30–100%) of ethanol and
mounted with VectaMount mounting medium (Vector
Laboratories).
b-Hexosaminidase assay
Mature BMMCs (from day 50 of culture), 2 · 10
6
of each
genotype (SG
+ ⁄ +
and SG
– ⁄ –
), were washed in Tyrode’s
buffer (130 mm NaCl ⁄ 5mm KCl ⁄ 1.4 mm CaCl
2
⁄ 1mm
MgCl
2
⁄ 5.6 mm glucose ⁄ 10 mm Hepes ⁄ 0.1% BSA, pH 7.4)
and were then incubated at 37 °Cin5%CO
2
in 300 lLof

Tyrode’s buffer. To half of the cells, A23187 was added
(final concentration 2 lm), whereas the other half was left
without additions. Cells were incubated for 1, 2 or 4 h.
Samples (50 lL) were taken at each time point, and cells
were centrifuged at 300 g for 10 min (Megafuge 1.0R;
Heraeus; equipped with a swing out rotor). The remaining
supernatant was incubated with 100 lLof1mm p-nitro-
phenyl-N-acetyl-b-d-glucosaminide (Sigma-Aldrich) in
0.05 m citrate buffer (pH 4.5) at 37 °C for 1 h. As a control
for total b-hexosaminidase content, cells were lysed with
Tyrode’s buffer ⁄ 1% Triton X-100 and incubated as above.
All reactions were quenched by addition of 100 lLof
0.05 m Na
2
CO
3
(pH 10.0). The absorbance of each reaction
was read at 405 nm. In addition, cell pellets containing
1 · 10
6
cells recovered at six different time points (days 0,
6, 12, 20, 26 and 33) were lysed with 1% Triton X-100 in
Tyrode’s buffer and assayed for total b-hexosaminidase
content as described above.
Acknowledgements
This work was supported by grants from the Swedish
Research Council, Formas, Mizutani Foundation for
Glycoscience and King Gustaf V’s 80th Anniversary
Fund.
References

1 Metcalfe DD, Baram D & Mekori YA (1997) Mast
cells. Physiol Rev 77, 1033–1079.
2 Galli SJ, Nakae S & Tsai M (2005) Mast cells in the
development of adaptive immune responses. Nat Immu-
nol 6, 135–142.
3 Stevens RL & Austen KF (1989) Recent advances
in the cellular and molecular biology of mast cells.
Immunol Today 10, 381–386.
4 Miller HR & Pemberton AD (2002) Tissue-specific
expression of mast cell granule serine proteinases and
their role in inflammation in the lung and gut. Immunol-
ogy 105, 375–390.
5 Caughey G (2001) New developments in the genetics
and activation of mast cell proteases. Mol Immunol 38,
1353–1357.
6 Kolset SO, Prydz K & Pejler G (2004) Intracellular pro-
teoglycans. Biochem J 379, 217–227.
7 Dayton ET, Pharr P, Ogawa M, Serafin WE, Austen KF,
Levi-Schaffer F & Stevens RL (1988) 3T3 fibroblasts
induce cloned interleukin 3-dependent mouse mast cells
to resemble connective tissue mast cells in granular con-
stituency. Proc Natl Acad Sci USA 85, 569–572.
8 Blank U & Rivera J (2004) The ins and outs of IgE-
dependent mast-cell exocytosis. Trends Immunol 25,
266–273.
9 Turner H & Kinet JP (1999) Signalling through the
high-affinity IgE receptor Fc epsilonRI. Nature 402,
B24–B30.
10 Abrink M, Grujic M & Pejler G (2004) Serglycin is
essential for maturation of mast cell secretory granule.

J Biol Chem 279, 40897–40905.
11 Tsai M, Takeishi T, Thompson H, Langley KE, Zsebo
KM, Metcalfe DD, Geissler EN & Galli SJ (1991)
Induction of mast cell proliferation, maturation, and
heparin synthesis by the rat c-kit ligand, stem cell fac-
tor. Proc Natl Acad Sci USA 88, 6382–6386.
12 Razin E, Ihle JN, Seldin D, Mencia-Huerta JM, Katz
HR, LeBlanc PA, Hein A, Caulfield JP, Austen KF &
Stevens RL (1984) Interleukin 3: a differentiation and
growth factor for the mouse mast cell that contains
chondroitin sulfate E proteoglycan. J Immunol 132,
1479–1486.
13 Henningsson F, Yamamoto K, Saftig P, Reinheckel T,
Peters C, Knight SD & Pejler G (2005) A role for cathe-
psin E in the processing of mast-cell carboxypeptidase
A. J Cell Sci 118, 2035–2042.
14 Diamant B & Patkar SA (1975) Stimulation and inhibi-
tion of histamine release from isolated rat mast cells.
Dual effects of the ionophore A23187. Int Arch Allergy
Appl Immunol 49, 183–207.
15 Dragonetti A, Baldassarre M, Castino R, Demoz M,
Luini A, Buccione R & Isidoro C (2000) The lysosomal
F. Henningsson et al. Role of serglycin in secretory granule assembly
FEBS Journal 273 (2006) 4901–4912 ª 2006 The Authors Journal compilation ª 2006 FEBS 4911
protease cathepsin D is efficiently sorted to and secreted
from regulated secretory compartments in the rat baso-
philic ⁄ mast cell line RBL. J Cell Sci 113, 3289–3298.
16 Wolters PJ, Laig-Webster M & Caughey GH (2000)
Dipeptidyl peptidase I cleaves matrix-associated proteins
and is expressed mainly by mast cells in normal dog air-

ways. Am J Respir Cell Mol Biol 22, 183–190.
17 Humphries DE, Wong GW, Friend DS, Gurish MF, Qiu
WT, Huang C, Sharpe AH & Stevens RL (1999) Heparin
is essential for the storage of specific granule proteases
in mast cells [see comments]. Nature 400, 769–772.
18 Forsberg E, Pejler G, Ringvall M, Lunderius C, Toma-
sini-Johansson B, Kusche-Gullberg M, Eriksson I,
Ledin J, Hellman L & Kjellen L (1999) Abnormal mast
cells in mice deficient in a heparin-synthesizing enzyme
[see comments]. Nature 400, 773–776.
19 Arvan P & Castle D (1998) Sorting and storage during
secretory granule biogenesis: looking backward and
looking forward. Biochem J 332, 593–610.
20 Ghosh P, Dahms NM & Kornfeld S (2003) Mannose
6-phosphate receptors: new twists in the tale. Nat Rev
Mol Cell Biol 4, 202–212.
21 Dvorak AM (2005) Piecemeal degranulation of basoph-
ils and mast cells is effected by vesicular transport of
stored secretory granule contents. Chem Immunol
Allergy 85, 135–184.
22 Pejler G & Maccarana M (1994) Interaction of heparin
with rat mast cell protease 1. J Biol Chem 269, 14451–
14456.
23 Hallgren J, Ba
¨
ckstro
¨
m S, Estrada S, Thuveson M &
Pejler G (2004) Histidines are critical for heparin-depen-
dent activation of mast cell tryptase. J Immunol 173,

1868–1875.
24 Serafin WE, Katz HR, Austen KF & Stevens RL (1986)
Complexes of heparin proteoglycans, chondroitin sulfate
E proteoglycans, and [3H]diisopropyl fluorophosphate-
binding proteins are exocytosed from activated mouse
bone marrow-derived mast cells. J Biol Chem 261,
15017–15021.
25 Springman EB, Dikov MM & Serafin WE (1995)
Mast cell procarboxypeptidase A. Molecular
modeling and biochemical characterization of its
processing within secretory granules. J Biol Chem 270,
1300–1307.
Role of serglycin in secretory granule assembly F. Henningsson et al.
4912 FEBS Journal 273 (2006) 4901–4912 ª 2006 The Authors Journal compilation ª 2006 FEBS