Tài liệu Báo cáo khoa học: Golgi reassembly stacking protein 55 interacts with membrane-type (MT) 1-matrix metalloprotease (MMP) and furin and plays a role in the activation of the MT1-MMP zymogen pdf

Golgi reassembly stacking protein 55 interacts with
membrane-type (MT) 1-matrix metalloprotease (MMP) and
furin and plays a role in the activation of the MT1-MMP
zymogen
Christian Roghi
1,2
, Louise Jones
2
*, Matthew Gratian
2
, William R. English
1,2
and Gillian Murphy
1,2
1 Cancer Research UK Cambridge Research Institute, The Li Ka Shing Centre, UK
2 Cambridge Institute for Medical Research, UK
Keywords
furin; GRASP55; intracellular traffic;
MT1-MMP; protease
Correspondence
C. Roghi, Cancer Research UK Cambridge
Research Institute, The Li Ka Shing Centre,
Robinson Way, Cambridge CB2 0RE, UK
Fax: +44 (0)1223 404573
Tel: +44 (0)1223 404472
E-mail:
*Present address
KuDOS Pharmaceuticals Ltd, Cambridge
Science Park, UK
(Received 25 March 2010, revised 14 May
2010, accepted 28 May 2010)

doi:10.1111/j.1742-4658.2010.07723.x
Membrane-type 1 matrix metalloproteinase (MT1-MMP) is a proteinase
involved in the remodelling of extracellular matrix and the cleavage of a
number of substrates. MT1-MMP is synthesized as a zymogen that requires
intracellular post-translational cleavage to gain biological activity. Furin,
a member of the pro-protein convertase family, has been implicated in the
proteolytic removal of the MT1-MMP prodomain sequence. In the present
study, we demonstrate a role for the peripheral Golgi matrix protein
GRASP55 in the furin-dependent activation of MT1-MMP. MT1-MMP
and furin were found to co-localize with Golgi reassembly stacking protein
55 (GRASP55). Further analysis revealed that GRASP55 associated with
the cytoplasmic domain of both proteases and that the LLY
573
motif in the
MT1-MMP intracellular domain was crucial for the interaction with
GRASP55. Overexpression of GRASP55 was found to enhance the forma-
tion of a complex between MT1-MMP and furin. Finally, we report that
disruption of the interaction between GRASP55 and furin led to a reduc-
tion in pro-MT1-MMP activation. Taken together, these data suggest that
GRASP55 may function as an adaptor protein coupling MT1-MMP with
furin, thus leading to the activation of the zymogen.
Structured digital abstract
l
MINT-7897990: Furin (uniprotkb:P09958) and GRASP55 (uniprotkb:Q9H8Y8) colocalize
(
MI:0403)byfluorescence microscopy (MI:0416)
l
MINT-7897801: GRASP55 (uniprotkb:Q9R064) physically interacts (MI:0915) with MT2-
MMP (uniprotkb:
P51511)bytwo hybrid (MI:0018)

l
MINT-7897821: GRASP55 (uniprotkb:Q9R064) physically interacts (MI:0915) with MT3-
MMP (uniprotkb:
P51512)bytwo hybrid (MI:0018)
l
MINT-7897577: GRASP55 (uniprotkb:Q9R064) and MT1-MMP (uniprotkb:P50281) coloca-
lize (
MI:0403)byfluorescence microscopy (MI:0416)
l
MINT-7897366: MT1-MMP (uniprotkb:P50281) physically interacts (MI:0915) with
GRASP55 (uniprotkb:
Q9H8Y8)byanti bait coimmunoprecipitation (MI:0006)
Abbreviations
ECM, extracellular matrix; EGFP, enhanced green fluorescent protein; EYFP, enhanced yellow fluorescent protein; FACS, fluorescence-
activated cell sorting; GFP, green fluorescent protein; GRASP, Golgi reassembly stacking protein; GRASP55F, FLAG-tagged GRASP55;
IB, immunoblotting; ICD, intracellular domain; M2H, mammalian two-hybrid; MMP, matrix metalloprotease; MT1/EYFP, EYFP-tagged
MT1-MMP; MT1/MYC, Myc-tagged MT1-MMP; MT-MMP, membrane-type MMP; PDZ, PSD-95/SAP90 Drosophila septate junction protein
discs-large and epithelial tight junction ZO-1; TGF, transforming growth factor; TGN, trans-Golgi network.
3158 FEBS Journal 277 (2010) 3158–3175 ª 2010 The Authors Journal compilation ª 2010 FEBS
Introduction
Extracellular matrix (ECM) remodelling is a crucial
process occurring during cell migration and invasion
in various physiological (i.e. embryonic development,
ovulation, angiogenesis, wound healing) and patho-
logical processes, including rheumatoid arthritis,
tumour growth, invasion and metastasis [1]. Of all
the different proteolytic systems involved in ECM
turnover, the matrix metalloproteinases (MMPs) have
been reported to exert a dominant effect [2]. MMPs
are a large family of structurally and functionally

related multi-domain zinc-dependent endopeptidases
that collectively are able to degrade virtually all pro-
teins of the ECM. MMPs are mainly soluble
enzymes released by the cell in the extracellular
milieu, although membrane-bound MMPs (membrane
type-MMPs or MT-MMPs) have also been identified
and are ideally positioned for regulating pericellular
proteolysis [3].
Membrane-type 1 matrix metalloproteinase (MT1-
MMP; MMP14; EC 3.4.24.80) is by far the most
extensively studied member of the MT-MMP sub-
family. MT1-MMP is a type 1 transmembrane MMP
involved in pericellular ECM turnover [4], as well as in
the proteolytic processing of cell surface receptors
[4,5]. MT1-MMP is also involved in the activation of
pro-MMP2 and pro-MMP13, leading to the indirect
increase in its repertoire of substrates [6,7].
MT1-MMP has a wide spectrum of cellular func-
tions [5,8]. Elevated MT1-MMP expression, which is
well documented in many tumours, has been correlated
with key processes of tumour progression [9,10],
including angiogenesis [11], cell migration and invasion
[12], cell growth [13], and metastatic spread. Inhibition
or silencing of the protease has been found to
significantly reduce the invasive phenotype of tumour
cells, implicating a leading role for MT1-MMP in such
processes [12,14].
There is mounting evidence that the short intracellu-
lar domain (ICD) of MT1-MMP (21 amino acids)
plays an important role in multiple MT1-MMP-medi-

ated cellular events [15]. MT1-MMP ICD has been
involved in cell migration [16] and invasion into recon-
stituted basement membrane [17,18]. The MT1-MMP
ICD is also critical for the intracellular trafficking of
the enzyme [19–23] and its targeting to invadopodia in
invasive cells [24]. The ICD of MT1-MMP has been
found to modulate multiple signal transduction path-
ways [16,25–27] and participates in the homophilic
interaction between MT1-MMP monomers [28].
Recently, the LL
572
di-leucine motif has been reported
l
MINT-7897617, MINT-7897659, MINT-7897681, MINT-7897702, MINT-7897725, MINT-
7898032, MINT-7898011, MINT-7897907, MINT-7897884: GRASP55 (uniprotkb:Q9R064)
physically interacts (
MI:0915) with MT1-MMP (uniprotkb:P50281)bytwo hybrid (MI:0018)
l
MINT-7898002: MT1-MMP (uniprotkb:P50281) physically interacts (MI:0914) with Furin
(uniprotkb:
P09958)byanti bait coimmunoprecipitation (MI:0006)
l
MINT-7897500: MT1-MMP (uniprotkb:P50281) and Giantin (uniprotkb:Q14789) colocalize
(
MI:0403)byfluorescence microscopy (MI:0416)
l
MINT-7897750, MINT-7897394: GRASP55 (uniprotkb:Q9R064) physically interacts
(
MI:0915) with MT1-MMP (uniprotkb:P50281)byanti tag coimmunoprecipitation (MI:0007)
l

MINT-7897562: MT1-MMP (uniprotkb:P50281) and GRASP55 (uniprotkb:Q9H8Y8) coloca-
lize (
MI:0403)byfluorescence microscopy (MI:0416)
l
MINT-7897512: TGN46 (uniprotkb:O43493) and MT1-MMP (uniprotkb:P50281) colocalize
(
MI:0403)byfluorescence microscopy (MI:0416)
l
MINT-7897921, MINT-7897975: GRASP55 (uniprotkb:Q9R064) physically interacts
(
MI:0915) with Furin (uniprotkb:P09958)bytwo hybrid (MI:0018)
l
MINT-7898052, MINT-7897410: MT1-MMP (uniprotkb:P50281) physically interacts
(
MI:0915) with GRASP55 (uniprotkb:Q9R064)byanti bait coimmunoprecipitation (MI:0006)
l
MINT-7897951: GRASP55 (uniprotkb:Q9R064) physically interacts (MI:0915) with PC7
(uniprotkb:
Q16549)bytwo hybrid (MI:0018)
l
MINT-7897866: GRASP55 (uniprotkb:Q9R064) physically interacts (MI:0915) with MT5-
MMP (uniprotkb:
Q9Y5R2)bytwo hybrid (MI:0018)
l
MINT-7897633: GRASP55 (uniprotkb:Q9R064) physically interacts (MI:0915) with TGFA
(uniprotkb:
P01135)bytwo hybrid (MI:0018)
l
MINT-7897551: GRASP55 (uniprotkb:Q9H8Y8) and Giantin (uniprotkb:Q14789) colocalize
(

MI:0403)byfluorescence microscopy (MI:0416)
l
MINT-7897938: GRASP55 (uniprotkb:Q9R064) physically interacts (MI:0915) with PC5/6B
(uniprotkb:
Q04592)bytwo hybrid (MI:0018)
C. Roghi et al. Role of GRASP55 in MT1-MMP activation
FEBS Journal 277 (2010) 3158–3175 ª 2010 The Authors Journal compilation ª 2010 FEBS 3159
to influence the O-glycosylation pattern of MT1-MMP
[29]. Post-translational modifications of the MT1-
MMP ICD have also been reported with the palmitoy-
lation of the cysteine 574 (C
574
) residue [30] and the
phosphorylation of the tyrosine 573 (Y
573
) [31] and
threonine 567 (T
567
) [32] residues. The MT1-MMP
ICD has been reported to interact with the multifunc-
tional protein p32/gC1qR [21], a protein with homol-
ogy to members of the Cupin superfamily (MTCBP-1)
[33], as well as with the l2 subunit of the clathrin-
coated pits adapter protein 2 (AP-2) [18] and phospho-
caveolin-1 in src overexpressing cells [34].
Golgi reassembly stacking protein 55 (GRASP55) is
a peripheral Golgi matrix protein that has been impli-
cated, in vitro, in the post-mitotic stacking of Golgi cis-
ternae [35]. Cryo-electron microscopy has shown that
GRASP55 is found predominantly in the medial-cister-

nae of the Golgi complex of HeLa cells [35].
GRASP55 interacts with Golgin-45 and the complex is
crucial for maintaining Golgi structure [36]. In addi-
tion to its contribution to the Golgi exoskeleton,
GRASP55 has also been reported to be involved in the
intracellular transport of pro-transforming growth fac-
tor (TGF)-a [37], CD8a or the frizzled receptor Fz4
[38], as well as in the Golgi retention of p24a, a mem-
ber of the p24 family of cargo receptors [39].
In the present study, we report in detail on the inter-
action between the MT1-MMP ICD and GRASP55
using a mammalian two-hybrid (M2H) system. Using
this approach, we have identified the GRASP55 bind-
ing site in the ICD of MT1-MMP, as well as the
GRASP55 domains involved in the interaction with
MT1-MMP ICD. We also describe the GRASP55
interaction with the furin ICD, and provide evidence
that GRASP55 could play an important role in the
furin-mediated proteolytic activation of the MT1-
MMP zymogen.
Results
MT1-MMP co-immunoprecipitates with GRASP55
Although the presence of MT1-MMP and GRASP55
(p59) in the same complex has been suggested by Kuo
et al. [37], the functional implications of this interac-
tion have yet to be fully investigated. In steady-state
HT1080 cells, MT1-MMP is mainly present at the cell
surface and in the endosomal compartment [22] and
virtually no protease can be detected in the Golgi
apparatus. We therefore transfected these cells with an

exogenous wild-type MT1-MMP cDNA (Fig. 1A),
aiming to detect the protease in the early secretory
pathway. Cells were then lysed and the extract was
A
60
50
40
IB: MT1-MMP
MT1-MMP
pCDNA3.1 Zeo+
kDa
+–
–+
+–
–+
60
50
40
IB: GRASP55
50
40
IB: β-actin
12
IgG
anti MT1-MMP
B
60
50
40
IB: GRASP55

1
–
––
–
+
+
+
+
2
kDa
GRASP55F
MT1/MYC
IP: FLAG
IB: MYC
kDa
60
C
IB: MYC
IB: FLAG
Input
lysates
60
60
1234
Fig. 1. MT1-MMP co-immunoprecipitates with GRASP55F. (A) Pro-
tein extracts prepared from HT1080 cells transiently transfected
with pCDNA3.1 Zeo+ (lane 1) or full-length MT1-MMP construct
were analyzed by IB with antibodies directed against MT1-MMP,
GRASP55 and b-actin. (B) Protein extract prepared from HT1080
transiently transfected with MT1-MMP were immunoprecipitated

with rabbit control IgGs (lane 1) or with the rabbit polyclonal anti-
body directed against MT1-MMP (lane 2) and analyzed by IB using
a monoclonal antibody to GRASP55. The arrow identifies immuno-
precipitated MT1-MMP. (C) Protein extracts prepared from HT1080
cells transiently transfected with pCDNA3.1 Zeo+ and MT1/MYC
(lane 2), pCDNA3.1 Zeo+ and GRASP55F (lane 3) and MT1/MYC
and GRASP55F (lane 4) were immunoprecipitated with the FLAG
M2 monoclonal antibody and the associated MT1-MMP was
detected by IB using the MYC tag monoclonal antibody. Expression
of the transfected construct was monitored in the input lysates
using specific antibodies. The black arrowhead indicates IgG (immu-
noglobulin heavy chain).
Role of GRASP55 in MT1-MMP activation C. Roghi et al.
3160 FEBS Journal 277 (2010) 3158–3175 ª 2010 The Authors Journal compilation ª 2010 FEBS
ABC
DEF
GHI
JKL
Fig. 2. Subcellular localization of GRASP55 and MT1-MMP in bbHT1080. Fixed and permeabilized bbHT1080 cells were incubated with anti-
bodies directed against MT1-MMP (A, D, J), GRASP55 (G, K), TGN46 (E) and giantin (B, H). The co-localization can be observed in yellow in
the merged panels (C, F, I, L). Arrowheads depict membranous structures where the proteins co-localize. Scale bar = 5 lm.
C. Roghi et al. Role of GRASP55 in MT1-MMP activation
FEBS Journal 277 (2010) 3158–3175 ª 2010 The Authors Journal compilation ª 2010 FEBS 3161
immunoprecipitated with nonspecific rabbit IgGs or
with the rabbit polyclonal antibody directed against
MT1-MMP. As shown in Fig. 1B, immunoprecipita-
tion of MT1-MMP led to the co-precipitation of a
small amount of endogenous GRASP55, as detected
by immunoblotting (IB). No GRASP55 was detected
when the pre-immune IgGs were used. Co-precipita-

tion between MT1-MMP and GRASP55 was also
observed in HT1080 cells expressing exogenous Myc-
tagged MT1-MMP (MT1/MYC) and FLAG-tagged
GRASP55 (GRASP55F) (Fig. 1C, lane 4) or HeLa
cells (Fig. S1). No MT1-MMP was detected in control
immunoprecipitations (Fig. 1C, lanes 1–3 and Fig. S1).
MT1-MMP co-localizes with GRASP55
The co-immunoprecipitation of MT1-MMP and
GRASP55 prompted us to investigate whether these
two proteins co-localized in the same membranous
compartment. To investigate this, we used HT1080 cells
stably expressing wild-type MT1-MMP (bbHT1080). In
these cells, MT1-MMP (Fig. 2A, 2D) co-localized
extensively with the medial Golgi marker giantin
(Fig. 2B) and with the trans-Golgi network (TGN)
membrane protein marker TGN46 (Fig. 2E). Endoge-
nous GRASP55 (Fig. 2G, 2K) was also found to
co-localize with giantin (Fig. 2H) [40] and a clear
co-localization with MT1-MMP (Fig. 2J) could also be
observed (Fig. 2L) in these cells. The co-localization of
MT1-MMP and GRASP55 was next assessed using live
video microscopy. We generated an enhanced yellow
fluorescent protein (EYFP)-tagged MT1-MMP con-
struct (MT1/EYFP), where the EYFP tag replaced the
entire MT1-MMP catalytic domain. In HT1080, the
intracellular trafficking of MT1/EYFP was indistin-
guishable from that of wild-type MT1-MMP and both
constructs were found to accumulate in the Golgi appa-
ratus and the TGN in these cells (C. Roghi, unpublished
data). HT1080 cells were then transiently co-transfected

with MT1/EYFP and the GRASP55-green fluorescent
protein (GFP) fusion protein as previously described
[35] and the localization of both proteins was studied in
live cells. Separation of the GFP and EYFP signals was
achieved using a Zeiss META confocal microscope (see
Fig. 3. Co-localization of MT1-MMP and GRASP55 in live cells. Four consecutive frames (4 s apart) of time lapse sequence collected from
HT1080 co-transfected cells with EYFP/MT1 and GRASP55-GFP. The arrowheads are examples of dynamic vesicles containing both fluore-
scent proteins. MT1/EYFP was pseudo-coloured in red during post-acquisition processing. The co-localization between GRASP55-GFP and
MT1/EYFP can be observed in yellow in the merged panels. Scale bar = 16 lm.
Role of GRASP55 in MT1-MMP activation C. Roghi et al.
3162 FEBS Journal 277 (2010) 3158–3175 ª 2010 The Authors Journal compilation ª 2010 FEBS
PGGGFFFRRHGTPRRLLYCQRSLLDKV
VP16-MT1
VP16
PGG
G
FFAAAHGTPRRLLYCQRSLLDKV
VP16VP16-MT1 FRR
A
PGGGFFFRRAAAPRRLLYCQRSLLDKV
VP16VP16-MT1 HGT
PGGGFFFRRHGTAAALLYCQRSLLDKV
VP16
VP16-MT1 PRR
PGGGFFFRRHGTPRRAAACQRSLLDKV
VP16
VP16-MT1 LLY
PGGGFFFRRHGTPRRLLYAAASLLDKV
VP16
VP16-MT1 CQR

PGGGFFFRRHGTPRRLLYCQRAAADKV
VP16VP16-MT1 SLL
PGGG
FFFRRHGTPRRLLYCQRSLL
AAA
VP16
VP16
-
MT1 DKV
B
KHCEWCRALICRHEKPSALLKGRTACCHSETVV
VP16-TGF-α VP16
PGGG
FFFRRHGTPRRLLYCQRSLL
AAA
VP16
VP16
MT1 DKV
VP16-MT1 Y
PGGGFFFRRHGTPRRLLACQRSLLDKV
VP16
VP16-MT1 LL
PGGGFFFRRHGTPRRAAYCQRSLLDKV
VP16
500
0 100 200 300 400
500
0 100 200 300 400
Luminescence (arbitrary units)
1

2
3
4
GAL4 VP16GRASP55
MT1
VP16
GAL4
GAL4 GRASP55
MT1VP16
GAL4 VP16
+
+
+
+
C
VP16
+
GAL4 GRASP55
+
TGF-α
VP16GAL4
+
GAL4 VP16
1
2
3
Luminescence (arbitrary units)
GAL4 GRASP55
TGF-α
VP16

+
4
Fig. 4. MT1-MMP interaction with
GRASP55. (A) Schematic representation of
VP16-TGF-a, VP16-MT1 and the VP16-MT1
mutant constructs. The mutated amino
acids are shown in bold and the PGGG
linker is shown in italics. (B) Interaction
between full-length GRASP55 (GAL4-
GRASP55) and MT1-MMP ICD (VP16-MT1)
or (C) TGF-a ICD (VP16-TGF-a) using the
M2H assay.
+
MT1
VP16
GAL4
500
0 100 200 300 400
1
Luminescence (arbitrary units)
A
B
GAL4 GRASP55
+
MT1 FRR
VP16
GAL4 GRASP55
+
MT1 HGT
VP16

GAL4 GRASP55
+
MT1 PRR
VP16
GAL4 GRASP55
+
MT1
VP16
GAL4 GRASP55
MT1 LLY
VP16
+
2
3
4
5
6
*
*
***
+
GAL4 GRASP55
+
MT1 CQR
VP16
GAL4 GRASP55
+
MT1 SLL
VP16
+

GAL4 GRASP55
MT1 DKV
VP16
7
8
9
**
**
MT1/MYC
+
+
+
+
++
+
GRASP55F
MT1 LLY/MYC
IP: FLAG
IB: MYC
50
+–––
–– –
––
WB: MYC
WB: FLAG
Input
lysates
12345
60
60

Fig. 5. The LLY motif in the MT1-MMP ICD
is crucial for the interaction with GRASP55.
(A) Interaction between GAL4-GRASP55 and
VP16-MT1 or MT1-MMP ICD triple mutants.
(B) Cell lysates prepared from HT1080 cells
transfected with pCDNA3.1 Zeo+ and MT1/
MYC (lane 1), pCDNA3.1 Zeo+ and
GRASP55F (lane 2), pCDNA3.1 Zeo+
and MT1 LLY/MYC (lane 3), GRASP55F and
MT1 LLY/MYC (lane 4) and GRASP55F
and MT1/MYC (lane 5) were immunoprecipi-
tated with the FLAG M2 antibody.
MT1-MMP present in the immunoprecipi-
tate was detected by IB using the MYC tag
antibody. Levels of transfected proteins
were monitored in input lysates using
specific antibodies.
C. Roghi et al. Role of GRASP55 in MT1-MMP activation
FEBS Journal 277 (2010) 3158–3175 ª 2010 The Authors Journal compilation ª 2010 FEBS 3163
Materials and methods). As previously described in
fixed bbHT1080 (Fig. 2), we clearly observed, in live
HT1080 cells, the presence of both tagged proteins in
the same membrane compartment (Fig. 3, merged), thus
confirming the results that were observed previously in
fixed bbHT1080 cells (Fig. 2). Interestingly, we also
noted that MT1/EYFP and GRASP55-GFP also
co-localized in very dynamic unidentified cytoplasmic
membranous structures (Fig. 3, arrowheads).
MT1-MMP intracellular domain is involved in the
interaction with GRASP55

If the co-immunoprecipitation between MT1-MMP and
GRASP55 is functionally relevant, there should be an
interaction between the peripheral scaffolding protein
and the ICD of MT1-MMP. To investigate this, we used
an M2H system. MT1-MMP ICD flanked by an N-ter-
minal PGGG linker was fused to the VP16 activation
domain (VP16-MT1; Fig. 4A) and full-length
GRASP55 was fused to the GAL4 DNA binding
domain (GAL4-GRASP55). Both constructs were tran-
siently co-transfected in HT1080 cells together with the
reporter plasmid pG5luc, which contains the firefly lucif-
erase gene under the control of five GAL4 binding sites.
After 24 h of transfection, the firefly luciferase activity
was measured, as described in the Materials and meth-
ods, and the values obtained were normalized according
to transfection efficiency using the Renilla reniformis
luciferase expressed by the pBIND vector. Co-expres-
sion of VP16-MT1 and GAL4-GRASP55 fusion
proteins (Fig. 4B, lane 4) in HT1080 resulted in the
production of significantly higher firefly luciferase lumi-
nescence compared to the controls (Fig. 4B, lanes 1–3),
demonstrating an interaction between the MT1-MMP
ICD and GRASP55 in the M2H system. Using this
assay, we also observed an interaction between the
VP16-TGF-a ICD (Fig. 4A) and GAL4-GRASP55
(Fig. 4C, lane 4), therefore confirming the interaction of
these two proteins previously observed biochemically
[37] or using a yeast two-hybrid assay [39]. Interestingly,
in the M2H system, the interaction between GRASP55
and TGF-a ICD did not require the oligomerization of

the TGF-a ICD as previously observed using a yeast
two-hybrid assay [39].
The MT1-MMP LLY motif is important for the
interaction with GRASP55
We next sought to define the nature of the GRASP55
binding site in the MT1-MMP ICD. pACT plasmids
driving the expression of MT1-MMP ICDs containing
single-, double- and triple-point mutations were gener-
ated (Fig. 4A) and used in the M2H system. System-
atic analysis of the interaction between the triple
MT1-MMP ICD mutants and GAL4-GRASP55
revealed that, apart from the VP16-MT1 FFR
(Fig. 5A, lane 3) and VP16-MT1 DKV mutants
(Fig. 5A, lane 9), all the other triple mutants (Fig. 5A,
lanes 4–8) displayed a marked and significant reduc-
tion of luciferase activity compared to the wild-type
VP16-MT1 construct (Fig. 5A, lane 2). In particular,
the mutation of the LLY
573
motif (LLY571-573AAA)
(Fig. 5A, lane 6) resulted in a complete inhibition of
MT3
MT3 ILY
MT5
VP16-MT1
VP16-MT2
PGGGFFFRRHGTPRRLLYCQRSLLDKV
VP16
PGGGVQMQRKGAPRVLLYCKRSL QEW V
VP16

% homology with
MT1-MMP ICD
-
57.1%
VP16-MT3
VP16-MT5
PGGGFQFKRKGTPRHILYCKRSMQEWV
VP16
PGGGFQFKNKTGPQPVTYYK RPVQEWV
VP16
57.1%
23.8%
500
A
B
Luminescence (arbitrary units)
0 100 200 300 400
+
GAL4
VP16
+
GAL4 GRASP55
VP16
+
GAL4
GRASP55
VP16
4
5
6

+
GAL4
VP16
GAL4 GRASP55
VP16
+
+
GAL4 GRASP55 VP16
1
2
3
**
GAL4
GRASP55

VP16
+
GAL4
VP16
+
GAL4 GRASP55 MT5VP16
+
GAL4 GRASP55
MT5 VTY
VP16
7
8
9
MT2
MT3

MT2
MT2 LLY
Fig. 6. The MT2-MMP LLY motif is involved
in the interaction with GRASP55. (A)
Schematic representation of the VP16-MT1,
VP16-MT2, VP16-MT3 and VP16-MT5
constructs. Amino acids conserved between
MT1-MMP ICD and either MT2-, MT3- or
MT5-MMP are shown in bold. (B)
Interactions between GAL4-GRASP55 and
VP16-MT2, VP16-MT2 LLY, VP16-MT3,
VP16-MT3 ILY, VP16-MT5 and VP16-MT5
VTY were tested using the M2H system.
**P < 0.001.
Role of GRASP55 in MT1-MMP activation C. Roghi et al.
3164 FEBS Journal 277 (2010) 3158–3175 ª 2010 The Authors Journal compilation ª 2010 FEBS
the interaction between MT1-MMP ICD and
GAL4-GRASP55. IB analysis revealed that all triple
mutants were expressed to a similar level (data not
shown), indicating that the differences in interaction
observed were not a result of impaired protein produc-
tion or stability. Our data therefore suggest that most
of the MT1-MMP ICD is implicated in the interaction
with GRASP55, with the LLY
573
motif playing a criti-
cal role in the interaction between the two proteins.
A reduction of luciferase activity was also observed
using the VP16-MT1 Y (MT1Y573A) (Fig. S2, lane 5)
and VP16-MT1 LL (LL570-572AA) mutants (Fig. S2,

lane 4), although not to the level observed with the
LLY570-573AAA triple mutant (Fig. S2, lane 3), dem-
onstrating that the mutation of the whole LLY
573
motif is needed to abolish the interaction of MT1-
MMP ICD with GRASP55.
The importance of the LLY
573
motif in the MT1-
MMP ICD observed in the M2H assay was next con-
firmed by co-immunoprecipitation. HT1080 cells were
transiently co-transfected with GRASP55F together
with MT1/MYC or MT1 LLY/MYC (LLY570-573
AAA) triple mutant and total cell lysates were
subjected to immunoprecipitation using the FLAG tag
monoclonal antibody. As previously observed, we
detected a clear interaction between GRASP55F and
the wild-type MYC-tagged MT1-MMP (Fig. 5B, lane
5) when both proteins were expressed in HT1080 cells.
Mutation of the LLY
573
motif to AAA
573
in MT1-
MMP ICD led to a marked reduction in the amount
of MT1-MMP present in the immunoprecipitated
material (Fig. 5B, lane 4), thus confirming the impor-
tant role of the LLY
573
motif in the interaction

between the protease and GRASP55. Interestingly,
mutation of the LLY
573
motif led to the detection
of pro-MT1-MMP in the immunoprecipitate, sugges-
ting that the disruption of the interaction between
MT1-MMP and GRASP55 could affect the activation
of the protease.
Taken together, and having been obtained using dif-
ferent experimental approaches, our data demonstrate
that the LLY
573
motif in MT1-MMP ICD plays an
important role in the interaction between MT1-MMP
with GRASP55. Interestingly, we were unable to
co-immunoprecipitate MT1-MMP and the soluble
G2A GRASP55F mutant (Fig. S3) [35,37,41], suggest-
ing that the Golgi localization of GRASP55 is crucial
for its interaction with MT1-MMP.
GRASP55 interacts with MT2-, MT3- and
MT5-MMP
The sequences of cytoplasmic domains of the four
MT-MMPs are conserved (Fig. 6A). Interestingly, the
LLY
573
motif in MT1-MMP ICD was completely
conserved in MT2-MMP (LLY
660
), whereas ILY
598

and VTY
636
sequences were found in MT3-MMP and
MT5-MMP ICDs, respectively (Fig. 6A). To test
whether MT2-, MT3- and MT5-MMP ICDs could also
interact with GRASP55, we generated VP16-MT2,
-MT3 and -MT5 chimeras (Fig. 6A). As shown in
Fig. 6B, all three ICDs (Fig. 6B, lanes 2, 5 and 8)
showed a clear interaction with GRASP55. We also
tested whether the LLY
660
motif in MT2-MMP, the
ILY
598
motif in MT3-MMP or the VTY
636
motif in
MT5-MMP could also be involved in the interaction
with GRASP55. Accordingly, VP16-MT2 LLY, VP16-
MT3 ILY and VP16-MT5 VTI triple mutants were
generated and used in the M2H assay. As previously
observed for MT1-MMP, mutation of the MT2-MMP
LLY
660
motif to AAA
660
significantly decreased
the interaction with GRASP55 (Fig. 6B, lane 3). By
Luminescence (arbitrary units)
500

0 100 200 300 400
+
MT1VP16
GAL4
VP16
+
GAL4 GRASP55
GAL4 GRASP55
+
MT1VP16
P1GAL4
+
MT1VP16
+
MT1VP16GAL4 P2
+
MT1
VP16
GAL4
1
2
3
4
5
6
+
MT1
VP16
GAL4
6

+
TGF-α
VP16GAL4
+
GAL4
GRASP55
VP16
Luminescence (arbitrary units)
500
0 100 200 300 400
1
2
+
GAL4
GRASP55
VP16
+
P1GAL4
TGF-α
VP16
2
3
TGF-α
A
B
R3
Fig. 7. GRASP55 PDZ2 and region 3 are
important for the interaction with the
MT1-MMP ICD. (A) Interactions between
VP16-MT1 and GAL4-GRASP55, GRASP55

PDZ1 (GAL4-P1), GRASP55 PDZ2 (GAL4-P2)
or GRASP55 region 3 (GAL4-R3) and
(B) between VP16-TFG a and GAL4-
GRASP55 or GAL4-P1 were tested using
the M2H system.
C. Roghi et al. Role of GRASP55 in MT1-MMP activation
FEBS Journal 277 (2010) 3158–3175 ª 2010 The Authors Journal compilation ª 2010 FEBS 3165
contrast, mutation of MT3-MMP ILY
598
(Fig. 6B,
lane 6) and MT5-MMP VTY
636
to AAA (Fig. 6B, lane
9) had no effect on GRASP55 binding.
MT1-MMP ICD binds to PDZ2 domain and
region 3 of GRASP55
GRASP55 contains two non-overlapping and structur-
ally independent PSD-95/SAP90 Drosophila septate
junction protein discs-large and epithelial tight junction
ZO-1 (PDZ) domains in its N-terminal half, followed
by a third region of approximately 250 amino acids
without known structural motif (region 3). We next
aimed to identify the region(s) of GRASP55 that inter-
acts with MT1-MMP ICD. GRASP55 PDZ1 (amino
acids 1–107), GRASP55 PDZ2 (amino acids 84–172)
and GRASP55 region 3 (amino acids 173–454) were
each fused to the GAL4 DNA binding domain and
used together with VP16-MT1 in the M2H system.
MT1-MMP ICD was found to interact with full-length
GRASP55 (Fig. 7A, lane 3), as well as with GRASP55

PDZ2 (P2; Fig. 7A, lane 5) and GRASP55 region 3
(R3; Fig. 7A, lane 6). However, no interaction was
found between VP16-MT1 and GAL4-GRASP55
PDZ1 (P1; Fig. 7A, lane 4), despite the expression of
the GAL4-GRASP55 PDZ1 chimera in HT1080 (data
not shown). TGF-a was previously reported to co-
immunoprecipitate with a very small amount of flagged
tagged GRASP55 PDZ1 domain [37]. In our hands, no
interaction between TGF-a ICD and GRASP55 PDZ1
could be observed in the M2H assay (Fig. 7B, lane 3).
The lack of interaction could result from a mis-folding
of GRASP55 PDZ1 subsequent to its fusion to the
GAL4 DNA binding domain. We therefore cannot rule
out an interaction between MT1-MMP ICD and the
GRASP55 PDZ1 domain.
GRASP55 binds to furin, PC5/6B and PC7
intracellular domains
The pro-convertase furin has previously been impli-
cated in the activation of pro-MT1-MMP [42].
Because MT1-MMP activation occurs during the
intracellular traffic of the protease, we tested whether
furin could interact, via its ICD, with GRASP55.
Accordingly, we generated a VP16-furin construct
500
0 100 200 300 400
1
*
+
FurinVP16GAL4
Luminescence (arbitrary units)

2
3
4
5
6
GAL4
+
PC7VP16
+
GAL4 PC5/6BVP16
GAL4 GRASP55 FurinVP16
+
GAL4 GRASP55 PC5/6BVP16
+
PC7VP16GAL4 GRASP55
+
Luminescence (arbitrary units)
*
*
+
GAL4 FurinVP16
+
GAL4 GRASP55 VP16
GAL4 GRASP55
+
FurinVP16
P1GAL4
+
FurinVP16
500

0 100 200 300 400
1
2
3
4
+
GAL4 P2 FurinVP16
+
R3GAL4 FurinVP16
5
6
A
B
C
Furin
GRASP55
Fig. 8. GRASP55 interacts with furin, PC5/
6B and PC7. (A) Interactions between
GAL4-GRASP55 and furin, PC5/6B or PC7
ICDs were tested using the M2H system.
(B) Interaction between VP16-furin and
GAL4-GRASP55, GRASP55 PDZ1 (GAL4-
P1), GRASP55 PDZ2 (GAL4-P2) or GRASP55
Region 3 (GAL4-R3). (C) Furin co-localized
with GRASP55. bbHT1080 cells, transfected
with a full-length furin cDNA, were
permeabilized and stained with polyclonal
antibodies against furin and GRASP55.
Arrows show examples of membrane
compartment containing GRASP55 and

furin. Scale bar = 10 lm.
Role of GRASP55 in MT1-MMP activation C. Roghi et al.
3166 FEBS Journal 277 (2010) 3158–3175 ª 2010 The Authors Journal compilation ª 2010 FEBS
where the furin ICD was fused to VP16. We also
generated the VP16-PC5/6B and VP16-PC7 chimeras
where the ICDs of PC5/6B and PC7 type I trans-
membrane pro-convertases were also fused to VP16.
Comparable to furin, PC5/6B and PC7 have previ-
ously been reported to process tetrabasic cleavage
sites [43] and their roles in pro-MT1-MMP activation
have been suggested both in vitro [44] and in vivo [42].
As shown in Fig. 8A, a clear interaction between
furin ICD and GRASP55 could be observed (Fig. 8A,
lane 2) compared to the control (Fig. 8A, lane 1). We
also observed a significant interaction between
GRASP55 and PC5/6B (Fig. 8A, lane 4) or PC7
(Fig. 8A, lane 6) ICDs. We next tested the interaction
between furin ICD and the GRASP55 domain con-
structs described previously. As shown in Fig. 8B, we
detected an interaction between furin ICD and PDZ2
(Fig. 8B, lane 5) and region 3 (Fig. 8B, lane 6) of
GRASP55. No interaction was detected between
GRASP55 PDZ1 (Fig. 8B, lane 4) and furin ICD,
A
GRASP55F
Furin
IP: MT1-MMP
IB: Furin
kDa
98

64
*
C
IB: MT1-MMP
EGFP-furin ICD
64
50
kDa
pro-MT1-MMP
active-MT1-MMP
IB: FLAG
IB: Furin
Input
lysate
IB: MT1-MMP
50
64
64
98
IB: β-actin
12
50
36
GRASP55F
MT1/MYC
EGFP-furin ICD (μg)
B
1234
50
05 10

D
IP: FLAG
IB: MYC
EGFP furin ICD (μg)
60
60
50
kDa
0.5 1.0
5
10
15
20
MMP2
n.s
*** **
***
IB: FLAG
IB: GFP
Input
lysate
60
50
60
60
0
(Arbritary units × 10
5
)
IB: MYC

12 3 4
50
–
––++
+–+
–
–
––
++ +
++ +
–+
Fig. 9. GRASP55 is important for MT1-MMP–furin complex formation and activation of pro-MT1-MMP. (A) Lysates of bbHT1080 cells trans-
fected with pCDNA3.1 Zeo+ vector control (lane 1), pCDNA3.1 Zeo+ and furin (lane 2), pCDNA3.1 Zeo+ and GRASP55 (lane 3) or with furin
and GRASP55F (lane 4) were immunoprecipitated using the affinity-purified anti-MT1-MMP IgGs and the associated furin was detected by
IB. Levels of transfected proteins were monitored in input lysates using specific antibodies. Asterisks marks endogenous furin immunopre-
cipitated by MT1-MMP in bbHT1080. (B) Expression of EGFP-furin ICD disrupted the formation of the complex between MT1-MMP and
GRASP55. Lysates of HT1080 cells transfected with pCDNA3.1 Zeo+ vector control (lane 1), MT1/MYC and GRASP55F (lane 2), MT1/MYC
and GRASP55F and 0.5 lg EGFP-furin ICD (lane 3) or MT1/MYC and GRASP55F and 1.0 lg of EGFP-furin ICD (lane 4) were immunoprecipi-
tated with the FLAG antibody and the associated MT1-MMP was detected by IB using the MYC tag antibody. Top black arrowheads indicate
IgGs. The bottom black arrowhead indicates a crossreaction. (C) IB analysis of protein extracts prepared from the EGFP-negative (lane 1)
and -positive (lane 2) cell population sorted in Fig. S4. Equal amounts of total protein (6 lg) were loaded and the expression of MT1-MMP
(pro and active) was analyzed by IB. Protein loading was controlled using the b-actin polyclonal antibody. (D) Expression of furin decrease
MT1-MMP cell surface activity. HT1080 cells were transiently transfected with empty vector pCDNA3.1 Zeo+, furin, GRASP55, fu-
rin + GRASP55, MT1-MMP, MT1-MMP + furin, MT1-MMP + GRASP55 and MT1-MMP + GRASP55 + furin. 4b-Phorbol 12-myristate 13-ace-
tate was used at 50 ngÆlL
)1
. After 24 h, supernatants were collected and analyzed by zymography. Data represent the mean ± SEM of two
independent experiments.
C. Roghi et al. Role of GRASP55 in MT1-MMP activation
FEBS Journal 277 (2010) 3158–3175 ª 2010 The Authors Journal compilation ª 2010 FEBS 3167

possibly as a result of the same reasons previously
reported for MT1-MMP and TGF-a. The interaction
observed between the furin ICD and GRASP55 lead
us to test whether both proteins could be detected in
the same membrane compartment. We were unable to
detecte the endogenous level of furin in HT1080 cells.
This is in agreement with previous observations [45].
To localize the proprotein-convertase in HT1080 cells,
we transfected the cells with a low amount of the
furin cDNA (250 ng). In furin overexpressing HT1080
cells, we could clearly detect the pro-convertase in
cytoplasmic vesicles scattered throughout the cell
cytoplasm, as well as in the TGN (Fig. 8C, furin).
Staining of GRASP55 (Fig. 8C, GRASP55) revealed
a limited co-localization between the two proteins.
Interestingly, in fixed samples, we also observed
co-localization between furin and GRASP55 in cyto-
plasmic vesicular structures (Fig. 8C, arrows), as pre-
viously observed for MT1-MMP and GRASP55 in
HT1080 cells by video microscopy (Fig. 3).
GRASP55 expression induces formation of a
MT1-MMP–furin complex
The interaction on the one hand between MT1-MMP
and GRASP55 and on the other hand between
GRASP55 and furin led us to hypothesize that
GRASP55 could potentially play a role in the forma-
tion of a complex between MT1-MMP and furin. To
test this hypothesis, bbHT1080 cells, which stably
overexpress MT1-MMP, were transfected with furin,
GRASP55F or with both furin and GRASP55F. Total

cell lysates were subjected to immunoprecipitation
using the MT1-MMP antibody and immunocomplexes
were then probed with the furin polyclonal antibody
(Fig. 9A). No furin was immunoprecipitated with
MT1-MMP in bbHT1080 cells transfected with
pCDNA 3.1 Zeo+ alone (Fig. 9A, lane 1) or express-
ing only GRASP55 (Fig. 9A, lane 3). When furin was
expressed in bbHT1080 cells, a small amount of the
proprotein-convertase was found in the same complex
as MT1-MMP (Fig. 9A, lane 2, asterisk). Expression
of GRASP55F together with furin resulted in a signifi-
cant increase in the amount of furin co-immunoprecip-
itated with MT1-MMP (Fig. 9A, lane 4), therefore
suggesting that GRASP55 could enhance the forma-
tion of a complex between MT1-MMP and furin.
Disruption of the GRASP55–furin complex
reduces processing of pro-MT1-MMP
GRASP55 appears to play an important role in the
formation of the MT1-MMP–furin complex; therefore,
disruption of the interaction between GRASP55 and
furin should perturb the activation of pro-MT1-MMP
and reduce MT1-MMP activity at the cell surface. To
test this hypothesis, we generated an enhanced green
fluorescent protein (EGFP)-furin ICD construct in
which the extracellular domain of the proprotein-con-
vertase was replaced by EGFP. Expression of the
EGFP-furin ICD construct (1 lg) in HT1080 express-
ing MT1/MYC and GRASP55F led to a disruption of
the complex between these two proteins (Fig. 9B). We
next tested whether expression of the EGFP-furin chi-

mera could exert a dominant negative effect on pro-
MT1-MMP processing by disrupting the complex
between GRASP55 and MT1-MMP. To test
this, bbHT1080 cells were transfected for 24 h with
EGFP-furin ICD. Transfected cells were then sorted by
fluorescence-activated cell sorting (FACS) into EGFP-
positive and EGFP-negative populations (Fig. S4) and
the expression of pro- and active MT1-MMP was
assessed by IB (Fig. 9C). Using this approach, we
found that high expression of the EGFP-furin ICD
construct in bbHT1080 cells (Fig. 9C, top panel, lane
2) led to an increased amount of pro-MT1-MMP being
detected compared to EGFP-negative cells (Fig. 9C,
top panel, lane 1). This observation prompted us to test
whether the reduction in pro-MT1-MMP activation
observed following furin expression would result in a
decrease of MT1-MMP activity at the cell surface.
HT1080 cells were transiently transfected with empty
vector pCDNA3.1 Zeo+ (and treated without or with
50 ngÆlL
)1
4b-phorbol 12-myristate 13-acetate), furin,
GRASP55, furin + GRASP55, MT1-MMP, MT1-
MMP + furin, MT1-MMP + GRASP55, and MT1-
MMP + GRASP55 + furin. After 24 h, the activation
of pro-MMP2 was analyzed by zymography in the cell
supernatants. As shown in Figs 9D and S5, expression
of MT1-MMP, but not furin, GRASP55 or
GRASP55 + furin, significantly increased endogenous
pro-MMP2 activation. Co-expression of furin with

MT1-MMP led to a significant reduction in the levels
of MMP2 generated compared to cells expressing only
the protease. A significant decrease was also observed
when furin was co-expressed with MT1-MMP and
GRASP55. Taken together, our observations suggest
that furin-mediated disruption of the MT1-MMP
GRASP55 complex can lead to a reduction of the
MT1-MMP, and a consequent decrease of protease
activity at the cell surface. Our data also revealed that
intracellular furin levels are critical for the efficient acti-
vation of MT1-MMP. The results obtained suggest that
GRASP55 might act as a molecular bridge between
MT1-MMP and furin and is involved in the furin-medi-
ated activation of pro-MT1-MMP.
Role of GRASP55 in MT1-MMP activation C. Roghi et al.
3168 FEBS Journal 277 (2010) 3158–3175 ª 2010 The Authors Journal compilation ª 2010 FEBS
Discussion
In the present study, we have investigated the interac-
tion between GRASP55, a Golgi matrix protein, and
MT1-MMP. Using the M2H system and a series of
MT1-MMP ICD triple mutants, we discovered that
mutation of the LLY
573
motif to AAA
573
completely
inhibited the interaction between MT1-MMP ICD and
GRASP55, suggesting that this motif is important for
the interaction of MT1-MMP ICD with the PDZ2
domain of GRASP55. The interaction of PDZ

domains with ICD internal motifs has previously been
reported. For example, the interaction of the ErbB2
ICD with the Erbin PDZ domain was found to be
dependent on a tyrosine at position )7 (the C-terminal
residue is referred to as the P
0
residue and subsequent
residues towards the N-terminal are termed P
)1
,P
)2
,
etc.) [46]. Similarly, a reduced interaction was observed
between GRASP55 and TGF-aD152 (lacking the last
eight amino acids of the ICD) compared to the TGF-a
D158 construct lacking only the last two valine resi-
dues, suggesting a role for an internal motif in the
interaction between TGF-a ICD and the GRASP55
PDZ1 domain [37]. Phosphorylation has been found to
modulate interactions between PDZ domains and their
ligands [47–49]. The affinity of the Erbin–ErbB2 inter-
action has also been found to decrease by 2.5-fold
when P
)7
tyrosine is phosphorylated [50]. In MT1-
MMP, tyrosine 573 (P
)9
) was recently reported to be
phosphorylated [31]. This opens the possibility that
binding of MT1-MMP to GRASP55 could be regu-

lated via tyrosine phosphorylation.
Surprisingly, MT1-MMP and furin ICDs were also
found to interact with the C-terminal region (region 3)
of GRASP55. Protein interactions with the region
located outside the PDZ domain have previously been
observed. Sox-4 and eIF5A [51,52] were reported to
interact only with the N-terminal region of syntenin-1,
thus allowing for the potential binding of different
proteins through the PDZ domain(s) and the N-termi-
nal region [53]. Our observation suggests that
GRASP55 could potentially interact with two MT1-
MMP molecules, therefore allowing for the homophilic
complex formation of the protease, as previously
reported [28,54]. Forcing the MT1-MMP ICD to
dimerize by fusing it to the coil-coiled region of
GM130 [39] was found to significantly enhance the
interaction between MT1-MMP and GRASP55
(Fig. S6). Furthermore, PDZ-containing proteins have
previously been reported to self-associate, generating
macromolecular complexes. This multimerization can
involve either of the PDZ domains, as for example in
glutamate receptor interacting protein 1 (GRIP1) [55]
or inactivation no afterpotential D (INAD) protein
[56]. However, it can also be a PDZ-independent
mechanism, as for example in the case of the PSD-95
protein, where dimer formation is mediated by the N-
terminal region of the protein [57]. GRASP65, which
is structurally related to GRASP55, has been found to
form dimers that can organize into higher-order oligo-
mers in interphase cells [58]. GRASP65 dimerization

involved the N-terminal GRASP domain (amino acids
1–201), which has been found to be highly conserved
between both GRASPs [35]. It is therefore possible
that, similar to GRASP65, GRASP55 dimerizes or
even oligomerizes into multimeric structures and could
therefore be involved in the oligomerization of MT1-
MMP [28,54]. MT1-MMP oligomerization, which has
been found to facilitate pro-MMP2 activation [54],
could therefore provide an explanation for the
previously obs erved intracellular activation o f pro-MMP2
[59].
Mutation of the LLY
573
motif to AAA
573
in the
MT1-MMP ICD has been reported to significantly
decrease the internalization of the protease from the
cell surface [18]. Because the expression of this mutant
at the cell surface is significantly lower than the wild-
type MT1-MMP [18], it is reasonable to suggest that
this motif also has a role in the exocytosis of protease.
We have shown that the LLY
573
motif is important for
the interaction between MT1-MMP ICD and
GRASP55 and therefore its mutation to AAA
573
could
explain the reduction of the level of cell surface expres-

sion of this mutated version of MT1-MMP. It is
important to note that perturbation of the GRASP55
interaction with TGF-a [37], as well as p24 proteins
[39], has been reported to affect the normal trafficking
of these proteins. The expression of the MT1-MMP
AAA
573
mutant at the cell surface [18] could result
from the intracellular traffic of MT1-MMP via an
alternative pathway, as previously described [60].
Alternatively, GRASP65, which is structurally related
to GRASP55, has previously been reported to interact
with transmembrane TGF- a, p24a, protein CD8a or
the frizzled receptor Fz4 [37–39] and therefore could
be involved in the intracellular traffic of MT1-MMP
AAA
573
to the cell surface. It would be interesting to
assess the role of GRASP65 with respect to MT1-
MMP trafficking to the cell surface.
MT1-MMP, similar to all the other MMPs, is syn-
thesized as a latent zymogen (pro-MT1-MMP) that is
activated by endoproteolytic cleavage of its N-terminal
inhibitory pro-domain peptide [6,61]. Furin has been
widely reported to be an activator of pro-MT1-MMP
and is considered to be physiologically relevant [42].
In the present study, we have reported that both
C. Roghi et al. Role of GRASP55 in MT1-MMP activation
FEBS Journal 277 (2010) 3158–3175 ª 2010 The Authors Journal compilation ª 2010 FEBS 3169
MT1-MMP and furin can interact with GRASP55.

Overexpression of GRASP55 has been found to
increase the amount of complex containing MT1-
MMP and furin, and the expression of a catalytically
inactive dominant negative furin construct affected the
processing of pro-MT1-MMP. Taken together, these
data reveal a new cellular function for GRASP55 as a
molecular bridge between MT1-MMP and furin.
Disruption of the MT1-MMP interaction with
GRASP55 only led to a small inhibition of pro-MT1-
MMP activation, suggesting that only a small propor-
tion of the MT1-MMP zymogen is activated by a
GRASP55-dependent furin-dependent mechanism, with
the bulk of activation being GRASP55-independent.
The interaction observed between GRASP55 and PC5/
6B or PC7 together with the potential role of these
proprotein-convertases in MT1-MMP activation
[42,44,62] would support our observation. This would
suggest that different pro-MT1-MMP activation path-
ways may co-exist within a cell.
Finally, MT2-, MT3- and MT5-MMP were also
found to interact with GRASP55. Our data led us to
hypothesize a role for GRASP55 in the furin-mediated
activation of pro-MT2-MMP, pro-MT3-MMP and
pro-MT5-MMP [63–65] because all these MT-MMPs
harbour a tetrabasic motif sandwiched between the
pro- and catalytic domains. For MT2-MMP and
MT3-MMP, the LLY
660
and ILY
598

motifs, respec-
tively, were found to be important for the interaction
of these MT-MMPs with GRASP55. By contrast,
mutation of the VTY
636
motif to AAA
636
in MT5-
MMP did not affect the binding of the protease with
GRASP55. Previous studies [66,67] have identified the
EWV
614
motif in MT5-MMP ICD as comprising an
important sequence for the interaction with the PDZ
proteins: Mint-3, AMPA binding protein and gluta-
mate receptor interacting protein (GRIP). It would
therefore be interesting to determine whether this motif
is also involved in the interaction between MT5-MMP
and GRASP55.
GRASP55 could potentially be considered as a
molecular bridge involved in connecting furin with var-
ious substrates. So far, it is unknown whether this
mechanism is specific for the type I transmembrane
MT-MMPs or whether it comprises a more general
mechanism for furin-mediated activation of transmem-
brane substrates. The activation of ADAM17 by furin
[68] and our observation of an interaction between
GRASP55 and the ICD of ADAM17 (C. Roghi and
L. Jones, unpublished results) would suggest a much
wider activation mechanism. A role of GRASP55 in

bridging MT1-MMP with other transmembrane sub-
strates could also be considered.
Materials and methods
Antibodies
FLAG mouse monoclonal antibody (IB: 8 lgÆmL
)1
) was
obtained from Sigma-Aldrich Company Ltd (Poole, UK).
GRASP55 sheep polyclonal antibody (FBA34) [immuno-
fluorescence (IF): 3 lgÆmL
)1
] was obtained from F. Barr
(University of Liverpool, UK) [35]. GRASP55 mouse
monoclonal antibody (B01P, IB: 0.48 lgÆmL
)1
) was
obtained from Abnova (Taipei, Taiwan). Anti-human
MT1-MMP sheep heavy chain IgG (N175/6) (IF:
3 lgÆmL
)1
, IB: 1.5 lgÆmL
)1
) was described previously [69].
MT1-MMP rabbit polyclonal antibody and control rabbit
IgGs were obtained from Insight Biotechnology Ltd
(Middlesex, UK). MYC (4A6) (IB:1.5 lgÆmL
)1
) and MT1-
MMP (LEM-2/15.8) (IF: 10 lgÆmL
)1

) mouse monoclonal
antibody were obtained from Millipore (UK) Ltd
(Watford, UK). Polyclonal anti-TGN46 sheep (IF:
5 lgÆmL
)1
) serum was obtained from Serotec Ltd (Oxford,
UK). b-actin (IB: 0.1 ngÆmL
)1
) and anti-furin (IF:
8.6 lgÆmL
)1
, IB: 8.6 l gÆmL
)1
) rabbit Polyclonal antibody
were from Abcam plc (Cambridge, UK). Giantin mouse
monoclonal antibody was obtained from H P. Hauri
(University of Basel, Switzerland) [70]. All secondary
antibodies were obtained from Jackson ImmunoResearch
Europe Ltd (Soham, UK) and used in accordance with
the manufacturer’s instructions.
Cell culture conditions and transfections
All cell culture reagents were obtained from Invitrogen Ltd
(Paisley, UK), unless otherwise indicated. HT1080 Human
fibrosarcoma cells (from Cancer Research UK Research
Services, London, UK) were maintained in DMEM con-
taining 10% (v/v) fetal bovine serum (Hyclone Laborato-
ries Inc., UT, USA), 2 mml-glutamine, 100 UÆmL
)1
penicillin and 100 lgÆmL
)1

streptomycin at 37 °Cin5%
CO
2
. HT1080 stably expressing wild-type MT1-MMP
(bbHT1080 clone #2) were obtained from J. Clements
(British Biotech plc, Oxford, UK) [71]. Transient transfec-
tions were performed using FuGENE 6 (Roche Diagnostics
Ltd, Lewes, UK). Total DNA transfected was kept
constant.
DNA constructs
Full-length wild-type human MT1-MMP cDNA was cloned
in pCDNA3.1 Zeo+ vector (Invitrogen Ltd). VP16-MT1
was generated by cloning MT1-MMP intracellular domain
(ICD; amino acids 562–582) in frame with the activation
domain of herpes simplex virus type 1 (VP16) in the pACT
Checkmate Mammalian Two-Hybrid System vector (Pro-
mega, Southampton, UK). VP16-MT2, VP16-MT3 and
Role of GRASP55 in MT1-MMP activation C. Roghi et al.
3170 FEBS Journal 277 (2010) 3158–3175 ª 2010 The Authors Journal compilation ª 2010 FEBS
VP16-MT5 were generated using MT2-MMP (amino acids
647–669), MT3-MMP (amino acids 585–607) and MT5-
MMP (amino acids 623–645) ICDs. MT1-MMP (VP16-
MT1 Y, VP16-MT1 LL, VP16-MT1 FRR, VP16-MT1
HGT, VP16-MT1 PRR, VP16-MT1 LLY, VP16-MT1
CQR, VP16-MT1 SLL, VP16-MT1 DKV), MT2-MMP
(VP16-MT2 LLY), MT3-MMP (VP16-MT3 ILY) and
MT5-MMP (VP16-MT5 VTY) mutants were generated by
PCR using mutagenic primers. All these constructs have a
Pro-Gly-Gly-Gly linker (PGGG) between VP16 and the
ICD. TGF-a ICD (amino acids 126–160) was generated by

annealing overlapping oligonucleotides and ‘filling up’ the
single strand regions with KOD DNA polymerase
(2.5 units; Merck Biosciences Ltd, Nottingham, UK). The
resulting double-stranded DNA was cloned in frame with
VP16. MT1/MYC has a MYC tag between P
312
and T
313
in the hinge domain of MT1-MMP. LLY570-573AAA
MT1/MYC and G2A GRASP55F mutants were generated
using the QuikChange II site-directed mutagenesis kit
(Stratagene, Amsterdam, the Netherlands). Full-length rat
GRASP55 cDNA and N-terminally EGFP-tagged rat
GRASP55 (GRASP55-GFP) were obtained from F. Barr
and colleagues [35]. GRASP55F in pCDNA3.1 Zeo+ con-
tained a C-terminal FLAG tag. Full-length GRASP55,
GRASP55 PDZ1 (amino acids 1–107), GRASP55 PDZ2
(amino acids 84–172) and GRASP55 region 3 (amino acids
173–454) were fused to GAL4 DNA binding domain in the
pBIND Checkmate Mammalian Two-Hybrid System vector
(Promega).
To generate the MT1/EYFP, oligonucleotides coding for
MT1-MMP signal sequence (amino acids 1–30) were
annealed and inserted upstream of EYFP in pEYFP-N1
(Clontech-Takara Bio Europe, Saint-Germain-en-Laye,
France). DNA coding for MT1-MMP hinge, hemopexin,
stalk, transmembrane and cytoplasmic domains (amino
acids 283–582) was amplified by PCR and cloned down-
stream of EYFP. VP16-PC5/6B, VP16-PC7 and VP16-furin
were generated by fusing VP16 to ICDs of mouse PC5/6B

(amino acids 1461–1548), human PC7 (amino acids 684–
785) or furin (amino acids 738–794). To create EGFP-furin
ICD, the EGFP (pEGFP-N1; Clontech-Takara Bio Europe)
stop codon was replaced by PCR with the Pro-Gly-Gly
linker. The TIMP-3 signal peptide (amino acids 1–23, from
D. Edwards, UEA, Norwich, UK) was then inserted
upstream of EGFP-PGG. The furin stalk, transmembrane
and intracellular domains (amino acids 708–794) were
amplified by PCR and inserted downstream of EGFP-
PGG. This construct lacks the intramolecular cleavage site
for furin shedding [72]. All constructs were confirmed by
sequencing.
SDS/PAGE and IB
All reagents were purchased from Bio-Rad (Hemel Hemp-
stead, UK). SDS/PAGE (10%) and IB were carried out as
previously described [69]. Membranes were then stripped
using the ReblotÔ Plus kit [Millipore (UK) Ltd].
M2H system
HT1080 cells (10
5
) were seeded per well of a six-well plates
for 24 h prior to transfection (1 lg of pG5luc DNA, 1 lg
of DNA for the construct in pBIND and 1 lg of DNA for
the construct in pACT). After 16 h at 37 °C, cells were
washed in ice-cold NaCl/P
i
(137 mm NaCl, 4.3 mm
Na
2
HPO

4
, 2.7 mm KCl, 1.47 mm KH
2
PO
4
) and lysed in
passive lysis buffer (200 lL; Promega) for 15 min at room
temperature. After centrifugation (13 000 g for 5 min at
room temperature), Firefly and Renilla reniformis luciferase
activities were measured in 20 lL of lysate using the
Dual-Luciferase Reporter Assay System (Promega) and a
SpectraFluor Plus plate reader (Tecan UK, Reading, UK).
Firefly luciferase activity was normalized to Renilla lucifer-
ase activity and is presented as relative luminescence units.
Each transfection was performed in duplicate and two inde-
pendent measurements were read per sample. All numerical
values are shown as the mean ± SEM. The graphs pre-
sented are representative of at least three experiments.
Immunoprecipitation
HT1080 cells (2 · 10
5
per well in a six-well plate) were
transfected (1–2 lg per construct) for 16 h. Cells were then
washed twice in ice-cold NaCl/P
i
and lysed for 5 min in
600 lL of lysis buffer per well [10 mm Tris–HCl, pH 7.4,
150 mm NaCl, 1% (v/v) Triton X-100, 0.5% (v/v) Nonidet
P-40, 1 mm EDTA, 1 mm EGTA, 1 mm sodium vanadate]
containing protease inhibitor (Roche Diagnostics Ltd). For

immunoprecipitation of endogenous GRASP55, cells were
lysed in 1% Brij96, 5 mm CaCl
2
,5mm MgCl
2
,10mm
Tris–HCl, pH 7.4, 150 mm NaCl containing protease inhib-
itor (Roche Diagnostics Ltd). After centrifugation (13 000 g
for 15 min at 4 °C), the supernatant volume was adjusted
to 800 lL and 150 lL (input) was kept for IB analysis.
Antibodies or control IgG (3 lg) were bound to Dynabeads
protein G (Invitrogen) for 16 h at 4 °C in NaCl/P
i
contain-
ing 5 mgÆmL
)1
BSA. Beads were washed with lysis buffer
and mixed with the protein extracts for 2 h at 4 °C under
constant rotation. Beads were washed three times for
15 min with lysis buffer and resuspended in 2 · Laemmli
sample buffer (30 lL). Denatured samples were then
resolved by SDS/PAGE.
Indirect immunofluorescence microscopy
Cells (1 · 10
5
) seeded on 13 mm round glass coverslips
(Agar Scientific, Stansted, UK) were transfected for 16 h.
Cells were then processed for immunofluorescence micro-
scopy, as described previously [69].
C. Roghi et al. Role of GRASP55 in MT1-MMP activation

FEBS Journal 277 (2010) 3158–3175 ª 2010 The Authors Journal compilation ª 2010 FEBS 3171
Live cell time-lapse microscopy
Transfected HT1080 cells (500 ng of each construct) were
imaged with a confocal Zeiss LSM510 META microscope
(Zeiss Axiovert 200M platform, · 63/1.4 NA Oil Plan-Apo-
chromat lens; Carl Zeiss Ltd, Welwyn Garden City, UK).
The Argon Ion laser (Lasos Lasertechnik GmbH, Jena,
Germany) output was restricted to 3% of maximum
(< 0.3 mW at focal plane) to minimize phototoxicity and
bleaching. Coverslips (PeCon GmbH, Erbach, Germany)
were placed in a POC-R imaging chamber (PeCon GmbH)
containing phenol-red free Hepes-buffered medium and
kept at 37 ° C using a Heating Insert-P (PeCon GmbH).
Time series images (optical section thickness of 1.7 lm)
were collected in a lambda acquisition mode at one frame
every 4 s for 10–20 min. Reference EGFP and EYFP emis-
sion spectra were captured from singly transfected cells.
EGFP/EYFP signal unmixing was carried out using Zeiss
lsm510 software, version 3.2SP2. Confocal images were
post-processed with volocity software (Improvision Sys-
tems, Coventry, UK). Movies are presented at ten frames
per second, representing a speed increase of ·40.
FACS
bbHT1080, transfected for 24 h with the EGFP-furin ICD
construct, were washed with NaCl/P
i
and detached from the
plastic using NaCl/P
i
containing 5 mm EDTA (5 min at

37 °C). After centrifugation (500 g for 5 min), cells were
washed in NaCl/P
i
, resuspended DMEM containing 0.5%
(v/v) fetal bovine serum and kept on ice. EGFP-negative cells
were sorted from the EGFP-positive cells using an Aria
SORP flow cytometer (Becton Dickenson, San Jose, CA,
USA). Sorting efficiency was controlled by a post-sort analy-
sis. Cells were lysed with lysis buffer. Protein concentration
was determined using a BCA protein assay kit (Perbio Sci-
ence UK Ltd, Cramlington, UK). The western blot presented
is a representative of three independent experiments.
Pro-MMP2 activation and zymography
HT1080 cells (1 · 10
5
cells) in DMEM and 10% fetal bovine
serum were transfected in suspension using 0.5 lg of each
DNA construct and 1.5 lL of FuGENEÔ 6 transfection
reagent (Roche Diagnostics Ltd) in accordance with the
manufacturer’s instructions in 24-well culture dishes. The
total amount of DNA per transfection was kept constant.
After 24 h at 37 °C, the medium was removed and replaced
with 300 lL per well of serum-free DMEM containing insu-
lin, transferrin and selenium supplements (Sigma-Aldrich)
and the cells were incubated at 37 °C for a further 24 h. Cell
supernatants were harvested and analyzed by gelatin zymog-
raphy, as described previously [73]. Quantification of MMP2
was performed using imagequant tl software, version 7.0
(GE Heathcare, Little Chalfont, UK).
Statistical analysis

Statistical analysis was performed using the graphpad
prism, version 5 (GraphPad Software, Inc., San Diego, CA,
USA). Statistical significance was calculated using Student’s
t-test. Statistical significance was defined as P < 0.05 (*),
P < 0.001 (**) and P < 0.0001 (***).
Acknowledgements
We would like to thank Drs F. Barr, J. Creemers,
K. Shennan, D. Edwards and H P. Hauri for providing
reagents used in the present study. We thank Greg
Veltri, Therese Martin and Michele Bones (CRI Flow
Cytometry core unit) and Jane Gray (CRI equipment
park) for their technical assistance. We thank Neil
Taylor, Sue Atkinson, Patricia Eisenach and Helen
Gillingham for their helpful discussions and for criti-
cally reading the manuscript. We would like to
acknowledge the support of Cancer Research UK and
Hutchison Whampoa Limited (C.R. and G.M.; CRUK
grant C100/A8243), the Medical Research Council
(L.J.; grant 71393), the Wellcome Trust (M.G.; grant
079895/Z/06/Z), the British Heart Foundation
(W.R.E.; intermediate fellowship FS/03/055/15910)
and the University of Cambridge.
References
1 Zucker S, Pei D, Cao J & Lopez-Otin C (2003)
Membrane type-matrix metalloproteinases (MT-MMP).
Curr Top Dev Biol 54, 1–74.
2 Noe
¨
l A, Gilles C, Bajou K, Devy L, Kebers F, Lewalle
JM, Maquoi E, Munaut C, Remacle A & Foidart JM

(1997) Emerging roles for proteinases in cancer.
Invasion Metastasis 17, 221–239.
3 Ellerbroek SM & Stack MS (1999) Membrane
associated matrix metalloproteinases in metastasis.
Bioessays 21, 940–949.
4 Sato H, Takino T & Miyamori H (2005) Roles of
membrane-type matrix metalloproteinase-1 in tumor
invasion and metastasis. Cancer Sci 96, 212–217.
5 Itoh Y & Seiki M (2006) MT1-MMP: a potent modifier
of pericellular microenvironment. J Cell Physiol 206,
1–8.
6 Will H, Atkinson SJ, Butler GS, Smith B & Murphy G
(1996) The soluble catalytic domain of membrane type
1 matrix metalloproteinase cleaves the propeptide of
progelatinase A and initiates autoproteolytic activation.
Regulation by TIMP-2 and TIMP-3. J Biol Chem 271,
17119–17123.
7 Knauper V, Will H, Lopez-Otin C, Smith B, Atkinson
SJ, Stanton H, Hembry RM & Murphy G (1996) Cellu-
lar mechanisms for human procollagenase-3 (MMP-13)
Role of GRASP55 in MT1-MMP activation C. Roghi et al.
3172 FEBS Journal 277 (2010) 3158–3175 ª 2010 The Authors Journal compilation ª 2010 FEBS
activation. Evidence that MT1-MMP (MMP-14) and
gelatinase A (MMP-2) are able to generate active
enzyme. J Biol Chem 271, 17124–17131.
8 Holmbeck K, Bianco P, Yamada S & Birkedal-Hansen
H (2004) MT1-MMP: a tethered collagenase. J Cell
Physiol 200, 11–19.
9 Seiki M (2003) Membrane-type 1 matrix metalloprotein-
ase: a key enzyme for tumor invasion. Cancer Lett 194,

1–11.
10 Vihinen P & Kahari VM (2002) Matrix metalloprotein-
ases in cancer: prognostic markers and therapeutic
targets. Int J Cancer 99, 157–166.
11 Genis L, Galvez BG, Gonzalo P & Arroyo AG (2006)
MT1-MMP: universal or particular player in angiogene-
sis? Cancer Metastasis Rev 25, 77–86.
12 Sabeh F, Ota I, Holmbeck K, Birkedal-Hansen H,
Soloway P, Balbin M, Lopez-Otin C, Shapiro S, Inada
M, Krane S et al. (2004) Tumor cell traffic through the
extracellular matrix is controlled by the membrane-
anchored collagenase MT1-MMP. J Cell Biol 167,
769–781.
13 Hotary KB, Allen ED, Brooks PC, Datta NS, Long
MW & Weiss SJ (2003) Membrane type I matrix
metalloproteinase usurps tumor growth control imposed
by the three-dimensional extracellular matrix. Cell 114,
33–45.
14 Ueda J, Kajita M, Suenaga N, Fujii K & Seiki M
(2003) Sequence-specific silencing of MT1-MMP
expression suppresses tumor cell migration and
invasion: importance of MT1-MMP as a therapeutic
target for invasive tumors. Oncogene 22, 8716–8722.
15 Gingras D & Beliveau R (2009) Emerging concepts in
the regulation of membrane-type 1 matrix metallopro-
teinase activity. Biochim Biophys Acta 1803, 142–150.
16 Gingras D, Bousquet-Gagnon N, Langlois S, Lacham-
bre MP, Annabi B & Beliveau R (2001) Activation of
the extracellular signal-regulated protein kinase (ERK)
cascade by membrane-type-1 matrix metalloproteinase

(MT1-MMP). FEBS Lett 507, 231–236.
17 Lehti K, Valtanen H, Wickstrom SA, Lohi J &
Keski-Oja J (2000) Regulation of membrane-type-1
matrix metalloproteinase activity by its cytoplasmic
domain. J Biol Chem 275, 15006–15013.
18 Uekita T, Itoh Y, Yana I, Ohno H & Seiki M (2001)
Cytoplasmic tail-dependent internalization of
membrane-type 1 matrix metalloproteinase is important
for its invasion-promoting activity. J Cell Biol 155,
1345–1356.
19 Urena JM, Merlos-Suarez A, Baselga J & Arribas J
(1999) The cytoplasmic carboxy-terminal amino acid
determines the subcellular localization of proTGF-
(alpha) and membrane type matrix metalloprotease
(MT1-MMP). J Cell Sci 112, 773–784.
20 Jiang A, Lehti K, Wang X, Weiss SJ, Keski-Oja J &
Pei D (2001) Regulation of membrane-type matrix
metalloproteinase 1 activity by dynamin-mediated
endocytosis. Proc Natl Acad Sci USA 98, 13693–13698.
21 Rozanov DV, Ghebrehiwet B, Ratnikov B, Monosov
EZ, Deryugina EI & Strongin AY (2002) The
cytoplasmic tail peptide sequence of membrane type-1
matrix metalloproteinase (MT1-MMP) directly binds to
gC1qR, a compartment-specific chaperone-like
regulatory protein. FEBS Lett 527, 51–57.
22 Remacle A, Murphy G & Roghi C (2003) Membrane
type I-matrix metalloproteinase (MT1-MMP) is interna-
lised by two different pathways and is recycled to the
cell surface. J Cell Sci 116, 3905–3916.
23 Wang X, Ma D, Keski-Oja J & Pei D (2004) Co-recy-

cling of MT1-MMP and MT3-MMP through the trans-
Golgi network. Identification of DKV582 as a recycling
signal. J Biol Chem 279
, 9331–9336.
24 Nakahara H, Howard L, Thompson EW, Sato H, Seiki
M, Yeh Y & Chen WT (1997) Transmembrane/
cytoplasmic domain-mediated membrane type 1-matrix
metalloprotease docking to invadopodia is required for
cell invasion. Proc Natl Acad Sci USA 94, 7959–7964.
25 Sounni NE, Devy L, Hajitou A, Frankenne F, Munaut
C, Gilles C, Deroanne C, Thompson EW, Foidart JM
& Noel A (2002) MT1-MMP expression promotes
tumor growth and angiogenesis through an up-regula-
tion of vascular endothelial growth factor expression.
FASEB J 16, 555–564.
26 D’Alessio S, Ferrari G, Cinnante K, Scheerer W,
Galloway AC, Roses DF, Rozanov DV, Remacle AG,
Oh ES, Shiryaev SA et al. (2008) Tissue inhibitor of
metalloproteinases-2 binding to membrane-type 1
matrix metalloproteinase induces MAPK activation and
cell growth by a non-proteolytic mechanism. J Biol
Chem 283, 87–99.
27 Rozanov DV, Savinov AY, Williams R, Liu K, Golub-
kov VS, Krajewski S & Strongin AY (2008) Molecular
signature of MT1-MMP: transactivation of the
downstream universal gene network in cancer. Cancer
Res 68, 4086–4096.
28 Lehti K, Lohi J, Juntunen MM, Pei D & Keski-Oja J
(2002) Oligomerization through hemopexin and
cytoplasmic domains regulates the activity and turnover

of membrane-type 1 matrix metalloproteinase. J Biol
Chem 277, 8440–8448.
29 Ludwig T, Theissen SM, Morton MJ & Caplan MJ
(2008) The cytoplasmic tail dileucine motif LL572
determines the glycosylation pattern of membrane-type
1 matrix metalloproteinase. J Biol Chem 283, 35410–
35418.
30 Anilkumar N, Uekita T, Couchman JR, Nagase H,
Seiki M & Itoh Y (2005) Palmitoylation at Cys574 is
essential for MT1-MMP to promote cell migration.
FASEB J 19, 1326–1328.
31 Nyalendo C, Michaud M, Beaulieu E, Roghi C,
Murphy G, Gingras D & Beliveau R (2007)
C. Roghi et al. Role of GRASP55 in MT1-MMP activation
FEBS Journal 277 (2010) 3158–3175 ª 2010 The Authors Journal compilation ª 2010 FEBS 3173
Src-dependent phosphorylation of membrane type I
matrix metalloproteinase on cytoplasmic tyrosine 573:
role in endothelial and tumour cell migration. J Biol
Chem 282, 15690–15699.
32 Moss NMP, Wu YIP, Liu YM, Munshi HGM & Stack
MSP (2009) Modulation of the membrane type 1 matrix
metalloproteinase cytoplasmic tail enhances tumor cell
invasion and proliferation in three-dimensional collagen
matrices. J Biol Chem 284, 19791–19799.
33 Uekita T, Gotoh I, Kinoshita T, Itoh Y, Sato H,
Shiomi T, Okada Y & Seiki M (2004) Membrane-type 1
matrix metalloproteinase cytoplasmic tail-binding
protein-1 is a new member of the Cupin superfamily.
A possible multifunctional protein acting as an invasion
suppressor down-regulated in tumors. J Biol Chem 279,

12734–12743.
34 Labrecque L, Nyalendo C, Langlois S, Durocher Y,
Roghi C, Murphy G, Gingras D & Beliveau R (2004)
Src-mediated tyrosine phosphorylation of caveolin-1
induces its association with membrane type 1 matrix
metalloproteinase. J Biol Chem 279, 52132–52140.
35 Shorter J, Watson R, Giannakou ME, Clarke M,
Warren G & Barr FA (1999) GRASP55, a second
mammalian GRASP protein involved in the stacking of
Golgi cisternae in a cell-free system. EMBO J 18, 4949–
4960.
36 Short B, Preisinger C, Korner R, Kopajtich R,
Byron O & Barr FA (2001) A GRASP55-rab2 effector
complex linking Golgi structure to membrane traffic.
J Cell Biol 155, 877–883.
37 Kuo A, Zhong C, Lane WS & Derynck R (2000)
Transmembrane transforming growth factor-alpha
tethers to the PDZ domain-containing, Golgi mem-
brane-associated protein p59/GRASP55. EMBO J 19,
6427–6439.
38 D’Angelo G, Prencipe L, Iodice L, Beznoussenko G,
Savarese M, Marra P, Di Tullio G, Martire G, De
Matteis MA & Bonatti S (2009) GRASP65 and
GRASP55 sequentially promote the transport of
C-terminal valine-bearing cargos to and through the
Golgi complex. J Biol Chem 284, 34849–34860.
39 Barr FA, Preisinger C, Kopajtich R & Korner R (2001)
Golgi matrix proteins interact with p24 cargo receptors
and aid their efficient retention in the Golgi apparatus.
J Cell Biol 155, 885–891.

40 Jesch SA, Lewis TS, Ahn NG & Linstedt AD (2001)
Mitotic phosphorylation of Golgi reassembly stacking
protein 55 by mitogen-activated protein kinase ERK2.
Mol Biol Cell 12, 1811–1817.
41 Barr FA, Puype M, Vandekerckhove J & Warren G
(1997) GRASP65, a protein involved in the stacking of
Golgi cisternae. Cell 91, 253–262.
42 Yana I & Weiss SJ (2000) Regulation of membrane
type-1 matrix metalloproteinase activation by propro-
tein convertases. Mol Biol Cell 11, 2387–2401.
43 Nakayama K (1997) Furin: a mammalian subtilisin/
Kex2p-like endoprotease involved in processing of a
wide variety of precursor proteins. Biochem J 327, 625–
635.
44 Remacle AG, Rozanov DV, Fugere M, Day R &
Strongin AY (2006) Furin regulates the intracellular
activation and the uptake rate of cell surface-associated
MT1-MMP. Oncogene 25, 5648–5655.
45 Wouters S, Leruth M, Decroly E, Vandenbranden M,
Creemers JW, van de Loo JW, Ruysschaert JM &
Courtoy PJ (1998) Furin and proprotein convertase 7
(PC7)/lymphoma PC endogenously expressed in rat
liver can be resolved into distinct post-Golgi compart-
ments. Biochem J 336, 311–316.
46 Borg JP, Marchetto S, Le Bivic A, Ollendorff V, Jaulin-
Bastard F, Saito H, Fournier E, Adelaide J, Margolis B
& Birnbaum D (2000) ERBIN: a basolateral PDZ pro-
tein that interacts with the mammalian ERBB2/HER2
receptor. Nat Cell Biol 2, 407–414.
47 Cohen NA, Brenman JE, Snyder SH & Bredt DS

(1996) Binding of the inward rectifier K
+
channel Kir
2.3 to PSD-95 is regulated by protein kinase A
phosphorylation. Neuron 17, 759–767.
48 Cao TT, Deacon HW, Reczek D, Bretscher A & von
Zastrow M (1999) A kinase-regulated PDZ-domain
interaction controls endocytic sorting of the beta2-
adrenergic receptor. Nature 401, 286–290.
49 Hegedus T, Sessler T, Scott R, Thelin W, Bakos E,
Varadi A, Szabo K, Homolya L, Milgram SL &
Sarkadi B (2003) C-terminal phosphorylation of MRP2
modulates its interaction with PDZ proteins. Biochem
Biophys Res Commun 302, 454–461.
50 Birrane G, Chung J & Ladias JA (2003) Novel mode of
ligand recognition by the Erbin PDZ domain. J Biol
Chem 278, 1399–1402.
51 Geijsen N, Uings IJ, Pals C, Armstrong J, McKinnon
M, Raaijmakers JA, Lammers JW, Koenderman L &
Coffer PJ (2001) Cytokine-specific transcriptional regu-
lation through an IL-5Ralpha interacting protein.
Science 293, 1136–1138.
52 Li AL, Li HY, Jin BF, Ye QN, Zhou T, Yu XD,
Pan X, Man JH, He K, Yu M et al. (2004) A novel
eIF5A complex functions as a regulator of p53 and
p53-dependent apoptosis. J Biol Chem 279, 49251–
49258.
53 Gimferrer I, Ibanez A, Farnos M, Sarrias MR, Fenutria
R, Rosello S, Zimmermann P, David G, Vives J, Serra-
Pages C et al. (2005) The lymphocyte receptor CD6

interacts with syntenin-1, a scaffolding protein contain-
ing PDZ domains. J Immunol 175, 1406–1414.
54 Itoh Y, Takamura A, Ito N, Maru Y, Sato H, Suenaga
N, Aoki T & Seiki M (2001) Homophilic complex for-
mation of MT1-MMP facilitates proMMP-2 activation
on the cell surface and promotes tumor cell invasion.
EMBO J 20, 4782–4793.
Role of GRASP55 in MT1-MMP activation C. Roghi et al.
3174 FEBS Journal 277 (2010) 3158–3175 ª 2010 The Authors Journal compilation ª 2010 FEBS
55 Srivastava S, Osten P, Vilim FS, Khatri L, Inman G,
States B, Daly C, DeSouza S, Abagyan R, Valtschanoff
JG et al. (1998) Novel anchorage of GluR2/3 to the
postsynaptic density by the AMPA receptor-binding
protein ABP. Neuron 21, 581–591.
56 Xu XZ, Choudhury A, Li X & Montell C (1998) Coor-
dination of an array of signaling proteins through
homo- and heteromeric interactions between PDZ
domains and target proteins. J Cell Biol 142, 545–555.
57 Hsueh YP, Kim E & Sheng M (1997) Disulfide-linked
head-to-head multimerization in the mechanism of ion
channel clustering by PSD-95. Neuron 18, 803–814.
58 Wang Y, Seemann J, Pypaert M, Shorter J & Warren G
(2003) A direct role for GRASP65 as a mitotically regu-
lated Golgi stacking factor. EMBO J 22, 3279–3290.
59 Lee AY, Akers KT, Collier M, Li L, Eisen AZ &
Seltzer JL (1997) Intracellular activation of gelatinase A
(72-kDa type IV collagenase) by normal fibroblasts.
Proc Natl Acad Sci USA 94, 4424–4429.
60 Deryugina EI, Ratnikov BI, Yu Q, Baciu PC, Rozanov
DV & Strongin AY (2004) Prointegrin maturation follows

rapid trafficking and processing of MT1-MMP in furin-
negative colon carcinoma LoVo cells. Traffic 5, 627–641.
61 Pei D & Weiss SJ (1996) Transmembrane-deletion
mutants of the membrane-type matrix metalloprotein-
ase-1 process progelatinase A and express intrinsic
matrix-degrading activity. J Biol Chem 271, 9135–9140.
62 Stawowy P, Meyborg H, Stibenz D, Borges Pereira
Stawowy N, Roser M, Thanabalasingam U, Veinot JP,
Chretien M, Seidah NG, Fleck E et al. (2005) Furin-like
proprotein convertases are central regulators of the
membrane type matrix metalloproteinase-pro-matrix
metalloproteinase-2 proteolytic cascade in atherosclero-
sis. Circulation 111, 2820–2827.
63 Kang T, Nagase H & Pei D (2002) Activation of mem-
brane-type matrix metalloproteinase 3 zymogen by the
proprotein convertase furin in the trans-Golgi network.
Cancer Res 62, 675–681.
64 Wang X & Pei D (2001) Shedding of membrane type
matrix metalloproteinase 5 by a furin-type convertase:
a potential mechanism for down-regulation. J Biol
Chem 276, 35953–35960.
65 Will H & Hinzmann B (1995) cDNA sequence and
mRNA tissue distribution of a novel human matrix
metalloproteinase with a potential transmembrane
segment. Eur J Biochem 231, 602–608.
66 Wang P, Wang X & Pei D (2004) Mint-3 regulates the
retrieval of the internalized membrane-type matrix
metalloproteinase, MT5-MMP, to the plasma
membrane by binding to its carboxyl end motif EWV.
J Biol Chem 279, 20461–20470.

67 Monea S, Jordan BA, Srivastava S, DeSouza S & Ziff
EB (2006) Membrane localization of membrane type 5
matrix metalloproteinase by AMPA receptor binding
protein and cleavage of cadherins. J Neurosci 26,
2300–2312.
68 Srour N, Lebel A, McMahon S, Fournier I, Fugere M,
Day R & Dubois CM (2003) TACE/ADAM-17
maturation and activation of sheddase activity require
proprotein convertase activity. FEBS Lett 554, 275–283.
69 Atkinson SJ, Roghi C & Murphy G (2006) MT1-MMP
hemopexin domain exchange with MT4-MMP blocks
enzyme maturation and trafficking to the plasma
membrane in MCF7 cells. Biochem J 398, 15–22.
70 Schweizer A, Fransen JA, Bachi T, Ginsel L & Hauri
HP (1988) Identification, by a monoclonal antibody, of
a 53-kD protein associated with a tubulo-vesicular
compartment at the cis-side of the Golgi apparatus.
J Cell Biol 107, 1643–1653.
71 Green DR, Banuls MP, Gearing AJH, Needham LA,
White MRH & Clements JM (1994) Generation of
human umbilical vein endothelial cell lines which
maintain their differentiated phenotype. Endothelium 2,
191–201.
72 Plaimauer B, Mohr G, Wernhart W, Himmelspach M,
Dorner F & Schlokat U (2001) ‘Shed’ furin: mapping
of the cleavage determinants and identification of its
C-terminus. Biochem J 354
, 689–695.
73 Heussen C & Dowdle EB (1980) Electrophoretic analy-
sis of plasminogen activators in polyacrylamide gels

containing sodium dodecyl sulfate and copolymerized
substrates. Anal Biochem 102, 196–202.
Supporting information
The following supplementary material is available:
Fig. S1. MT1-MMP co-immunoprecipitates with
GRASP55F in HeLa cells.
Fig. S2. The LLY motif in the MT1-MMP ICD is
crucial for the interaction with GRASP55.
Fig. S3. MT1-MMP co-immunoprecipitates with
GRASP55F in HeLa cells.
Fig. S4. Selection of bbHT1080 cells expressing EGFP-
furin by FACS.
Fig. S5. Furin expression impairs activation of
pro-MMP2.
Fig. S6. Fusion of GM130 coiled-coil region to
MT1-MMP ICD increased its binding to GRASP55.
This supplementary material can be found in the
online version of this article.
Please note: As a service to our authors and readers,
this journal provides supporting information supplied
by the authors. Such materials are peer-reviewed and
may be re-organized for online delivery, but are not
copy-edited or typeset. Technical support issues arising
from supporting information (other than missing files)
should be addressed to the authors.
C. Roghi et al. Role of GRASP55 in MT1-MMP activation
FEBS Journal 277 (2010) 3158–3175 ª 2010 The Authors Journal compilation ª 2010 FEBS 3175