Tài liệu Báo cáo khoa học: Degradation of tropoelastin by matrix metalloproteinases – cleavage site specificities and release of matrikines pptx

Degradation of tropoelastin by matrix metalloproteinases –
cleavage site specificities and release of matrikines
Andrea Heinz
1
, Michael C. Jung
1
, Laurent Duca
2
, Wolfgang Sippl
1
, Samuel Taddese
1
,
Christian Ihling
1
, Anthony Rusciani
2
,Gu
¨
nther Jahreis
3
, Anthony S. Weiss
4
, Reinhard H. H. Neubert
1
and Christian E. H. Schmelzer
1
1 Institute of Pharmacy, Martin Luther University Halle-Wittenberg, Halle (Saale), Germany
2 Faculte
´
des Sciences, Laboratoire de Biochimie, Reims, France

3 Max Planck Research Unit for Enzymology of Protein Folding, Halle (Saale), Germany
4 School of Molecular and Microbial Biosciences, University of Sydney, Australia
Keywords
gelatinase B; GxxPG; macrophage elastase;
matrilysin; mass spectrometry
Correspondence
Christian E. H. Schmelzer, Martin Luther
University Halle-Wittenberg, Institute of
Pharmacy, Wolfgang-Langenbeck-Str. 4,
06120 Halle (Saale), Germany
Fax: +49 345 5527292
Tel: +49 345 5525215
E-mail:
(Received 7 January 2010, revised 3
February 2010, accepted 12 February
2010)
doi:10.1111/j.1742-4658.2010.07616.x
To provide a basis for the development of approaches to treat elastin-
degrading diseases, the aim of this study was to investigate the degradation
of the natural substrate tropoelastin by the elastinolytic matrix metallopro-
teinases MMP-7, MMP-9, and MMP-12 and to compare the cleavage site
specificities of the enzymes using complementary MS techniques and molec-
ular modeling. Furthermore, the ability of the three proteases to release
bioactive peptides was studied. Tropoelastin was readily degraded by all
three MMPs. Eighty-nine cleavage sites in tropoelastin were identified for
MMP-12, whereas MMP-7 and MMP-9 were found to cleave at only
58 and 63 sites, respectively. Cleavages occurred predominantly in the
N-terminal and C-terminal regions of tropoelastin. With respect to the
cleavage site specificities, the study revealed that all three MMPs similarly
tolerate hydrophobic and ⁄ or aliphatic amino acids, including Pro, Gly, Ile,

and Val, at P
1
¢. MMP-7 shows a strong preference for Leu at P
1
¢, which is
also well accepted by MMP-9 and MMP-12. Of all three MMPs, MMP-12
best tolerates bulky charged and aromatic amino acids at P
1
¢. All three
MMPs showed a clear preference for Pro at P
3
that could be structurally
explained by molecular modeling. Analysis of the generated peptides
revealed that all three MMPs show a similar ability to release bioactive
sequences, with MMP-12 producing the highest number of these peptides.
Furthermore, the generated peptides YTTGKLPYGYGPGG,
YGARPGVGVGGIP, and PGFGAVPGA, containing GxxPG motifs that
have not yet been proven to be bioactive, were identified as new matrikines
upon biological activity testing.
Structured digital abstract
l
MINT-7709630: MMP-7 (uniprotkb:P09237) cleaves (MI:0194) Tropoelastin (uniprotkb:
P15502)byprotease assay (MI:0435)
l
MINT-7709668: MMP-9 (uniprotkb:P14780) cleaves (MI:0194) Tropoelastin (uniprotkb:
P15502)byprotease assay ( MI:0435)
l
MINT-7709289: MMP-12 (uniprotkb:P39900) cleaves (MI:0194) Tropoelastin (uniprotkb:
P15502)byprotease assay ( MI:0435)
Abbreviations

ACN, acetonitrile; EBP, elastin-binding protein; ECM, extracellular matrix; EDP, elastin-derived peptide; i.d., internal diameter; MMP, matrix
metalloproteinase; qTOF, quadrupole time-of-flight; TFA, trifluoroacetic acid.
FEBS Journal 277 (2010) 1939–1956 ª 2010 The Authors Journal compilation ª 2010 FEBS 1939
Introduction
Matrix metalloproteinases (MMPs) form a large family
of multidomain zinc-dependent and calcium-dependent
endopeptidases that are known to cleave various com-
ponents of the extracellular matrix (ECM). MMPs
play a central role in connective tissue remodeling pro-
cesses and regulation of cell matrix composition
through their effects on cell migration, cell differentia-
tion, cell growth, wound healing, inflammation, angio-
genesis, and apoptosis. The disruption of the
physiological balance between MMP activation and
deactivation is connected with severe diseases such as
atherosclerosis, arthritis, pulmonary emphysema, myo-
cardial infarction, and tumor growth and metastasis
[1–5].
Three of the most widely studied MMPs with
respect to their biological actions are MMP-7
(EC 3.4.24.23), MMP-9 (EC 3.4.24.35), and MMP-12
(EC 3.4.24.65). MMP-7, which is also referred to as
matrilysin 1, is mainly expressed by epithelial cells and
processes various ECM constituents, such as collagens,
gelatin, and laminin, and also non-ECM proteins,
including pro-tumor necrosis factor-a and a
2
-macro-
globulin [2]. MMP-9 (gelatinase B) is secreted by neu-
trophils and macrophages and has also been found in

various malignant cells, ras-transformed murine cells,
and chemically stimulated fibroblasts. It has, for
instance, been shown to cleave native type IV and VII
collagens, gelatin, laminin, and plasminogen [2].
MMP-12 (macrophage elastase) is expressed mainly by
macrophages. The enzyme cleaves a variety of sub-
strates, including collagens, gelatin, laminin, pro-tumor
necrosis factor-a, and plasminogen [2]. Natural sub-
strates that are known to be degraded by MMP-7,
MMP-9, MMP-12, and a further member of the
MMP family, MMP-2, are the connective tissue pro-
tein elastin and its monomeric precursor tropoelastin
[6–11].
The biopolymer elastin, which provides elasticity
and resilience to several tissues, including lungs, arter-
ies, and skin, shows a unique chemical composition
characterized by the presence of large amounts of the
four hydrophobic amino acids Gly, Val, Ala, and Pro.
The protein consists of molecules of its soluble precur-
sor tropoelastin that are cross-linked at Lys residues.
Owing to its hydrophobicity and extensive cross-link-
ing, elastin is insoluble and highly resistant to proteo-
lytic degradation. Moreover, elastin does not undergo
substantial turnover in healthy tissue [12–16].
In the last two decades, studies have revealed that
elastin is not only a structural protein influencing the
architecture and biomechanical properties of the ECM
but also plays an active role in various physiological
processes [16]. Some elastin-derived peptides (EDPs),
which may occur upon proteolytic degradation of elas-

tin and tropoelastin, promote angiogenesis [17] and are
associated with the regulation of various cell activities,
including cell adhesion, chemotaxis, migration, prolif-
eration, protease activation, and apoptosis [18–21].
Such EDPs are matrikines; this name generally denotes
bioactive ligands that exist as part of an ECM protein.
The results of different studies suggest that EDPs con-
taining the GxxPG motif, in particular, are biologically
active, as these are able to interact with the elastin-
binding protein (EBP) [18,19,22–25].
In light of the diverse and complex biological func-
tions of elastin, EDPs, and MMPs, it is clear that the
previously mentioned aberrant expression of elastin-
degrading enzymes such as MMPs often leads to dam-
age to elastic fibers. This damage, together with other
biological processes triggered by EDPs and MMPs,
may support the development and progression of vari-
ous pathological conditions. It has, for instance, been
found that aortic stenosis is associated with increased
activity of MMP-2 and MMP-9 [26], atherosclerosis is
influenced by MMP-12, which promotes athero-
sclerotic plaque instability [27], and the development
of aortic aneurysms is enhanced by MMP-2, MMP-9,
and MMP-12 [28,29]. Studies have also indicated that
MMP-9 is involved in processes such as cardiac rup-
ture after myocardial infarction [30] and photoaging of
the skin [11,31]. Furthermore, overexpression of
MMP-12 has been found to be associated with the
development and progression of pulmonary emphy-
sema [32], photoaging of the skin [33], and granuloma-

tous skin diseases [34]. MMP-7 is strongly expressed in
tumors of almost every organ in the body and seems
to play a vital role in tumor progression and angiogen-
esis [35,36]. Taken together, these examples show that
it is of utmost importance to understand and charac-
terize elastin-degrading processes, including the cleav-
age behavior of elastinolytic MMPs and the nature of
the peptides released on degradation. This approach
may aid in the development of directed therapies to
treat pathologies related to elastin degradation, the
overexpression of MMPs, and the consequent release
of bioactive peptides.
Few studies have investigated the enzymatic degra-
dation of elastin or its precursor tropoelastin [10,37,38]
and the release of bioactive peptides upon enzymatic
degradation of elastin, tropoelastin or synthesized
domains derived from tropoelastin [39–41]. The aim of
the present study was to obtain detailed information
Tropoelastin degradation by elastinolytic MMPs A. Heinz et al.
1940 FEBS Journal 277 (2010) 1939–1956 ª 2010 The Authors Journal compilation ª 2010 FEBS
on the cleavage site specificities of MMP-7, MMP-9,
and MMP-12 in tropoelastin using complementary MS
techniques and to characterize and compare the cleav-
age behavior of the three enzymes using molecular
modeling. Tropoelastin was chosen as substrate in this
study because of its biological relevance during elastin
turnover and matrix remodeling and to increase the
number of identifiable peptides resulting from proteo-
lytic digestion. Mature elastin, in contrast, shows only
limited suitability for characterization of the cleavage

site specificity of proteolytic enzymes, owing to its
extensive cross-linking which restricts MS fragmenta-
tion and sequencing approaches [16,37]. In contrast to
previous studies, the present work, for the first time,
sought to obtain a comprehensive insight into the pre-
ferred amino acids at the cleavage site positions P
1
–P
4
and P
1
¢–P
4
¢ of the substrate and to give a structural
explanation of the amino acid preferences of MMP-7,
MMP-9, and MMP-12, which have not been described
to date. Moreover, the potential of the three MMPs to
produce bioactive peptides upon proteolytic digestion
was investigated, and peptides containing the GxxPG
motif resulting from the digestion of tropoelastin by
MMP-7, MMP-9, and MMP-12 were tested for their
bioactivity.
Results
Highest number of cleavages and highest
sequence coverage obtained for MMP-12
Sequence coverages of 71.1, 59.5, and 80.7% were
determined for degradation by MMP-7, MMP-9, and
MMP-12, respectively (Fig. 1). The cleavage sites iden-
tified for all three MMPs occurred mainly in amino or
carboxyl regions of the tropoelastin sequence, in agree-

ment with previous studies on bovine and human elas-
tin [10,37]. Altogether, for MMP-12, 89 cleavage sites
and 132 peptides were identified in almost all domains
of tropoelastin with the exception of domains 8, 9, and
11. In contrast, for MMP-7 and MMP-9, only 58 (84
peptides) and 63 (74 peptides) cleavage sites could be
determined, respectively. For MMP-7, no cleavages
were observed in domains 8–11, 17, 19–21, 23, and 36.
MMP-9 showed a similar cleavage behavior and did
not degrade domains 8–11, 16–20, and 36. Altogether,
23 cleavage sites and 20 peptides were found that were
common for all three MMPs. It is worth mentioning
that in MALDI-TOF experiments several unidentified
higher-mass peptides of between 10 kDa and 20 kDa
were observed for the three MMPs, underlining the
finding that some domains resisted proteolysis (data
not shown).
Aliphatic and
⁄
or hydrophobic residues favored at
P
1
¢
The P
1
¢–S
1
¢ interaction has been identified as the main
determinant of the cleavage position of MMPs in pep-
tide substrates [5,10,42,43]. The results of this work

are in agreement with previous studies proposing that
the three enzymes can accept a variety of amino acids
with hydrophobic and ⁄ or aliphatic residues including
Ala, Gly, Val, Leu, Ile, Tyr, and Phe at P
1
¢
[2,10,43,44]. Only the charged amino acid Lys, which
was found to be tolerated at P
1
¢ by MMP-12, has been
reported to be an exception to the previously men-
tioned preferences [43]. The present study appears to
confirm that MMP-12 shows a preference for x-Lys, as
the enzyme cleaved at 11 of 35 (31%) of such cleavage
sites (Fig. 2A; Table 1). It was also found that MMP-7
and MMP-9 cleaved N-terminal to Lys; however, this
was to a lesser extent than MMP-12, with MMP-9
cleaving 7 of 35 (20%), and MMP-7 cleaving only 1 of
the 35 (3%) possible cleavage sites. The interaction of
the hexapeptide substrate PQGKAG containing Lys at
P
1
¢ with the active sites of MMP-9 and MMP-12 as
investigated by molecular modeling is shown in Fig. 3
and confirms that both enzymes are able to accept Lys
at P
1
¢.
Another difference in the cleavage behavior of the
three MMPs was found at possible cleavage sites with

Leu. Figure 2A and Table 1 show that MMP-7 has a
strong preference for Leu at P
1
¢, which has also been
described previously [44,45]. MMP-7 cut at 29 of
40 (73%) possible cleavage sites with Leu, whereas
MMP-9 and MMP-12 only cleaved at 14 (35%) and
19 (48%) sites, respectively. The preference of MMP-7
is also shown in Fig. 3, where the interaction of the
hexapeptide substrate PQGLAG containing Leu at P
1
¢
is modeled. Small differences in the cleavage site speci-
ficities of the three MMPs were observed at cleavage
sites with bulky aromatic amino acids such as Tyr and
Phe at P
1
¢, which were cut, in particular, by MMP-12.
While MMP-12 hydrolyzed 53% (8 of 15) of the possi-
ble x-Tyr peptide bonds, MMP-7 and MMP-9 showed
similar cleavage behavior and only cut at 3 and 4 of
15 possible x-Tyr cleavage sites, respectively. x-Phe
bonds were found to be hydrolyzed by MMP-7 (2 of
16) and MMP-12 (4 of 16) but not by MMP-9.
The cleavage behavior of MMP-7, MMP-9, and
MMP-12 at x-Pro, x-Gly, x-Ile, and x-Val peptide
bonds is very similar. All three MMPs similarly toler-
ate amino acids with relatively small aliphatic and ⁄ or
hydrophobic residues at P
1

¢. Another interesting differ-
ence, however, can be found at x-Ala peptide bonds,
which are almost resistant to hydrolysis by MMP-7
A. Heinz et al. Tropoelastin degradation by elastinolytic MMPs
FEBS Journal 277 (2010) 1939–1956 ª 2010 The Authors Journal compilation ª 2010 FEBS 1941
Fig. 1. Cleavage sites identified after digestion of human tropoelastin isoform 2 (SwissProt accession number: P15502-2) with MMP-7,
MMP-9, and MMP-12. Cleavage sites are indicated by triangles (MMP-7, red; MMP-9, green; MMP-12, black), and all regions covered by
peptides are labeled with solid lines (MMP-7, red; MMP-9, green; MMP-12, black). Bioactive sequences [17,19,46–65] are shown in blue.
The sequence of the octapeptide 226–233 used to model the interaction between a natural substrate and the active site of MMP-12 (Fig. 4)
is marked with a blue bar. The sequences of three peptides containing GxxPG motifs (YTTGKLPYGYGPGG, YGARPGVGVGGIP, and
PGFGAVPGA) used for bioactivity tests are labeled with asterisks and orange arrows.
Tropoelastin degradation by elastinolytic MMPs A. Heinz et al.
1942 FEBS Journal 277 (2010) 1939–1956 ª 2010 The Authors Journal compilation ª 2010 FEBS
(2 of 157 possible x-Ala bonds) but are cleaved to a
small extent by MMP-9 (12 of 157) and even more by
MMP-12 (24 of 157).
The determined cleavage site preference of MMP-7
was based on the number of cleavages occurring N-ter-
minal to the respective amino acid and follows
A
B
Fig. 2. Number of amino acids found in P
1
¢
(A) and P
1
(B) in peptides of tropoelastin
after digestion with either MMP-7, MMP-9
or MMP-12.
Table 1. Occurrence of different amino acids at the substrate positions P

1
-P
4
and P
1
¢-P
4
¢ after digestion with MMP-7, MMP-9, and MMP-12.
Values are based on the total amounts of each of the amino acids in tropoelastin (isoform 2).
A. Heinz et al. Tropoelastin degradation by elastinolytic MMPs
FEBS Journal 277 (2010) 1939–1956 ª 2010 The Authors Journal compilation ª 2010 FEBS 1943
the order Leu (48% of all identified cleavage
sites) >> Val ⁄ Gly (each 12%) > Pro (10%) > Tyr
(5%), with Leu being clearly preferred over other
aliphatic and ⁄ or hydrophobic amino acids (Table 2).
The cleavage site specificity of MMP-9 follows the
order Leu (22% of all identified cleavage sites) > Ala
(19%) > Gly (14%) > Lys (11%) > Val (9%).
MMP-12, which is the most active of the three
enzymes, shows a cleavage site specificity according to
the order Ala (26% of all identified cleavage site-
s) > Leu (20%) > Lys (12%) > Val ⁄ Tyr (each
9%) > Gly (7%).
With respect to the charged or polar amino acids
Arg, Gln, Ser, Asp, Cys, and Glu, which together con-
stitute only 3.6% of the tropoelastin sequence, this
study revealed that hardly any cleavage occurred
N-terminal to these amino acids upon digestion with
MMP-7, MMP-9, and MMP-12. An exception is a sin-
gle cleavage between Ala and Cys, which was found

for MMP-12.
Mainly Gly and Ala are found at P
1
With regard to the preferred amino acids at P
1
, the
present study revealed strong similarities between the
three MMPs. After hydrolysis by MMP-7, MMP-9,
and MMP-12, Ala and Gly, which constitute 52%
of the tropoelastin sequence, were predominant at P
1
(Fig. 2B), which is in accordance with known P
1
specificities for MMPs [10,44]. In detail, the experi-
ments with MMP-7 showed that Gly occurred N-ter-
minal to 33 (55%) of the 60 identified cleavage sites
and Ala occurred N-terminal to 9 (15%) (Table 2;
Fig. 2B). After digestion with MMP-9, Gly was
found N-terminal to 24 (38%) of the 64 identified
cleavage sites, and Ala was found N-terminal to 21
(33%). Similar results were obtained for MMP-12,
where Gly was found N-terminal of 30 (33%) and
Ala N-terminal of 36 (39%) of the 92 identified
cleavage sites, respectively. In summary, it can be
stated that Gly and Ala occur at P
1
in about 70%
of the cleavages, whereas in the other 30% small
amino acids such as Pro and Val are mainly found
at P

1
(Table 2).
A
B
C
Fig. 3. Interaction of hexapeptide substrates with the binding
sites of MMP-7, MMP-9, and MMP-12. For clarity, only non-
conserved residues of the S
1
¢ pocket within the three
studied MMPs are shown. The zinc ion at the catalytic site is
shown as a yellow ball. (A) Interaction of the peptide substrate
PQGLAG containing a P
1
¢ Leu with the MMP-7 binding site. (B)
Interaction of the peptide substrate PQGKAG containing a P
1
¢
Lys with the MMP-9 binding site. (C) Interaction of the peptide
substrate PQGKAG containing a P
1
¢ Lys with the MMP-12
binding site.
Tropoelastin degradation by elastinolytic MMPs A. Heinz et al.
1944 FEBS Journal 277 (2010) 1939–1956 ª 2010 The Authors Journal compilation ª 2010 FEBS
Mainly Gly and Ala are found at P
2
–P
4
and P

2
¢–P
4
¢
Table 2 shows that following digestion with MMP-7,
MMP-9 or MMP-12, predominantly Gly and Ala were
found at P
2
–P
4
and P
2
¢–P
4
¢. Interestingly, a preference
for Pro at P
3
was found for all three MMPs. Pro
occurred at P
3
at 18% (MMP-7) to 24% (MMP-12) of
all identified cleavage sites. The preference for Pro at P
3
was also confirmed by molecular modeling, in which the
interaction of the natural substrate LPYGYGPG
containing Pro at P
3
with the active site of MMP-12 was
investigated (Fig. 4). Furthermore, it was observed that
MMP-7 tolerates Val at P

3
¢ better than MMP-9 and
MMP-12.
Peptides with bioactive sequences released upon
proteolytic digestion of tropoelastin with MMP-7,
MMP-9 or MMP-12
Table 3 and Fig. 1 show that some of the 42 bioactive
sequences [17,19,46–65] partly overlap in tropoelastin.
Altogether, 35 of these sequences were found: 15
sequences, of which 11 were nonrepetitive, were found
in 27 peptides of different lengths released by MMP-7;
22 sequences, of which 11 were nonrepetitive,
were found in 23 peptides released by MMP-9; and 20
sequences, of which 13 were nonrepetitive, were
found in 41 peptides released by MMP-12 (Tables 3
and 4).
Table 2. Occurrence of different amino acids at the substrate positions P
1
-P
4
and P
1
¢ -P
4
¢ after digestion with MMP-7, MMP-9, and
MMP-12. Values are based on the number of cleavage sites identified on MS analysis of the digests.
P3‘
P4‘
A B
Fig. 4. Interaction of the natural substrate LPYGYGPG (residues 226–233 from tropoelastin isoform 2; see Fig. 1) and the MMP-12 active

site. (A) The molecular surface of the binding pocket is colored according to electrostatic potential (red indicates negative electrostatic poten-
tial; blue indicates positive electrostatic potential). (B) The hydrogen bonds between the backbone residues of the substrate and the residues
of the MMP-12 binding site are highlighted (yellow dashed lines).
A. Heinz et al. Tropoelastin degradation by elastinolytic MMPs
FEBS Journal 277 (2010) 1939–1956 ª 2010 The Authors Journal compilation ª 2010 FEBS 1945
Bioactive sequences were predominantly released
from the N-terminal and C-terminal parts of tropoela-
stin, where most cleavages occurred (Fig. 1). From the
central part of the tropoelastin molecule, only three
peptides containing the matrikines VPGVG (341–345),
VGVPG (344–348), and GARPG (384–388) were
released by MMP-12 exclusively (Tables 3 and 4). The
smallest peptides found with bioactive sequences
Table 3. Bioactive sequences within tropoelastin isoform 2 and those that were identified as parts of peptides of different lengths after
digestion with MMP-7, MMP-9, and MMP-12. The bioactive sequences were selected on the basis of several publications [17,19,46–65].
Moreover, all sequences containing the GxxPG motif found in tropoelastin isoform 2 were listed, except for the peptide GGVPG, which is
known to show no bioactivity [19,47].
Tropoelastin degradation by elastinolytic MMPs A. Heinz et al.
1946 FEBS Journal 277 (2010) 1939–1956 ª 2010 The Authors Journal compilation ª 2010 FEBS
Table 4. Peptides containing bioactive sequences that were identified after digestion of recombinant tropoelastin with MMP-7, MMP-9, or
MMP-12. Bioactive sequences are in bold red letters.
A. Heinz et al. Tropoelastin degradation by elastinolytic MMPs
FEBS Journal 277 (2010) 1939–1956 ª 2010 The Authors Journal compilation ª 2010 FEBS 1947
included eight peptides of lengths between 9 and 18
amino acids (Table 4). The shortest peptide was identi-
fied after digestion with MMP-9 and MMP-12, and
displayed the sequence PGFGAVPGA (578–586). In
addition to the 8 relatively short peptides, 53 longer
peptides of lengths up to 105 amino acids were identi-
fied. The longest peptide (103–207) was released by

MMP-12 and contained 9 partly overlapping bioactive
sequences (Tables 3 and 4).
It is worth mentioning that domain 24, which can be
considered as one huge matrikine encompassing 15 par-
tially overlapping bioactive peptides, remained intact
upon proteolytic digestion with MMP-7, MMP-9, and
MMP-12 (Fig. 1). An interesting finding is a peptide
that was released after digestion with MMP-9 and
spans the sequence 442–523 and so contains all the 15
bioactive sequences within domain 24 (Table 4). This
includes GLVPG and repeats of VGVAPG and
GVAPGV, which have been reported to show biologi-
cal effects [46–60], as well as further GxxPG sequences
(GIGPG and GLAPG). Because none of the three
MMPs was capable of cleaving within domain 24, it is
also likely that treatment with MMP-7 and MMP-12
resulted in additional peptides comprising the whole of
domain 24; however, these could not be identified upon
MS analysis.
Among the different peptides released by MMP-7,
MMP-9, and MMP-12, three contain GxxPG motifs
for which no biological activity has yet been described:
YTTGKLPYGYGPGG (residues 221-234, released by
MMP-7, MMP-9, and MMP-12), YGARPGVGVG-
GIP (residues 383–395, released by MMP-12), and
Table 4. (continued)
Fig. 5. Zymography analysis of pro-MMP-2
secretion. Cells were stimulated for 24 h
with or without elastin peptides, and cell
culture media were subjected to gelatin

zymography. Lower panel: densitometric
analysis. The statistical test compares
control and elastin peptides. **P < 0.01.
Differences observed in biological activities
between the three peptides are not
significant.
Tropoelastin degradation by elastinolytic MMPs A. Heinz et al.
1948 FEBS Journal 277 (2010) 1939–1956 ª 2010 The Authors Journal compilation ª 2010 FEBS
PGFGAVPGA (residues 578–586, released by MMP-9
and MMP-12), containing GYGPG, GARPG, and
GAVPG, respectively (Fig. 1; Tables 3 and 4). In this
work, their ability to trigger pro-MMP-2 secretion by
human dermal fibroblasts in culture was evaluated.
The zymography results show that all three peptides
exhibit strong biological activity triggering a significant
and important increase in pro-MMP-2 secretion,
demonstrating that they are matrikines (Fig. 5).
Discussion
While most cleavages by the three MMPs occurred
predominantly in N-terminal or C-terminal regions of
tropoelastin, only a few cleavages were found in
domains 17–24, which constitute a region that is
known to be involved in aggregation through coacer-
vation and subsequently the cross-linking of tropoela-
stin during elastogenesis [66,67]. The lack of cleavage
in this area is consistent with the folded nature of this
part of the molecule [68,69] and its observed associa-
tion hot-spot properties [70], as the incubation condi-
tions used were conducive to tropoelastin association
[71]. Furthermore, the tertiary structure of tropoelastin

has been reported to contain pockets of hydrophobic
clusters [72] that may be only partially solvent-accessi-
ble and hence restrict cutting in this area.
Cleavage site specificity
To gain a better understanding of the cleavage site
specificities of the three MMPs, the active sites were
modeled in complex with peptide substrates (Figs 3
and 4). Previous studies of the crystal structures of dif-
ferent MMPs have indicated that, in general, S
1
¢ pock-
ets are relatively large, although they differ in size and
shape among various MMPs depending on the amino
acids constituting the S
1
¢ loop which varies in second-
ary structure and length [2,5,73–76]. In the case of
MMP-7, Tyr214 has been found to extend into the
opening of the S
1
¢ pocket, restricting the size of the
binding site (Fig. 3A) [5]. Thus, the S
1
¢ pocket of
MMP-7 is less accessible to large polar or aromatic
residues such as Ser, Lys, Arg, Phe, and Tyr, whereas
it easily accommodates medium-sized hydrophobic res-
idues, including Leu and the even bulkier Ile
[5,44,45,76,77], as confirmed in this study (Tables 1
and 2; Fig. 2A). Both MMP-9 and MMP-12 contain

Leu at the same position which is a smaller molecule
than Tyr214 in MMP-7 (Fig. 3). The S
1
¢ pocket there-
fore has an extended shape that is comparable to a
long tube running through the MMP molecule [5] and
is able to accommodate larger bulky residues; this is
found in particular for MMP-12 (Tables 1 and 2;
Fig. 2A).
The comparison of the available crystal structures of
MMP-7, MMP-9, and MMP-12 (Fig. 3) in complex
with modeled substrates revealed further differences in
the MMP binding cleft. The polar residue Thr214 of
MMP-12 is mutated to an aliphatic residue in MMP-9
(Val398) and MMP-7 (Ala215), and Glu219 of MMP-7
and MMP-12 is mutated to Gln402 in MMP-9. The
model of the bound substrate with Lys at P
1
¢ (Fig. 3C)
indicates the neighborhood of the polar residues
Thr214 and Glu219 and the e-amino group of the P
1
¢
Lys, whereas the environment in MMP-7 is more
hydrophobic, thus favoring apolar residues at P
1
¢
(Tables 1 and 2; Fig. 2A). In the MMP-9 crystal struc-
ture, the entrance of the S
1

¢ pocket is restricted by the
bulky and flexible Tyr393, whereas in MMP-12 the
smaller Thr210 is observed at the same position, indi-
cating that MMP-12 can accept larger residues.
Overall, the MS findings (Tables 1 and 2; Fig. 2A)
and the modeled structures (Fig. 3) show that all three
MMPs can accommodate medium-sized aliphatic
and ⁄ or hydrophobic amino acids at P
1
¢, including Ala,
Gly, Pro, Val, Leu, and Ile, which is in agreement with
previous studies on the cleavage site specificities of
MMPs [10,37,38,43]. The differences in the cleavage
behavior of the three MMPs at charged or bulky aro-
matic amino acids such as Lys, Tyr, and Phe (Tables 1
and 2; Fig. 2A) result from differences in the second-
ary structures and amino acid compositions of the
active sites of the three MMPs (Fig. 3). As described
above, MMP-12 has a more polar and larger active
site than MMP-7 and MMP-9 and hence tolerates such
residues better. The low number or lack of cleavages
by all three MMPs N-terminal to large and polar
amino acids, including Arg, Gln, Ser, Asp, Cys, and
Glu, is most likely a result of unfavorable interactions
and may also be due to the fact that these amino acids
only exist in small numbers in tropoelastin and consti-
tute only 3.6% of the tropoelastin sequence.
A difference between the present and three previous
studies on elastin [10,37,38] concerns the amino acid
Lys. While Lys was identified as a relatively well

accepted amino acid at P
1
¢ for MMP-12 in this and
another study [43], the studies on elastin have not
identified Lys at P
1
¢. The reasons for this lie in signifi-
cant structural differences between tropoelastin and
native elastin. Elastin is highly cross-linked at Lys resi-
dues, with about 88% of all Lys residues in tropoela-
stin participating in covalent linkages that are crucial
for the formation of mature elastin [78]. As elastin
cross-links are not enzymatically degraded and the
structure of Lys is significantly altered upon cross-link-
A. Heinz et al. Tropoelastin degradation by elastinolytic MMPs
FEBS Journal 277 (2010) 1939–1956 ª 2010 The Authors Journal compilation ª 2010 FEBS 1949
ing, Lys is not commonly found in high amounts at
P
1
¢ upon analysis of the cleavage site specificity of
MMPs in elastin.
The results of this study show that predominantly
Gly and Ala were found at P
1
–P
4
and P
2
¢–P
4

¢, which is
consistent with previous studies [10,38,44]. It has been
reported that Ala and Gly are found at P
1
for 75% of
the time when Leu or Ile is in P
1
¢ [10]. Furthermore, it
was found that the amino acids experimentally identi-
fied at P
1
¢ are, in most cases, preceded by Gly or Ala
[10]. Thus, the current and previous studies indicate
that the occurrence of a specific amino acid at P
1
might correlate with the total number of molecules of
the respective amino acid in tropoelastin and may also
be influenced by the amino acid at P
1
¢. As Ala and
Gly constitute more than 50% of the tropoelastin
sequence, it seems likely that the nature of the amino
acid at P
1
is governed not only by the amino acid pref-
erences of the elastinolytic enzyme at this position, but
also by the tropoelastin sequence itself. Furthermore,
the connection between P
1
¢ and P

1
can be shown, for
example, for MMP-12, which exhibits a higher prefer-
ence for Lys in P
1
than the other MMPs. This might
actually be due to the fact that the enzyme tolerates
Tyr in P
1
¢ better than the other MMPs, and almost all
cleavages with Lys in P
1
¢ occurred at Lys-Tyr bonds
(Table 1; Fig. 1).
An interesting finding is the preference of the three
MMPs for Pro at P
3
(Table 2), which has also previ-
ously been reported for MMP-3 (stromelysin), MMP-7,
MMP-9, and MMP-13 [44,45,79], but has not yet
been described for MMP-12. To understand this pref-
erence, the available crystal structure of MMP-12 was
graphically analyzed in complex with the natural sub-
strate LPYGYGPG (Fig. 4A). It was found that the
backbone of the bound peptide is forced into a turn
conformation by a Pro at P
3
which favors binding to
the active site. Furthermore, this turn conformation
allows the interaction of bulky residues with the S

2
pocket, as shown for Tyr (Fig. 4B, C) at P
2
, and can
also be seen from the experimental data (Tables 1
and 2).
Release of peptides containing bioactive
sequences
All three MMPs released peptides containing the
motifs GLGVGAGVP, PGAIPG, VPGVG, and
VGVPG, which have been reported to act as matrikin-
es by stimulating cell proliferation and showing
chemotactic activity on different cell types (Table 4)
[47,55,59,61–63]. Moreover, the bioactive sequence
VVPQ [64] was found in peptides after digestion with
MMP-7 and MMP-12. MMP-9 released a peptide con-
taining VGVAPG [46–60], GVAPGV [59] and their
repeats, which exhibit a range of biological activities,
in addition to a peptide containing bioactive GLVPG
[19,46]. Considering the fact that biological effects of
the above mentioned sequences have already been
reported, it is likely that the peptides containing these
sequences display bioactivity, provided that they are
liberated in vivo.
It has been suggested that the ability of EDPs to
exhibit biological activities is associated with their
structural conformation. In particular, GxxP sequences
in which x „ Gly are able to adopt a type VIII b-turn
conformation that enables the peptides to bind to the
EBP and thus show biological activities [19,61]. Most

of the various GxxPG motifs released in different
peptides by MMP-7, MMP-9, and MMP-12 (Table 3)
follow that rule. It can hence be assumed that these
motifs occur in the conformation required for binding
to EBP and that the released peptides display bioactiv-
ity [19]. The sequence GPQPG, which is released in
longer peptides by MMP-7 and MMP-12, seems to be
a good candidate for interaction with EBP, as a Pro
following a Gly (GPxPG) has been reported to stabi-
lize the type VIII b-turn owing to its unique hetero-
cyclic pyrrolidine ring structure [19]. This, however,
needs to be proven in further experiments.
To confirm the hypothesis that peptides containing
GxxPG motifs are likely to act as matrikines, bioactiv-
ity tests were performed on the three peptides,
YTTGKLPYGYGPGG, YGARPGVGVGGIP, and
PGFGAVPGA, that were released during digestion of
tropoelastin by MMP-7, MMP-9, and ⁄ or MMP-12
(Fig. 1; Tables 3 and 4). The three peptides contain the
GxxPG motifs GYGPG, GARPG, and GAVPG,
which have not yet been tested for their bioactivity. It
has previously been shown that HT-1080 tumor cells
and human dermal fibroblasts exhibit an increase in
pro-MMP-2 secretion when they are treated with elas-
tin peptides such as VGVAPG [65]. Consequently, an
increase in pro-MMP-2 secretion can be considered to
be an indicator for bioactivity of elastin peptides. When
dermal fibroblasts were incubated in the presence or
absence of three peptides, it was found that all three
peptides exhibited strong biological activity triggering a

significant increase in pro-MMP-2 secretion (Fig. 5). It
is important to note that all three peptides increased
pro-MMP-2 expression to the same extent, as no signif-
icant difference in bioactivity was observed between the
peptides. Overall, these results show that YTTGKL-
PYGYGPGG, YGARPGVGVGGIP and PGFGAV-
PGA can be considered as new matrikines that are
likely to be released by the action of elastinolytic
Tropoelastin degradation by elastinolytic MMPs A. Heinz et al.
1950 FEBS Journal 277 (2010) 1939–1956 ª 2010 The Authors Journal compilation ª 2010 FEBS
MMPs. The results of the bioactivity study confirm the
assumption that a variety of amino acids (in this case
Tyr, Ala, Gly, Arg, or Val) may occur in the x position
of active GxxPG motifs [19]. Interestingly, the peptide
YTTGKLPYGYGPGG, containing GYGPG, pos-
sesses a Gly in the x position and is nevertheless
biologically active. This could be explained by the
fact that the GYGPG motif is preceded by eight
amino acids that could facilitate the adoption of an
active conformation and its stabilization. This has
also been suggested by Moroy et al. [19] who showed
that a residue located before the GxxP motif increases
the stability of the bioactive type VIII b-turn confor-
mation.
With regard to the long peptide spanning more than
the entire length of domain 24 that was released by
MMP-9 and contains a variety of bioactive sequences,
it is likely that it has biological effects. As has previ-
ously been stated, the presence of multiple parts of a
peptide in the type VIII b-turn conformation increases

the probability of an interaction with EBP and hence
could support biological activity [65]. The same
applies to other peptides containing more than one
bioactive sequence, such as peptide 103–207 released
by MMP-12, comprising nine bioactive sequences, or
peptide 121–207, released by MMP-7 and MMP-12,
containing eight bioactive sequences (Tables 3 and 4).
All of these are able to form multiple type VIII b-turns
and could thus be more likely to interact with EBP
[65].
In summary, while all three MMPs show similar
ability to release peptides with bioactive sequences
owing to their similar cleavage behavior, MMP-12 pro-
duced the overall highest number of peptides and
therefore the greatest diversity of bioactive sequences
(Table 4). The liberation of a higher number of bioac-
tive sequences in a given time may increase the proba-
bility of an interaction with EBP [65]. The study also
revealed that three of the peptides containing the
GxxPG motif generated during digestion of tropoela-
stin by MMP-7, MMP-9, and MMP-12 show biologi-
cal activity and can therefore be considered to be new
matrikines.
Experimental procedures
Materials
The primary transcript of tropoelastin undergoes substan-
tial alternative splicing, resulting in the production of multi-
ple isoforms in nature [80]. In this work, tropoelastin
isoform 2 (also known as SHELD26A) prepared as previ-
ously described was used as substrate [81,82]. Its sequence

lacks domains encoded by the frequently spliced exons 22,
24A, and 26A (isoform 2) [83]. Recombinant MMP-7 and
MMP-12 expressed in Escherichia coli and full-length
human MMP-9 purified from human neutrophils were
obtained from Biomol (Plymouth Meeting, PA, USA). The
catalytic domains of MMP-7 and MMP-12 were used
because these are the biologically relevant forms of the pro-
teases. MMP-7, as the shortest member of the MMP
family, does not contain a hemopexin-like domain and acts
only with its catalytic domain, and MMP-12 is special in
that it autocatalytically loses its hemopexin-like domain
soon after activation without loss of its elastin-degrading
capacity [84]. Gelatin was purchased from Sigma (Saint-
Quentin Fallavier, France). All reagents for cell culture
were obtained from Gibco BRL (Invitrogen, Cergy
Pontoise, France). HPLC-grade acetonitrile (ACN) (VWR
Prolabo, Leuven, Belgium) and water purified by means of
a Millipore (Milford, MA, USA) filtration device (Milli-Q
grade) were used. Analytical-grade Tris, formic acid,
trifluoroacetic acid (TFA), sodium chloride, and calcium
chloride dihydrate were purchased from Merck (Darmstadt,
Germany).
Synthesis of peptides
The peptides YTTGKLPYGYGPGG, YGARPGVGVG-
GIP, and PGFGAVPGA were produced by solid-phase pep-
tide synthesis using 0.15 mm of preloaded Fmoc amino acid
Wang resins (NovaBiochem, La
¨
ufelfingen, Switzerland) on
a SYRO II multiple peptide synthesizer (MultiSynTech,

Witten, Germany). Assembly of the peptides was performed
with Fmoc chemistry using a standard protocol with Fmoc
amino acids as building blocks, and PyBOP (NovaBio-
chem) and N-methyl-morpholine as coupling reagents in
dimethylformamide. Piperidine (20%) ⁄ dimethylformamide
was the standard cleavage cocktail used for Fmoc detach-
ment. The resins were treated twice for 10 min. All cou-
plings were performed using a four-fold excess of the Fmoc
amino acid derivative, PyBOP, and N-methyl-morpholine
in dimethylformamide; a double coupling protocol was
used. After detachment of the peptides from the resins and
side chain deprotection with TFA ⁄ tri-isopropylsilan ⁄ water
(95 : 3: 2, v ⁄ v ⁄ v) for 2 h at room temperature, the crude
peptides were precipitated with diethylether and purified by
preparative RP-HPLC on a Gilson 306 system with an
SP 250 ⁄ 10 Nucleosil 100-7 C8 column (Macherey-Nagel,
Du
¨
ren, Germany) using a water ⁄ ACN gradient containing
0.1% TFA. The peptides were lyophilized, and their purity
was verified by analytical HPLC using a LiChroCART col-
umn (LiChrospher 100, RP8, 5 lm; 125 · 4 mm) (Merck)
with the following conditions: gradient of 5–100% ACN in
0.1% TFA at a flow rate of 1 mLÆmin
)1
over 30 min; and
detection at k = 220 nm. The molecular masses of the pep-
tides were confirmed by ESI or MALDI-TOF MS.
A. Heinz et al. Tropoelastin degradation by elastinolytic MMPs
FEBS Journal 277 (2010) 1939–1956 ª 2010 The Authors Journal compilation ª 2010 FEBS 1951

Proteolysis of tropoelastin
For proteolysis using the catalytic domains of either MMP-7
or MMP-12, recombinant tropoelastin was dissolved at a
concentration of 1 mgÆmL
)1
in 50 mm Tris containing
200 mm NaCl and 10 mm CaCl
2
at pH 7.5. The digestions
were performed for 4 h at 37 °C at an enzyme-to-substrate
ratio of 1 : 500 (m ⁄ m). Prior to proteolysis using MMP-9,
the enzyme was activated through incubation with 1.76 mm
aminophenylmercuric acetate for 1 h at 37 °C. Recombi-
nant tropoelastin was dissolved at a concentration of
1mgÆmL
)1
in 50 mm Tris containing 100 mm NaCl and
5mm CaCl
2
at pH 7.5, and subsequently digested with
activated MMP-9 at an enzyme-to-substrate ratio of
1 : 500 (m ⁄ m) for 4 h at 37 °C. Before MS analysis, all
samples were stored at )30 °C.
MALDI-TOF
⁄
TOF MS analysis
Analysis of the digests was conducted by offline nano-
HPLC ⁄ MALDI-TOF ⁄ TOF MS, using a nanoHPLC sys-
tem (Ultimate 3000; Dionex, Idstein, Germany) and an
Ultraflex III MALDI-TOF ⁄ TOF mass spectrometer

equipped with a SmartBeam laser (Bruker Daltonik,
Bremen, Germany). One microliter sample was mixed
with 49 lL of 0.1% TFA solution, loaded onto a trap-
ping column [Acclaim PepMap100 C18, 5 lm, 100 A
˚
,
300 lm internal diameter (i.d.) · 5 mm; Dionex], and
washed for 30 min with 0.1% TFA at a flow rate of
30 lL min
)1
. The trapped peptides were then eluted onto
the separation column (Acclaim PepMap100 C18, 3 lm,
100 A
˚
,75lm i.d. · 150 mm; Dionex), which had been
equilibrated with 95% solvent A (5% ACN ⁄ 95% H
2
O
containing 0.05% TFA). The peptides were separated using
a solvent system of solvent A and solvent B (80%
ACN ⁄ 20% H
2
O containing 0.04% TFA): linear gradient of
5–50% solvent B in 60 min, linear gradient of 50–100% sol-
vent B in 2 min, maintenance at 100% solvent B for 6 min,
and linear gradient of 100–5% solvent B in 2 min. The col-
umn was maintained at 30 °C and the flow rate was 300 nLÆ
min
)1
. Detection was performed by UV absorption at

214 nm and 280 nm. Between 8 min and 75 min of the run,
191 fractions of the effluent (each of 21 s) were spotted,
together with a solution of MALDI matrix (1.1 lL per spot
in total; 0.71 mgÆmL
)1
a-cyano-4-hydroxy-cinnamic acid in
90% ACN ⁄ 10% H
2
O containing 0.1% TFA and 1 mm
NH
4
H
2
PO
4
), onto an AnchorChip MALDI target (384
spots; 800 lm anchors; Bruker Daltonik) using a Protein-
eerFC HPLC ⁄ MALDI fraction collector (Bruker Daltonik).
Mass spectra in the m ⁄ z range 740–5000 were acquired in
positive ionization mode and reflectron mode by accumulat-
ing data from 1800 laser shots per spot. Data acquisition was
performed by flex control 3.0.101.1, which was controlled
by warplc 1.1 (Bruker Daltonik). Ion signals with signal-to-
noise ratio higher than 10 were automatically subjected to
MALDI-LIFT-TOF ⁄ TOF MS ⁄ MS experiments by applying
2000 laser shots.
NanoHPLC–nanoESI–qTOF MS analysis
On-line nanoHPLC, nanoESI MS and MS ⁄ MS experiments
were conducted using an Ultimate nanoHPLC system (LC
Packings ⁄ Dionex) coupled to a quadrupole TOF (qTOF)

mass spectrometer (Q-TOF-2; Waters ⁄ Micromass, Man-
chester, UK). The mass spectrometer was equipped with a
nanoESI Z-spray source and a tip adapter for PicoTips
(New Objective, Woburn, MA, USA). SilicaTips (FS360-
20-10-D-5, 10 lm i.d.) obtained from New Objective were
used. The digests were desalted by washing the sample
bound to an OMIX C18 unit (Varian, Darmstadt,
Germany) according to the manufacturer’s instructions,
prior to nanoHPLC. Chromatographic separation of the
peptides was performed by loading 1 lL of the sample onto
an Acclaim PepMap100 column (C18, 3 lm, 100 A
˚
,75lm
i.d. · 150 mm; Dionex) and elution using a linear binary
gradient: 5–65% ACN in 0.1% formic acid in 35 min,
maintenance at 65% ACN for 10 min, and 65–5% ACN in
10 min.
The typical operating conditions for the mass spectrome-
ter were as follows: positive ion mode; capillary voltage of
1700 V; sample cone voltage of 35 V; and source tempera-
ture of 80 °C. MS experiments were performed over the
m ⁄ z range of 300–2700, and peptides were selected for colli-
sion-induced dissociation either in data-dependent acquisi-
tion mode or after selecting them manually. The
quadrupole mass filter, followed by the TOF analyzer, was
adjusted with low-mass and high-mass resolution settings of
10 (arbitrary units), and the collison energy was varied
between 25 eV and 60 eV depending on charge state and
m/z value.
Data processing and peptide sequencing

The tandem mass spectra obtained from qTOF experiments
were processed by masslynx (version 4.0; Waters ⁄ Micro-
mass) with the add-on Maximum Entropy 3 (maxent3). The
fragment ion spectra obtained in MALDI-TOF ⁄ TOF
MS experiments were analyzed using flex analysis 3.0.54
(Bruker Daltonik). All fragment ion peak lists were submit-
ted to a local mascot server (version 2.2.1; Matrix Science,
London, UK) [85]. The searches were taxonomically
restricted to ‘Homo sapiens’, the enzyme was set to ‘none’,
and the mass error tolerances for precursor and fragment
ions were typically set to 0.15 u for qTOF MS ⁄ MS and
50 p.p.m. and 1 u, respectively, for MALDI-TOF ⁄ TOF MS
data. The formation of pyroglutamic acid moieties from Glu
and Gln were considered as variable modifications.
Moreover, automated de novo sequencing of the nano-
ESI–qTOF and MALDI-TOF ⁄ TOF MS data followed by
database matching (peaks protein id and spider) [86] was
Tropoelastin degradation by elastinolytic MMPs A. Heinz et al.
1952 FEBS Journal 277 (2010) 1939–1956 ª 2010 The Authors Journal compilation ª 2010 FEBS
performed using the in-chorus search of the software
peaks studio (version 4.5; Bioinformatics Solutions,
Waterloo, Canada) [87] with precursor and fragment mass
error tolerances as described for mascot searches. The
enzyme entry was set to ‘unknown’, and the same post-
translational modifications as used for the mascot searches
were considered. For searching purposes, swissprot
extended by the splice variant of the used tropoelastin was
implemented.
Molecular modeling
All calculations were performed on a Pentium IV 2.6 GHz-

based Linux cluster. The molecular modeling calculations
were performed using moe2008.10 modeling software
(Chemical Computing Group, Montreal, Canada). The
X-ray structures of MMP-7 (2ddy.pdb), MMP-9 (2ovx.pdb),
and MMP-12 (2oxw.pdb), were taken from the Protein
Data Bank and prepared using moe. The structures of the
substrate peptides LPYGYGPG (residues 226–233 from
tropoelastin isoform 2; Fig. 1), PQGLAG, and PQGKAG
were generated by taking the coordinates of the prodomain
(YPFALAPT, residues 96–103) of the crystallized proform
of MMP-9 (1lj6.pdb) and the coordinates of the peptide
IAG cocrystallized with MMP-12 (2oxw.pdb) as template.
As the prodomain of MMP-9 and the bound peptide at
MMP-12 show the same conserved hydrogen bonds
between the substrate backbone atoms and the residues of
the binding pocket, the prodomain interaction at
pro-MMP-9 could be used to model the interaction of the
peptide substrate at MMP-12. The prodomain of MMP-9
was inserted into the active site cleft blocking access to the
catalytic zinc. Hydrogen atoms were added and the MMP–
peptide structures were minimized using the amber all-
atom forcefield [88], until the default derivative conver-
gence criterion of 0.01 kcal ⁄ (mol A
˚
) was met. Graphical
analysis of the complexes was carried out using moe.
Fibroblast culture and treatments
Human skin fibroblast strains were established from ex-
plants of human adult skin biopsy samples obtained from
informed healthy volunteers (aged 21–49 years). Cells were

grown as monolayer cultures in Dulbecco’s modified Eagle’s
medium (DMEM) containing 1 gÆL
)1
glucose, glutamax I,
and pyruvate, supplemented with 10% fetal bovine serum in
the presence of 5% CO
2
. Cells at subcultures 4–8 were used.
For experiments, fibroblasts were grown to subconfluence in
the medium containing 10% fetal bovine serum. Before
stimulation, the cells were incubated for 18 h in DMEM
supplemented with 0.5% fetal bovine serum, washed twice
with NaCl ⁄ P
i
, and incubated in serum-free DMEM with or
without elastin peptides (200 lgÆmL
)1
) for 24 h. The cell
media were then harvested and centrifuged at 400 g to
eliminate cellular debris.
Gelatin zymography
To analyze pro-MMP-2 secretion from the conditioned
media of fibroblasts, equal amounts of proteins were ana-
lyzed using SDS ⁄ PAGE with 0.1% gelatin. After electro-
phoresis, the gels were soaked in 2.5% (v ⁄ v) Triton X-100
solution for 1 h to remove SDS. The gels were then incu-
bated in 50 mm Tris ⁄ HCl (pH 7.6) containing 5 mm CaCl
2
and 200 mm NaCl at 37 °C for 24 h. The gels were stained
with 0.1% (m ⁄ v) G250 Coomassie Brilliant Blue in a

solvent system containing 45% methanol and 55% of a
10% acetic acid solution (v/v). After 45 min, the gels were
destained with a solvent system containing 25% methanol
and 75% of a 10% acetic acid solution (v/v). The proteo-
lytic activity was detected as clear bands on a blue back-
ground of the Coomassie Brilliant Blue-stained gel.
Statistical analysis
All biological experiments were performed in triplicate. The
results are expressed as mean ± standard error of the
mean. Comparisons between the groups were made using
Student’s t-test. The results were considered significantly
different at P < 0.05.
Acknowledgements
A. Heinz wishes to thank the Landesgraduiertenfo
¨
rde-
rung Sachsen-Anhalt for financial support. S. Taddese
would like to acknowledge financial support from
Katholischer Akademischer Ausla
¨
nder-Dienst (KAAD).
A. S. Weiss acknowledges support from the Australian
Research Council and Heart Research Foundation.
References
1 Maskos K (2005) Crystal structures of MMPs in com-
plex with physiological and pharmacological inhibitors.
Biochimie 87, 249–263.
2 Visse R & Nagase H (2003) Matrix metalloproteinases
and tissue inhibitors of metalloproteinases: structure,
function, and biochemistry. Circ Res 92, 827–839.

3 Sternlicht MD & Werb Z (2001) How matrix metallo-
proteases regulate cell behavior. Annu Rev Cell Dev Biol
17, 463–516.
4 Nagase H & Woessner JF Jr (1999) Matrix metallopro-
teinases. J Biol Chem 274, 21491–21494.
5 Bode W (2003) Structural basis of matrix metallopro-
teinase function. Biochem Soc Symp 70, 1–14.
6 Verma RP & Hansch C (2007) Matrix metalloproteinas-
es (MMPs): chemical–biological functions and
(Q)SARs. Bioorg Med Chem 15 , 2223–2268.
7 Shipley JM, Doyle GAR, Fliszar CJ, Ye Q-Z, Johnson
LL, Shapiro SD, Welgus HG & Senior RM (1996)
A. Heinz et al. Tropoelastin degradation by elastinolytic MMPs
FEBS Journal 277 (2010) 1939–1956 ª 2010 The Authors Journal compilation ª 2010 FEBS 1953
The structural basis for the elastolytic activity of the
92-kDa and 72-kDa gelatinases. J Biol Chem 271,
4335–4341.
8 Filippov S, Caras I, Murray R, Matrisian LM,
Chapman HA Jr, Shapiro S & Weiss SJ (2003) Matrily-
sin-dependent elastolysis by human macrophages. J Exp
Med 198, 925–935.
9 Murphy G, Cockett MI, Ward RV & Docherty AJ
(1991) Matrix metalloproteinase degradation of elastin,
type IV collagen and proteoglycan. A quantitative com-
parison of the activities of 95 kDa and 72 kDa gelatin-
ases, stromelysins-1 and -2 and punctuated
metalloproteinase (PUMP). Biochem J 277, 277–279.
10 Mecham RP, Broekelmann TJ, Fliszar CJ, Shapiro SD,
Welgus HG & Senior RM (1997) Elastin degradation
by matrix metalloproteinases. Cleavage site specificity

and mechanisms of elastolysis. J Biol Chem 272,
18071–18076.
11 Herouy Y, Vanscheidt W, Scho
¨
pf E & Norgauer J
(1999) Gen- und Proteinfamilie der Matrix-Metallopro-
teinasen Struktur, Funktion und Expression in der
Haut. Der Hautarzt 50, 825–836.
12 Vrhovski B & Weiss AS (1998) Biochemistry of tropo-
elastin. Eur J Biochem 258, 1–18.
13 Mithieux SM & Weiss AS (2005) Elastin. Adv Protein
Chem 70, 437–461.
14 Li B & Daggett V (2002) Molecular basis for the exten-
sibility of elastin. J Muscle Res Cell Motil 23, 561–573.
15 Uitto J, Christiano AM, Kahari VM, Bashir MM &
Rosenbloom J (1991) Molecular biology and pathology
of human elastin. Biochem Soc Trans 19, 824–829.
16 Debelle L & Tamburro AM (1999) Elastin: molecular
description and function. Int J Biochem Cell Biol 31,
261–272.
17 Robinet A, Fahem A, Cauchard J-H, Huet E, Vincent
L, Lorimier S, Antonicelli F, Soria C, Crepin M,
Hornebeck W et al. (2005) Elastin-derived peptides
enhance angiogenesis by promoting endothelial cell
migration and tubulogenesis through upregulation of
MT1-MMP. J Cell Sci 118, 343–356.
18 Maquart FX, Pasco S, Ramont L, Hornebeck W &
Monboisse JC (2004) An introduction to matrikines:
extracellular matrix-derived peptides which regulate cell
activity. Implication in tumor invasion. Crit Rev Oncol

Hematol 49, 199–202.
19 Moroy G, Alix AJP & He
´
ry-Huynh S (2005) Structural
characterization of human elastin derived peptides con-
taining the GXXP sequence. Biopolymers 78, 206–220.
20 Jung S, Rutka JT & Hinek A (1998) Tropoelastin and
elastin degradation products promote proliferation of
human astrocytoma cell lines. J Neuropathol Exp Neurol
57, 439–448.
21 Antonicelli F, Bellon G, Debelle L & Hornebeck W
(2007) Elastin-elastases and inflamm-aging. Curr Top
Dev Biol 79, 99–155.
22 Duca L, Blanchevoye C, Cantarelli B, Ghoneim C,
Dedieu S, Delacoux F, Hornebeck W, Hinek A, Marti-
ny L & Debelle L (2007) The elastin receptor complex
transduces signals through the catalytic activity of its
Neu-1 subunit. J Biol Chem 282, 12484–12491.
23 Hinek A, Rabinovitch M, Keeley F, Okamura-Oho Y
& Callahan J (1993) The 67-kD elastin ⁄ laminin-binding
protein is related to an enzymatically inactive, alterna-
tively spliced form of beta-galactosidase. J Clin Invest
91, 1198–1205.
24 Privitera S, Prody CA, Callahan JW & Hinek A (1998)
The 67-kDa enzymatically inactive alternatively spliced
variant of beta-galactosidase is identical to the elas-
tin ⁄ laminin-binding protein. J Biol Chem 273, 6319–
6326.
25 Mecham RP, Hinek A, Entwistle R, Wrenn DS, Griffin
GL & Senior RM (1989) Elastin binds to a multifunc-

tional 67-kilodalton peripheral membrane protein. Bio-
chemistry 28, 3716–3722.
26 Helske S, Kupari M, Lindstedt KA & Kovanen PT
(2007) Aortic valve stenosis: an active atheroinflamma-
tory process. Curr Opin Lipidol 18, 483–491.
27 Johnson JL, George SJ, Newby AC & Jackson CL
(2005) Divergent effects of matrix metalloproteinases 3,
7, 9, and 12 on atherosclerotic plaque stability in mouse
brachiocephalic arteries. Proc Natl Acad Sci USA 102,
15575–15580.
28 Longo GM, Xiong W, Greiner TC, Zhao Y, Fiotti N &
Baxter BT (2002) Matrix metalloproteinases 2 and 9
work in concert to produce aortic aneurysms. J Clin
Invest 110, 625–632.
29 Curci JA, Liao S, Huffman MD, Shapiro SD &
Thompson RD (1998) Expression and localization of
macrophage elastase (matrix metalloproteinase-12) in
abdominal aortic aneurysms. J Clin Invest 102,
1900–1910.
30 Romanic AM, Harrison SM, Bao W, Burns-Kurtis CL,
Pickering S, Gu J, Grau E, Mao J, Sathe GM, Ohlstein
EH et al. (2002) Myocardial protection from ische-
mia ⁄ reperfusion injury by targeted deletion of matrix
metalloproteinase-9. Cardiovasc Res 54, 549–558.
31 Ka
¨
ha
¨
ri V-M & Saarialho-Kere U (1997) Matrix
metalloproteinases in skin. Exp Dermatol 6, 199–213.

32 Hautamaki RD, Kobayashi DK, Senior RM & Shapiro
SD (1997) Requirement for macrophage elastase for cig-
arette smoke-induced emphysema in mice. Science 277,
2002–2004.
33 Chung JH, Seo JY, Lee MK, Eun HC, Lee JH, Kang
S, Fisher GJ & Voorhees JJ (2002) Ultraviolet modula-
tion of human macrophage metalloelastase in human
skin in vivo. J Invest Dermatol 119, 507–512.
34 Vaalamo M, Kariniemi A-L, Shapiro SD & Saarialho-
Kere U (1999) Enhanced expression of human metallo-
elastase (MMP-12) in cutaneous granulomas and mac-
rophage migration. J Invest Dermatol 112, 499–505.
Tropoelastin degradation by elastinolytic MMPs A. Heinz et al.
1954 FEBS Journal 277 (2010) 1939–1956 ª 2010 The Authors Journal compilation ª 2010 FEBS
35 Wielockx B, Libert C & Wilson C (2004) Matrilysin
(matrix metalloproteinase-7): a new promising drug
target in cancer and inflammation? Cytokine Growth
Factor Rev 15, 111–115.
36 Nagashima Y, Hasegawa S, Koshikawa N, Taki A,
Ichikawa Y, Kitamura H, Misugi K, Kihira Y, Matuo
Y, Yasumitsu H et al. (1997) Expression of matrilysin
in vascular endothelial cells adjacent to matrilysin-pro-
ducing tumors. Int J Cancer 72, 441–445.
37 Taddese S, Weiss AS, Neubert RHH & Schmelzer CEH
(2008) Mapping of macrophage elastase cleavage sites
in insoluble human skin elastin. Matrix Biol 27, 420–
428.
38 Barroso B, Abello N & Bischoff R (2006) Study of
human lung elastin degradation by different elastases
using high-performance liquid chromatography ⁄ mass

spectrometry. Anal Biochem 358, 216–224.
39 Taddese S, Weiss AS, Jahreis G, Neubert RHH &
Schmelzer CEH (2009) In vitro degradation of human
tropoelastin by MMP-12 and the generation of matri-
kines from domain 24. Matrix Biol 28, 84–91.
40 Lombard C, Arzel L, Bouchu D, Wallach J & Saulnier
J (2006) Human leukocyte elastase hydrolysis of pep-
tides derived from human elastin exon 24. Biochimie 88,
1915–1921.
41 Lombard C, Bouchu D, Wallach J & Saulnier J (2005)
Proteinase 3 hydrolysis of peptides derived from human
elastin exon 24. Amino Acids 28, 403–408.
42 Schechter I & Berger A (1967) On the size of the active
site in proteases. I. Papain. Biochem Biophys Res
Commun 27, 157–162.
43 Gronski TJ Jr, Martin RL, Kobayashi DK, Walsh BC,
Holman MC, Huber M, Van Wart HE & Shapiro SD
(1997) Hydrolysis of a broad spectrum of extracellular
matrix proteins by human macrophage elastase. J Biol
Chem 272, 12189–12194.
44 Netzel-Arnett S, Sang QX, Moore WGI, Navre M,
Birkedal-Hansen H & Van Wart HE (1993) Compara-
tive sequence specificities of human 72- and 92-kDa
gelatinases (type IV collagenases) and PUMP (matrily-
sin). Biochemistry 32, 6427–6432.
45 Smith MM, Shi L & Navre M (1995) Rapid identifica-
tion of highly active and selective substrates for strom-
elysin and matrilysin using bacteriophage peptide
display libraries. J Biol Chem 270, 6440–6449.
46 Bisaccia F, Castiglione Morelli MA, De BiasiM,

Traniello S, Spisani S & Tamburro AM (1994) Migra-
tion of monocytes in the presence of elastolytic
fragments of elastin and in synthetic derivates.
Structure–activity relationships. Int J Pept Protein Res
44, 332–341.
47 Castiglione Morelli MA, Bisaccia F, Spisani S, De Biasi
M, Traniello S & Tamburro AM (1997) Structure–activ-
ity relationships for some elastin-derived peptide chemo-
attractants. J Pept Res 49, 492–499.
48 Hance KA, Tataria M, Ziporin SJ, Lee JK &
Thompson RW (2002) Monocyte chemotactic activity
in human abdominal aortic aneurysms: role of elastin
degradation peptides and the 67-kD cell surface elastin
receptor. J Vasc Surg 35, 254–261.
49 Senior R, Griffin G, Mecham R, Wrenn D, Prasad K &
Urry D (1984) Val-Gly-Val-Ala-Pro-Gly, a repeating
peptide in elastin, is chemotactic for fibroblasts and
monocytes. J Cell Biol 99, 870–874.
50 Kamoun A, Landeau J-M, Godeau G, Wallach J,
Duchesnay A, Pellat B & Hornebeck W (1995) Growth
stimulation of human skin fibroblasts by elastin-derived
peptides. Cell Commun Adhes 3, 273–281.
51 Tajima S, Wachi H, Uemura Y & Okamoto K (1997)
Modulation by elastin peptide VGVAPG of cell prolif-
eration and elastin expression in human skin fibroblasts.
Arch Dermatol Res
289, 489–492.
52 Mochizuki S, Brassart B & Hinek A (2002) Signaling
pathways transduced through the elastin receptor facili-
tate proliferation of arterial smooth muscle cells. J Biol

Chem 277, 44854–44863.
53 Simionescu A, Philips K & Vyavahare N (2005) Elastin-
derived peptides and TGF-[beta]1 induce osteogenic
responses in smooth muscle cells. Biochem Biophys Res
Commun 334, 524–532.
54 Debret R, Antonicelli F, Theill A, Hornebeck W,
Bernard P, Guenounou M & Le Naour R (2005) Elas-
tin-derived peptides induce a T-helper type 1 polariza-
tion of human blood lymphocytes. Arterioscler Thromb
Vasc Biol 25, 1353–1358.
55 Long MM, King VJ, Prasad KU, Freeman BA & Urry
DW (1989) Elastin repeat peptides as chemoattractants
for bovine aortic endothelial cells. J Cell Physiol 140,
512–518.
56 Pocza P, Su
¨
li-Vargha H, Darvas Z & Falus A (2008)
Locally generated VGVAPG and VAPG elastin-derived
peptides amplify melanoma invasion via the galectin-3
receptor. Int J Cancer 122, 1972–1980.
57 Fahem A, Robinet A, Cauchard JH, Duca L,
Soula-Rothhut M, Rothhut B, Soria C, Guenounou M,
Hornebeck W & Bellon G (2008) Elastokine-mediated
up-regulation of MT1-MMP is triggered by nitric oxide
in endothelial cells. Int J Biochem Cell Biol 40, 1581–1596.
58 Karnik SK, Wythe JD, Sorensen L, Brooke BS, Urness
LD & Li DY (2003) Elastin induces myofibrillogenesis via
a specific domain, VGVAPG. Matrix Biol 22, 409–425.
59 Brassart B, Fuchs P, Huet E, Alix AJP, Wallach J,
Tamburro AM, Delacoux F, Haye B, Emonard H,

Hornebeck W et al. (2001) Conformational dependence
of collagenase (matrix metalloproteinase-1) up-regula-
tion by elastin peptides in cultured fibroblasts. J Biol
Chem 276, 5222–5227.
60 Faury G, Garnier S, Weiss AS, Wallach J, Fu
¨
lo
¨
pT,
Jacob M-P, Mecham RP, Robert L & Verdetti J (1998)
Action of tropoelastin and synthetic elastin sequences
A. Heinz et al. Tropoelastin degradation by elastinolytic MMPs
FEBS Journal 277 (2010) 1939–1956 ª 2010 The Authors Journal compilation ª 2010 FEBS 1955
on vascular tone and on free Ca
2+
level in human vas-
cular endothelial cells. Circ Res 82, 328–336.
61 Fuchs P, Debelle L & Alix AJP (2001) Structural study
of some specific elastin hexapeptides activating MMP1.
J Mol Struct 565–566, 335–339.
62 Lin S-Y, Hsieh T-F & Wei Y-S (2005) pH- and ther-
mal-dependent conformational transition of PGAIPG, a
repeated hexapeptide sequence from tropoelastin.
Peptides 26, 543–549.
63 Wachi H, Seyama Y, Yamashita S, Suganami H, Uemura
Y, Okamoto K, Yamada H & Tajima S (1995) Stimula-
tion of cell proliferation and autoregulation of elastin
expression by elastin peptide VPGVG in cultured chick
vascular smooth muscle cells. FEBS Lett 368, 215–219.
64 Spezzacatena C, Pepe A, Green LM, Sandberg LB,

Bochicchio B & Tamburro AM (2005) Synthesis, solu-
tion structure and biological activity of Val-Val-Pro-
Gln, a bioactive elastin peptide. Eur J Org Chem 2005,
1644–1651.
65 Duca L, Floquet N, Alix AJ, Haye B & Debelle L
(2004) Elastin as a matrikine. Crit Rev Oncol Hematol
49, 235–244.
66 Toonkool P, Jensen SA, Maxwell AL & Weiss AS
(2001) Hydrophobic domains of human tropoelastin
interact in a context-dependent manner. J Biol Chem
276, 44575–44580.
67 Mithieux SM, Wise SG, Raftery MJ, Starcher B &
Weiss AS (2005) A model two-component system for
studying the architecture of elastin assembly in vitro.
J Struct Biol 149, 282–289.
68 Djajamuliadi J, Kagawa TF, Ohgo K & Kumashiro
KK (2009) Insights into a putative hinge region in elas-
tin using molecular dynamics simulations. Matrix Biol
28, 92–100.
69 Dyksterhuis LB, Baldock C, Lammie D, Wess TJ &
Weiss AS (2007) Domains 17–27 of tropoelastin contain
key regions of contact for coacervation and contain an
unusual turn-containing crosslinking domain. Matrix
Biol 26, 125–135.
70 Wise SG, Mithieux SM, Raftery MJ & Weiss AS (2005)
Specificity in the coacervation of tropoelastin: solvent
exposed lysines. J Struct Biol 149, 273–281.
71 Vrhovski B, Jensen S & Weiss AS (1997) Coacervation
characteristics of recombinant human tropoelastin. Eur
J Biochem 250, 92–98.

72 Muiznieks LD & Weiss AS (2007) Flexibility in the
solution structure of human tropoelastin. Biochemistry
46, 8196–8205.
73 Terp GE, Cruciani G, Christensen IT & Jorgensen FS
(2002) Structural differences of matrix metalloproteinas-
es with potential implications for inhibitor selectivity
examined by the GRID ⁄ CPCA approach. J Med Chem
45, 2675–2684.
74 Rowsell S, Hawtin P, Minshull CA, Jepson H, Brock-
bank SMV, Barratt DG, Slater AM, McPheat WL,
Waterson D, Henney AM et al. (2002) Crystal structure
of human MMP9 in complex with a reverse hydroxa-
mate inhibitor. J Mol Biol 319, 173–181.
75 Massova I, Kotra LP, Fridman R & Mobashery S
(1998) Matrix metalloproteinases: structures, evolution,
and diversification. FASEB J 12, 1075–1095.
76 Browner MF, Smith WW & Castelhano AL (1995)
Crystal structures of matrilysin–inhibitor complexes.
Biochemistry 34, 6602–6610.
77 Wilson CL & Matrisian LM (1996) Matrilysin: an epi-
thelial matrix metalloproteinase with potentially novel
functions. Int J Biochem Cell Biol 28, 123–136.
78 Kozel BA, Wachi H, Davis EC & Mecham RP
(2003) Domains in tropoelastin that mediate elastin
deposition in vitro and in vivo. J Biol Chem 278,
18491–18498.
79 Deng SJ, Bickett DM, Mitchell JL, Lambert MH,
Blackburn RK, Carter HL, Neugebauer J, Pahel G,
Weiner MP & Moss ML (2000) Substrate specificity of
human collagenase 3 assessed using a phage-displayed

peptide library. J Biol Chem 275, 31422–31427.
80 Wise SG & Weiss AS (2009) Tropoelastin. Int J Bio-
chem Cell Biol 41, 494–497.
81 Martin SL, Vrhovski B & Weiss AS (1995) Total syn-
thesis and expression in Escherichia coli of a gene
encoding human tropoelastin. Gene 154, 159–166.
82 Wu WJ, Vrhovski B & Weiss AS (1999) Glycosamino-
glycans mediate the coacervation of human tropoelastin
through dominant charge interactions involving lysine
side chains. J Biol Chem 274, 21719–21724.
83 Fazio MJ, Olsen DR, Kauh EA, Baldwin CT, Indik Z,
Ornstein-Goldstein N, Yeh H, Rosenbloom J & Uitto
J (1988) Cloning of full-length elastin cDNAs from a
human skin fibroblast recombinant cDNA library:
further elucidation of alternative splicing utilizing exon-
specific oligonucleotides. J Invest Dermatol 91, 458–464.
84 Lang R, Kocourek A, Braun M, Tschesche H, Huber R,
Bode W & Maskos K (2001) Substrate specificity deter-
minants of human macrophage elastase (MMP-12) based
on the 1.1 A
˚
crystal structure. J Mol Biol 312, 731–742.
85 Perkins DN, Pappin DJC, Creasy DM & Cottrell JS
(1999) Probability-based protein identification by
searching sequence databases using mass spectrometry
data. Electrophoresis 20, 3551–3567.
86 Han Y, Ma B & Zhang K (2005) SPIDER: software for
protein identification from sequence tags with de novo
sequencing error. J Bioinform Comput Biol 3, 697–716.
87 Ma B, Zhang K, Hendrie C, Liang C, Li M, Doherty-

Kirby A & Lajoie G (2003) PEAKS: powerful software
for peptide de novo sequencing by tandem mass spec-
trometry. Rapid Commun Mass Spectrom 17, 2337–2342.
88 Case DA, Cheatham TE III, Darden T, Gohlke H, Luo
R, Merz KM Jr, Onufriev A, Simmerling C, Wang B &
Woods RJ (2005) The Amber biomolecular simulation
programs. J Comput Chem 26, 1668–1688.
Tropoelastin degradation by elastinolytic MMPs A. Heinz et al.
1956 FEBS Journal 277 (2010) 1939–1956 ª 2010 The Authors Journal compilation ª 2010 FEBS