OUP user menu

Structure and function of matrix metalloproteinases and TIMPs

Hideaki Nagase , Robert Visse , Gillian Murphy
DOI: http://dx.doi.org/10.1016/j.cardiores.2005.12.002 562-573 First published online: 15 February 2006


Matrix metalloproteinases (MMPs), also called matrixins, function in the extracellular environment of cells and degrade both matrix and non-matrix proteins. They play central roles in morphogenesis, wound healing, tissue repair and remodelling in response to injury, e.g. after myocardial infarction, and in progression of diseases such as atheroma, arthritis, cancer and chronic tissue ulcers. They are multi-domain proteins and their activities are regulated by tissue inhibitors of metalloproteinases (TIMPs). This review introduces the members of the MMP family and discusses their domain structure and function, proenyme activation, the mechanism of inhibition by TIMPs and their significance in physiology and pathology.

1. Introduction

Timely degradation of extracellular matrix (ECM) is an important feature of development, morphogenesis, tissue repair and remodelling. It is precisely regulated under normal physiological conditions, but when dysregulated it becomes a cause of many diseases such as arthritis, nephritis, cancer, encephalomyelitis, chronic ulcers, fibrosis, etc. Uncontrolled ECM remodelling of the myocardium and vasculature are features of cardiovascular disorders such as atherosclerosis, stenosis, left ventricular hypertrophy, heart failure, and aneurysm [1–3]. Various types of proteinases are implicated in ECM degradation, but the major enzymes are considered to be matrix metalloproteinases (MMPs), also called matrixins [4]. Humans have 24 matrixin genes including duplicated MMP-23 genes; thus there are 23 MMPs in humans. The activities of most matrixins are very low or negligible in the normal steady-state tissues, but expression is transcriptionally controlled by inflammatory cytokines, growth factors, hormones, cell–cell and cell–matrix interaction [5]. Matrixin activities are also regulated by activation of the precursor zymogens and inhibition by endogenous inhibitors, tissue inhibitors of metalloproteinases (TIMPs). Thus, the balance between MMPs and TIMPs are critical for the eventual ECM remodelling in the tissue. This review describes the structure and function of MMPs and TIMPs, highlighting recent progress.

2 Biological activities of MMPs and cardiovascular diseases

The main function of matrixins has been considered to be the degradation and removal of ECM molecules from the tissue. However, it has been increasingly recognised that breakdown of ECM molecules or cell surface molecules alters cell–matrix and cell–cell interactions and the release of growth factors that are bound to the ECM makes them available for cell receptors. In addition, a number of non-ECM molecules are also potential substrates of MMPs. Table 1 lists examples of MMP actions that may affect cell migration, differentiation, growth, inflammatory processes, neovacularization, apoptosis, etc.

View this table:
Table 1

Biological activities mediated by MMP cleavage

Studies using MMP gene knockout mice have indicated that MMP-2 and MMP-9 play key roles in cardiac rupture after myocardial infarction [6–9]. A recent study showed that MT1-MMP (MMP-14) is also increased after ischemia-reperfusion [10]. TIMP-3 deficiency in mice disrupts matrix homeostasis and causes spontaneous left ventricular dilation, cardiomyocyte hypertrophy and contractile dysfunction [11]. A critical role of MMP-2 and MMP-9 has been also shown for the development of abdominal aortic aneurysm using MMP gene deletion mice [12]. Both MMP-2 and MMP-9 cleave elastin, type IV collagen, and several other ECM molecules and MMP-2 digests interstitial collagen types I, II and III. Intracellular MMP-2 found in cardiomyocytes digests troponin I [13], myosin light chain [14] and poly(ADP-ribose) polymerase [15], which may contribute cardiac dysfunction. Recent studies on atherosclerotic plaque stability using a series of apoE/MMP double knockout mice have indicated that MMP-3 and MMP-9 have protective roles by limiting plaque growth and enhancing plaque stability, but MMP-12 promotes lesion expansion and destabilization [16]. A number of MMP gene polymorphisms have shown to contribute to inter-individual susceptibility and outcome of several cardiovascular diseases including coronary artery disease, stenosis, myocardial infarction, coronary aneurysm, stroke, large artery stiffness (see [17]). In many cases they are mostly attributed to different promoter activities due to polymorphic sites.

3 General structural features of MMPs

MMPs are classified as the matrixin subfamily of zinc metalloprotease family M10 in the MEROPS database (http://www.merops.sanger.ac.uk/). Those in vertebrates are assigned with MMP numbers and some members have trivial names (Table 2). MMP-4, MMP-5, MMP-6 and MMP-22 are missing in the list since they were shown to be identical to other members. MMPs are extracellular proteins, but recent studies have indicated that MMP-1 [18], MMP-2 [19] and MMP-11 [20] are also found intracellularly and may act on intracellular proteins. A typical MMP consists of a propeptide of about 80 amino acids, a catalytic metalloproteinase domain of about 170 amino acids, a linker peptide of variable lengths (also called the ‘hinge region’) and a hemopexin (Hpx) domain of about 200 amino acids. Exceptions to this are MMP-7 (matrilysin 1), MMP-26 (matrilysin 2) and MMP-23; they lack the linker peptide and the Hpx domain and MMP-23 has a unique cysteine-rich domain and an immunoglobulin-like domain after the metalloproteinase domain. Two gelatinases, gelatinase A (MMP-2) and gelatinase B (MMP-9), have three repeats of a fibronectin type II motif in the metalloproteinase domain. The zinc binding motif HEXXHXXGXXH in the catalytic domain, and the “cysteine switch” motif PRCGXPD in the propeptide are common structural signatures, where three histidines in the zinc binding motif coordinate and the cysteine in the propetide coordinate with the catalytic zinc ion. This Cys-Zn2+ coordination keeps proMMPs inactive by preventing a water molecule essential for catalysis from binding to the zinc atom. The catalytic domain also contains a conserved methionine, forming a “Met-turn” eight residues after the zinc binding motif, which forms a base to support the structure around the catalytic zinc [21]. The zinc binding motif and the Met-turn are also conserved in members of the ADAM (a disintegrin and metalloproteinase) family, the ADAMTS (ADAM with thrombospondin motifs) family, the astacin family, the bacterial serralysin family, a protozoan proteinase leishmanolysin and 2 pregnancy associated plasma proteins. A (pappalysins, PAPP-A), and they are collectively called “metzincins” [21]. While the primary structures of these metalloproteinase domains have little homology among the families, the overall protein folds are similar [22].

View this table:
Table 2

Matrix metalloproteinases and their domain composition

4 MMPs and their domain arrangements

Based on domain organization and substrate preference, MMPs are grouped into collagenases, gelatinases, stromelysins, matrilysins, membrane-type (MT)-MMPs and others. Fig. 1 shows variations in MMP domain arrangement and Table 2 describes the composition of each MMP member.

Fig. 1

Domain structures of the MMP family. See Table 1 for the domain arrangement for each MMP. sp, signal sequence; pro, pro-domain; cat, catalytic domain, FNII, fibronectin type II motif; L1, linker 1; Hpx, hemopexin domain; L2, linker 2; Mb, plasma membrane; TM, transmembrane domain; Cy, cytoplasmic tail; CysR, cysteine rich; Ig, immunoglobulin domain; GPI, glycosylphosphatidilyinositol anchor.

Collagenases (MMP-1, MMP-8, MMP-13 and MMP-18 in Xenopus) cleave interstitial collagens I, II and III into characteristic 3/4 and 1/4 fragments but they can digest other ECM molecules and soluble proteins (see [4,23] for detail]. Recent studies indicated that MMP-1 activates protease activated receptor (PAR) 1 by cleaving the same Arg-Ser bond cleaved by thrombins, which promotes growth and invasion of breast carcinoma cells [24]. Two other matrixins, ie, MMP-2 and MMP-14 (MT1-MMP), have collagenolytic activity, but they are classified into other subgroups because of their domain compositions.

Gelatinases (MMP-2 and MMP-9) readily digest gelatin with the help of their three fibronectin type II repeats that binds to gelatin/collagen. They also digest a number of ECM molecules including type IV, V and XI collagens, laminin, aggrecan core protein, etc. MMP-2, but not MMP-9, digests collagens I, II and III in a similar manner to the collagenases [25,26]. The collagenolytic activity of MMP-2 is much weaker than MMP-1 in solution, but because proMMP-2 is recruited to the cell surface and activated by the membrane-bound MT-MMPs, it may express reasonable collagenolytic activity on or near the cell surface.

Stromelysins (MMP-3, MMP-10 and MMP-11) have a domain arrangement similar to that of collagenases, but they do not cleave interstitial collagens. MMP-3 and MMP-10 are similar in structure and substrate specificity, but MMP-11 (stromelysin 3) is distantly related. The MMP-11 gene is located on chromosome 22, whereas MMP-3 and MMP-10 are on chromosome 11, along with MMP-1, -7, -8, -12, -20, -26 and -27. MMP-3 and MMP-10 digest a number of ECM molecules and participate in proMMP activation. MMP-11, on the other hand, has very weak activity toward ECM molecules [27], but cleaves serpins more readily [28]. MMP-11 has a furin recognition motif RX[R/K]R at the C-terminal end of the propeptide and therefore it is activated intracellularly [29]. An intracellular 40-kDa MMP-11 isoform (β-stromelysin 3) is found in cultured cells and placenta [30]. This transcript, resulting from alternative splicing and promoter usage, lacks the signal peptide and the pro-domain. The function of this isoenzyme is not known.

Matrilysins (MMP-7 and -26) lack a hemopexin domain. MMP-7 is synthesized by epithelial cells and is secreted apically. Besides ECM components it processes cell surface molecules such as pro-α-defensin, Fas-ligand, pro-tumor necrosis factor α, and E-cadherin. MMP-26 is expressed in normal cells such as those of the endometrium and in some carcinomas. It digests several ECM molecules, and unlike most other MMPs, it is largely stored intracellularly [31].

MT-MMPs in mammals includes four type I transmembrane proteins (MMP-14, -15, -16, and -24) and two glycosylphosphatidylinositol-anchored proteins (MMP-17 and -25). They all have a furin recognition sequence RX[R/K]R at the C-terminus of the propeptide. They are therefore activated intracellularly and active enzymes are likely to be expressed on the cell surface. All MT-MMPs, except MT4-MMP (MMP-17) [32] can activate proMMP-2. MT1-MMP (MMP-14) has collagenolytic activity on collagens I, II, and III [33]. MT1-MMP null mice exhibit skeletal abnormalities during post-natal development, which is attributed to the lack of collagenolytic activity [34].

Seven MMPs are not grouped in the above categories although MMP-12, MMP-20 and MMP-27 have similar structures and chromosome location as stromelysins (see Table 1). Metalloelastase (MMP-12) is expressed primarily in macrophages, but is also found in hypertrophic chondrocytes [35] and osteoclasts [36]. It digests elastin and a number of ECM molecules. It is essential in macrophage migration [37].

MMP-19 digests many ECM molecules including the components of basement membranes [38]. It is also called RASI (rheumatoid arthritis synovial inflammation) as it is found in the activated lymphocytes and plasma from patients with rheumatoid arthritis and it is also recognised as an autoantigen in patients with rheumatoid arthritis and systemic lupus erythematosis [39]. It is, however, widely expressed in many organs including proliferating keratinocytes in healing wounds [40].

Enamelysin (MMP-20) is expressed in newly formed tooth enamel and digests amelogenin [41].

MMP-21 was originally found in Xenopus [42] and more recently in mice and humans [43]. It is expressed in various fetal and adult tissues and in basal and squamous cell carcinomas [44]. It digests gelatin, but information about the action on ECM components is not known.

MMP-23 is a unique member as it has unique C-terminal cysteine-rich immunoglobulin-like domains instead of a hemopexin domain [45]. The propeptide lacks a cysteine switch. It is proposed to be a type II membrane protein having a transmembrane domain at the N-terminal of the propeptide, but the enzyme is released from the cell as the membrane anchored propeptide is cleaved by a proprotein convertase [46].

MMP-27 was first found in chicken embryo fibroblasts [47]. Chicken MMP-27 digests gelatin and casein and causes autolysis of the enzyme, but little information is available on the activity of mammalian enzyme.

Epilysin (MMP-28) is expressed in many tissues such as lung, placenta, heart, gastrointestinal tract and testis. The enzyme expressed in basal keratinocytes in skin is considered to function in wound repair [48]. It is also elevated in cartilage from patients with osteoarthritis [49] and rheumatoid arthritis [50].

MMP-21, MMP-23 and MMP-28 have a furin recognition sequence before the catalytic domain and therefore they are likely to be activated intracellularly and secreted as active enzymes.

5 Three-dimensional (3D) structures of MMPs

In 1994 3D structures of the catalytic domain of collagenases (MMP-1 and MMP-8) were determined by X-ray crystallography by several groups [51–54] and subsequently crystal structures of proMMP-3 lacking the hemopexin domain [55] and the active full-length MMP-1 [56] in 1995. Since then, a large number of 3D structures of MMPs have been determined both by X-ray crystallography and by NMR spectroscopy (see [23,57]) including full-length proMMP-1 [58], proMMP-2 [59] and the proMMP-2–TIMP-2 complex [60].

The polypeptide folds of MMP catalytic domains are essentially superimposable. The domain consists of a 5-stranded β-pleated sheet, three α-helices and connective loop (see Fig. 2). It contains two zinc ions (one catalytic and one structural), and up to three calcium ions which stabilize the structure. MT1-MMP (MMP-14), MT2-MMP (MMP-15), MT3-MMP (MMP-16) and MT5-MMP (MMP-24) have an additional 8 residues between β-strand II and III. This loop is critical for activation of proMMP-2 as shown for MT1-MMP [61]. In the orientation shown in Fig. 2A-II, a peptide substrate binds to the catalytic domain from left to right with respect to the N- and C-termini, and the carbonyl group of the peptide bond to be cleaved coordinates with the catalytic zinc. The substrate docking is dictated by the structure of the substrate binding site, including a pocket called the “S1′ pocket”, located to the right of the zinc atom. This pocket is hydrophobic in nature, but variable in depth among MMPs. It is therefore one of the determining factors of substrate specificity of MMPs. Binding of a substrate to the enzyme displaces the water molecule from the zinc. The peptide bond hydrolysis is then facilitated by the carboxylate group of the glutamate in the active site which draws a proton from the displaced water and allows a nucleophilic attack of the polarized water on the carbonyl carbon of the peptide bond.

Fig. 2

(A) Ribbon diagram of human proMMP-1 and active pig MMP-1. The pro-domain is shown in red, catalytic domain in pink, the linker region in yellow, the hemopexin domain in green, zinc ions in purple, calcium ions in grey. The dotted circle indicates the region where the catalytic and hemopexin domains intereact. Note that proMMP-1 has a larger area of contact sites than MMP-1. This results in the active form has an “open” configuration compared the “closed” configuration of proMMP-1. It is predicted that this region is where triple helical collagens binds. (B) Ribon structure of the complex of proMMP-1 and TIMP-2. Domains of proMMP-2 are shown as in (A) and the finbronectin type II (FNII) motif is in purple. TIMP-2 is shown in blue. The image was prepared from Bookhaven Protein Data bank entries ISU3 (proMMP-1), 1FBL(MMP-1) and 1GXD (proMMP-2-TIMP-2) using the Pymol (http://www.pymol.org).

The propeptide domain consists of three α chains and connecting loops. The proteinase susceptible “bait region” is located between helix 1 and helix 2, but the structure has not been resolved due to flexible nature of the region, except that of proMMP-2 which is stabilised by a disulphide bond. The “cysteine switch” lies in the substrate binding pocket but the orientation of this polypeptide segment is opposite from that of a peptide substrate. The SH group of the cysteine interacts with the catalytic zinc ion. Upon activation the interaction of Cys–Zn2+ is disrupted, which allows a water molecule to bind to the zinc atom. In the case of proMMP-1, the pro-domain interacts with the hemopexin domain [58], rendering it in a “closed” configuration of proMMP-1 in contrast to the “open” configuration of the active MMP-1 [56]. This change in configuration is considered to create the collagen binding site. The linker region connects the catalytic domain and the hemopexin domain, and mutation of this region in MMP-1 [62] and MMP-8 [63] significantly reduced the collagenolytic activity. This may be due to the restrictions of movement between the catalytic domain and the hemopexin domain introduced by mutagenesis. In proMMP-2, the propeptide interacts with the third motif of the three fibronectin type II repeats as well as with the catalytic domain [59] (see Fig. 2B).

Three repeats of fibronectin type II domains present in the loop between the fifth β-strand and the catalytic site helix consist of two antiparallel β-sheets connected with a short α-helix and stabilized by two disulfide bonds. Domains 2 and 3 are flexible and possibly interact simultaneously with multiple sites of the ECM [64].

The hemopexin domain has a 4-bladed β-propeller structure with a single disulfide bond between the first and the fourth blades. The center of the propeller generally contains one calcium ion and a chloride (Fig. 2A,B). TIMP-2 binds to the hemopexin domain of proMMP-2 through interaction of its C-terminal domain and blades III and IV of the hemopexin domain. This leaves the N-terminal inhibitory domain of TIMP-2 free to interact with an active MMP (Fig. 2B).

6 Activation of proMMPs

MMPs are synthesized as pre-proenzymes. The signal peptide is removed during translation and proMMPs are generated. Activation of these zymogens is therefore an important regulatory step of MMP activity.

Thirteen MMPs are secreted from the cell as proMMPs. The presence of a proteinase susceptible “bait” region in the propeptide allows tissue and plasma proteinases or opportunistic bacterial proteinases to activate proMMPs. Cleavage of the bait region removes only a part of the propeptide and complete removal of the propeptide is often conducted in trans by the action of the MMP intermediate or by other active MMPs. This is the mechanism referred to as “stepwise activation” [65]. Ten proMMPs possess a furin-like proprotein convertase recognition sequence RX[K/R]R at the end of the propeptide and they are likely to be activated intracellularly and secreted or cell surface-bound as active enzymes. The activities of these MMPs are regulated by tissue specific location of the enzyme and inactivation by endogenous inhibitors or proteolysis. In the case of MT1-MMP, it is rapidly endocytosed and partially recycled to the cell surface [66,67]. In atherosclerotic plaque macrophages furin and PC5 are proprotein convertases which activate pro-MT1-MMP intracellularly [68].

Another unique feature of MMPs is that many of them are readily activated by treatment with mercurial compounds, SH reagents and chaotropic agents, probably due to perturbation of the molecule. This property is used to activate proMMPs in the laboratory. On the other hand oxidants such as HOCl and ONOO activate proMMPs by reacting with the cysteine of the cysteine switch in the propeptide and this activation process may take place under inflammatory conditions [69,70]. However, an extended reaction may inactivate the enzyme [71].

7 The role of non-catalytic domain of MMPs in substrate specificity: An insight into collagenolysis

Non-catalytic ancillary domains play a key role for some matrixins in determining substrate specificity. Fibronectin type II domains are important for gelatinases to effectively cleave type IV collagen, elastin, and gelatins, but they do not affect hydrolysis of small peptides [72,73]. The hemopexin domain is essential for collagenolytic activity of collagenases. The catalytic domains of these MMPs alone can hydrolyze non-collagenous proteins and synthetic substrates [74], but they cannot cleave triple helical collagens without the hemopexin domain. Recent studies by Chung et al. [75] have shown that collagenases unwind the triple helical chains before they cleave each α chain. It is postulated that the collagen groove consists of the catalytic domain and the hemopexin domain (see Fig. 2A). This groove is narrower in the proMMP-1 (‘closed’ form) and the pro-domain partially blocks the putative collagen binding site (Fig. 2A). Upon activation the groove is opened up due to the removal of the pro-domain. This explains the earlier observation that proMMP-1 does not bind to collagen I, but the activated collagenase does [76,77]. However, it is still not clear how collagenases unwind the triple helical strands, but effective unwinding is probably due to cooperative molecular interactions of the two domains and the triple helical substrate.

8 The role of hemopexin domains in proMMP-2 activation by MT1-MMP (MMP-14)

ProMMP-2 forms a tight complex with TIMP-2 through hemopexin domain interactions with the non-inhibitory C-terminal domain of the inhibitor. This complex formation is essential for proMMP-2 activation by MT1-MMP on the cell surface [78]. The complex binds to an active MT1-MMP through the free N-terminal MMP inhibitory domain of TIMP-2, which orients the propeptide of proMMP-2 to an adjacent active MT1-MMP. Two MT1-MMP molecules interact with each other on the cell surface through their hemopexin domains, thus forming a tetrameric quaternary proMMP-2 activation complex; one MT1-MMP molecule acts as a receptor of the proMMP-2–TIMP-2 complex and the other as an activator of proMMP-2. An excess TIMP-2 prevents this activation process by inhibiting the second MT1-MMP. ProMMP-2 also forms a complex with TIMP-3 and TIMP-4 possibly in a similar manner. The proMMP-2–TIMP-4 complex interacts with MT1-MMP, but this complex is non-productive in terms of MMP-2 activation [79]. The biological significance of the proMMP-2–TIMP-3 complex is not known.

The hemopexin domain of proMMP-9 also forms a tight complex with TIMP-1 and TIMP-3 through their C-terminal domains. ProMMP-9 in neutrophils partially binds to neutrophil gelatinase associated lipocalin-like molecule (NGAL) through an intermolecular disulfide bond [80]. However, the biological significance of these complexes is not known, except that proMMP-9–TIMP complexes are potential inhibitors of metalloproteinases.

9 Palmitoylation of MT1-MMP and cell migration

While MT1-MMP can activate proMMP-2 and proMMP-13 on the cell surface, it also degrades interstitial collagens and other ECM molecules [78]. These activities are often found at the invasive front when cells migrate through ECM. How MT1-MMP is localized to a particular section of the cell surface is yet to be investigated, but a recent study by Anilkumar et al. [81] showed that a unique cysteine in the cytoplasmic domain is posttranslationally palmitoylated and this modification is essential for the function of MT1-MMP to promote cell migration.

Once appeared on the cell surface, MT1-MMP is rapidly internalized. This is occurs via clathrin-coated pits and involves the binding of the cytoplasmic domain of MMP-14 to μ2 subunit of adaptor protein 2 through the LLY (571–573) motif [82]. Internalized MT1-MMP is recycled back to the plasma membrane directed by the DKV (580–582) motif in the cytoplasmic domain [83]. Such dynamic trafficking may be important for regulation of MT-MMP activity and its role in cell migration.

10 Endogenous inhibitors of MMPs

MMP activities are regulated by two major types of endogenous inhibitors: α2-macroglobulin and TIMPs. Human α2-macroglobin is a plasma glycoprotein of 725 kDa consisting of four identical subunits of 180 kDa. It inhibits most proteinases by entrapping the proteinase within the macroglobulin and the complex is rapidly cleared by the receptor (low density lipoprotein receptor-related protein-1)-mediated endocytosis [84]. MMP activities in the fluid phase are primarily regulated by α2-macroglobulin and related proteins.

TIMPs, consisting of 184–194 amino acids, are inhibitors of MMPs. They are subdivided into an N-terminal and a C-terminal subdomain. Each domain contains three conserved disulfide bonds and the N-terminal domain folds as an independent unit with MMP inhibitory activity. TIMPs inhibit all MMPs tested so far, but TIMP-1 is a poor inhibitor for MT1-MMP, MT3-MMP, MT5-MMP and MMP-19. TIMP-3 has been shown to inhibit ADAMs (ADAM-10, -12 and -17) and ADAMTSs (ADAMTS-1, -4 and -5). TIMP-1 inhibits ADAM-10. While TIMP-1-null mice and TIMP-2-null mice do not exhibit obvious abnormalities, TIMP-3 ablation in mice causes lung emphysema-like alveolar damage [85] and faster apoptosis of mammary epithelial cells after weaning [86], indicating that TIMP-3 is a major regulator of metalloproteinase activities in vivo. Impaired liver regeneration after partial hepatectomy in TIMP-3-null mice is caused by abnormal inflammation, with elevated tumor necrosis factor-α release, suggesting that ADAM17 (TACE) is one of the major targets of TIMP-3 in vivo [87].

Several other proteins have been reported to inhibit selected members of MMPs: the secreted form of β-amyloid precursor protein inhibits MMP-2 [88]; A C-terminal fragment of procollagen C-proteinase enhancer protein inhibits MMP-2 [89], and RECK, a GPI-anchored glycoprotein that suppresses angiogenesis inhibits MMP-2, MMP-9 and MMP-14 [90]. However the inhibition mechanisms of these proteins are not known. Tissue factor pathway inhibitor-2, a serine proteinase inhibitor, was reported to inhibit MMP-1 and MMP-2 [91], but this effect is controversial [92].

11 Inhibition mechanism of TIMPs and TIMP variants

The mechanism of TIMP inhibition of MMPs has been elucidated based on the crystal structures of the TIMP-MMP complexes [93,94]. The overall shape of the TIMP molecule is “wedge-like” and the N-terminal four residues Cys1-Thr-Cys-Val4 and the residues Glu67-Ser-Val-Cys70 (residues are in TIMP-1) that are linked by a disulfide from a contiguous ridge that slots into the active site of the MMPs. This region occupies about 75 % of the protein–protein interaction in case of the complex of the catalytic domain of MMP-3 and TIMP-1. The catalytic zinc atom is bidentately chelated by the N-terminal amino group and the carbonyl group of Cys1, which expels the water molecule bound to the zinc atom.

Mutation of the position 2 (Thr in TIMP-1) greatly affects the affinity of TIMPs for MMPs and substitution to glycine essentially inactivates TIMP-1 for MMP inhibition [95]. Additional double and triple mutations at position Val4 and Ser68 generated inhibitors discriminatory between MMP-1, -2 and -3 [96]. The major difference between TIMP-1 and TIMP-2 is that TIMP-2 has a longer AB loop (Fig. 3). The latter contains Ile35, Tyr36 and Asn38 that fit into a special cavity on the surface of the MT1-MMP molecule, and of these Tyr36 play a key role in interaction with MT1-MMP [97]. TIMP-1 is a very poor inhibitor of MT1-MMP, MMP-19 and ADAM17, but it gains reactivity for all three metalloproteinases by replacing Thr98 with Leu, along with modifications to the AB loop [98]. TIMP-2 which does not inhibit ADAM17 may be converted to a functional inhibitor by replacing the AB loop with that of with TIMP-3, in combination with S2T/A70S/V71L mutations [99]. Full-length TIMP-4 is a weak inhibitor of ADAM17, but truncation of the C-terminal domain increased the reactivity. Further replacement with the TIMP-3 AB loop significantly improved the inhibitory activity with a sub-nanomolar Ki [100].

Fig. 3

(A) Ribbon structure of TIMP-1 and TIMP-2. The residues of Cys1, Thr2, Cys3, Val4, Ser68 and Thr98 in TIMP-1 and the corresponding residues in TIMP-2 are shown as spheres. Disulfide bonds are in yellow. The image was prepared from Bookhaven Protein Data bank entry 1UEA (TIMP-1) and 1BQQ (TIMP-2) using Pymol.

Carbamylation of the N-terminal amino group [101] or an addition of an extra Ala at the N-terminus [102] inactivates TIMPs with respect to MMP inhibition. These alterations weaken the metal ion chelating ability of TIMPs. However, recent studies of Wei et al. [103] have shown that addition of Ala to the N-terminus of TIMP-3 did not significantly alter the ability to inhibit ADAM-17, although it impaired the inhibition of MMPs. Mutation of Thr2 of TIMP-3 to Gly also inactivates MMP inhibition, but retained the inhibitory activity for ADAM17. Those studies suggest that the TIMP-3 mechanism to inhibit ADAMs may be different from that elucidated for MMPs.

12 Conclusions and future prospects

A wealth of knowledge has been accumulated to show that matrixins play many roles in both biological and pathological processes. Biochemical studies of MMPs have characterized their functions and the 3D structures have provided the molecular basis for our understanding of how these multi-domain proteinases function and interact with ECM molecules and inhibitors. Structural and functional studies have also provided us with clues as to how to manipulate their enzymatic activities.

Based on those studies a large number of MMP inhibitors have been designed and synthesized and some were clinically tested for the treatment of patients with cancer or arthritis, but they showed little efficacy so far [104]. The failure of these clinical trials may be partially due to our limited knowledge of the function of MMPs in biology and pathology and the lack of selective inhibitors. As shown in Table 1, we have begun to appreciate the complexity of MMPs in terms of their biological activities, some of which seem paradoxical as they may exhibit pro- or anti-inflammatory effects, or pro- or anti-angiogenic activities. As discussed above MMP-3 and MMP-9 have protective roles in atherogenesis. Challenges in designing selective metalloproteinase inhibitors includes not only the identification of the enzymes critical in disease progression, but also the fact that there are more than 50 similar metalloproteinases in human (23 MMPs, 13 ADAMs and 19 ADAMTSs) and, hence, how to screen inhibitors for a particular enzyme or a set of enzymes. Elucidation of the inhibition mechanism of TIMP has provided opportunities to design forms that are more selective for different MMPs and ADAMs. An attraction of this approach is to locally express such inhibitors in the affected tissue using gene transfer techniques, further increasing inhibition specificity. The critical roles of non-catalytic domains of some MMPs in recognition of ECM substrates suggest that such domains may also be good targets for the development of inhibitors of the enzyme. Other possibilities may be agents that block dimerization of MT1-MMP or inhibit the unwinding activity of collagenases. For these purposes, further structural work is necessary to understand how MT1-MMP is assembled on the cell surface and interacts with the proMMP-2–TIMP-2 complex and how collagenases interact with collagens and unwind triple helical collagens. These studies will provide new insights into screening for and designing of novel types of allosteric inhibitors of matrixins for therapeutic intervention.


The work is supported by grants from the Wellcome Trust, Arthritis Research Campaign, Medical Research Council, Cancer Research UK and National Institutes of Health (AR 40994).


  • Time for primary review 12 days


  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
  28. [28]
  29. [29]
  30. [30]
  31. [31]
  32. [32]
  33. [33]
  34. [34]
  35. [35]
  36. [36]
  37. [37]
  38. [38]
  39. [39]
  40. [40]
  41. [41]
  42. [42]
  43. [43]
  44. [44]
  45. [45]
  46. [46]
  47. [47]
  48. [48]
  49. [49]
  50. [50]
  51. [51]
  52. [52]
  53. [53]
  54. [54]
  55. [55]
  56. [56]
  57. [57]
  58. [58]
  59. [59]
  60. [60]
  61. [61]
  62. [62]
  63. [63]
  64. [64]
  65. [65]
  66. [66]
  67. [67]
  68. [68]
  69. [69]
  70. [70]
  71. [71]
  72. [72]
  73. [73]
  74. [74]
  75. [75]
  76. [76]
  77. [77]
  78. [78]
  79. [79]
  80. [80]
  81. [81]
  82. [82]
  83. [83]
  84. [84]
  85. [85]
  86. [86]
  87. [87]
  88. [88]
  89. [89]
  90. [90]
  91. [91]
  92. [92]
  93. [93]
  94. [94]
  95. [95]
  96. [96]
  97. [97]
  98. [98]
  99. [99]
  100. [100]
  101. [101]
  102. [102]
  103. [103]
  104. [104]
  105. [105]
  106. [106]
  107. [107]
  108. [108]
  109. [109]
  110. [110]
  111. [111]
  112. [112]
  113. [113]
  114. [114]
  115. [115]
  116. [116]
  117. [117]
  118. [118]
View Abstract