What Are Metalloproteases and How Do They Work?

Metalloproteases are enzymes that use a metal ion, almost always zinc, to cut other proteins apart. They are among the most widespread protein-degrading enzymes in nature, found in every kingdom of life from bacteria to plants to humans, and they participate in processes as different as embryonic development, immune defense, wound healing, and the toxicity of snake venom. In human biology alone, well over a hundred metalloproteases have been identified, and their misregulation is linked to cancer, arthritis, cardiovascular disease, and neurodegeneration. Understanding what these enzymes do, how the body controls them, and why they so often feature in disease has been a central thread in biomedical research for decades.

How a Zinc Ion Cuts a Protein

The defining feature of a metalloprotease is a metal atom at its core, positioned precisely to break a specific bond in a target protein. In the vast majority of cases that metal is zinc. The zinc ion sits in a pocket formed by conserved amino acids and coordinates a water molecule, which the enzyme converts into a potent chemical weapon. In matrix metalloproteinase-1, for example, computational studies show that the reaction begins when a nearby glutamate residue pulls a proton off the zinc-bound water, creating a hydroxide ion. That hydroxide then attacks the target bond, forming a short-lived intermediate before the bond snaps apart. The energy barrier for this step is around 22 kilocalories per mole, making it the slowest part of the reaction and effectively the speed limit for collagen breakdown.

A surprising detail is that a second water molecule participates, acting as a shuttle that moves protons into the right positions and lowers the overall energy cost by several kilocalories per mole.¹ Earlier computational work had already pointed to this helper water as important, finding it donated a hydrogen bond to the target’s carbonyl oxygen and cut the activation energy by about 5 kilocalories per mole.² More recent modeling of the same enzyme has added a wrinkle: the proton that ultimately breaks the bond may come from the zinc-bound hydroxide itself rather than from a separate solvent water, suggesting the mechanism is still being refined at the atomic level.³ The practical takeaway is that the zinc center, the glutamate base, and at least one water molecule form a remarkably efficient cutting machine, and small changes in how those pieces interact determine how fast and how selectively the enzyme works.

The Major Families

Metalloproteases are not one enzyme but a sprawling collection of families. One useful way to organize them is by shared structural features around the active site. A large group of zinc-dependent endopeptidases share a signature motif in their catalytic site and a conserved methionine residue tucked beneath the zinc, forming a structural element called the Met-turn. This shared architecture groups them into a superfamily called the metzincins, which includes four families: the astacins, the adamalysins, the serralysins, and the matrix metalloproteinases (often shortened to MMPs or matrixins).⁴ Despite low sequence similarity from one family to the next, their catalytic domains fold into remarkably similar shapes, an indication that the architecture is ancient and works well enough that evolution has kept it largely intact.⁵

Beyond the metzincins, two other families come up constantly in biomedical research. The ADAMs (a disintegrin and metalloproteinase) are membrane-anchored enzymes that specialize in clipping proteins off the cell surface, a process called ectodomain shedding. ADAM10 and ADAM17 are the best-studied members, and between them they release at least six growth-factor ligands that activate the epidermal growth factor receptor, each handled predominantly by one or the other.⁶ Then there are the ADAMTS enzymes (ADAMs with thrombospondin motifs), which are secreted rather than membrane-bound and specialize in degrading proteoglycans, the large sugar-coated proteins that give cartilage and other connective tissues their structure.⁷

Built-In Safety Switches

An enzyme that can chew through collagen, the most abundant protein in the body, is dangerous if left unregulated. So cells control metalloproteases at multiple levels. One of the most elegant controls is that many MMPs are made as inactive precursors, with a built-in plug that blocks the active site until it is deliberately removed. The mechanism is called the cysteine switch: a cysteine residue in the enzyme’s pro-domain directly binds to the catalytic zinc, physically occupying the spot where a water molecule would normally sit. Until that cysteine is pulled away, the enzyme cannot activate a water molecule and cannot cut anything.⁸

Activation happens when the cysteine–zinc bond is disrupted, typically through cleavage by another protease. Molecular simulations of MMP-9 activation by a serine protease called kallikrein revealed something unexpected: the cysteine–zinc interaction starts breaking apart even before the pro-domain is cut off, driven by proton transfers triggered by the two enzymes simply docking together.⁹ And at least one family member, MMP-26, has a variant cysteine-switch motif that appears nonfunctional, meaning it uses a completely different activation logic.¹⁰ So the cysteine switch is a strong general principle but not universal.

Once active, metalloproteases are kept in check by a family of natural inhibitors called TIMPs (tissue inhibitors of metalloproteinases). There are four human TIMPs, and they work by wedging themselves into the enzyme’s active-site cleft. Crystal structures show that TIMP-1, for example, forms an elongated wedge that spans the full length of the cleft, with its very first residue coordinating the catalytic zinc, effectively locking the enzyme shut.¹¹ The interactions are extensive enough that researchers have been able to engineer TIMP variants that selectively target certain metalloproteinases while sparing others, opening a door to therapeutic applications.¹²

Normal Roles in the Body

The reason metalloproteases exist at all is that tissues need to be remodeled constantly. The extracellular matrix, the scaffolding of collagen, fibronectin, and other structural proteins that surrounds cells, is not permanent. It gets reshaped during embryonic development, wound repair, and the formation of new blood vessels. MMPs are the primary enzymes responsible for this turnover, and they can collectively degrade almost every component of the extracellular matrix.¹³ They also do more than just demolition: by cutting matrix proteins and cell-surface receptors, they release signaling fragments that influence how cells migrate, divide, and differentiate.¹⁴

Cancer and Metalloprotease Overproduction

Cancer cells are notoriously good at co-opting metalloproteases. Tumor growth, invasion into surrounding tissue, and the establishment of blood supply all depend on breaking down the matrix barriers that normally keep cells in place. Two gelatinases, MMP-2 and MMP-9, are especially prominent in this context. They degrade basement membrane collagens and help remodel the tumor microenvironment to favor new blood vessel growth.¹⁵ In hypopharyngeal carcinoma, the expression of both MMP-2 and MMP-9 was found to increase with the degree of lymph node metastasis.¹⁶

A striking animal experiment demonstrated that MMP-2 and MMP-9 work cooperatively. Mice lacking only one of the two enzymes still developed invasive, well-vascularized tumors from malignant skin cells. But mice lacking both showed a dramatic failure of tumor vascularization and invasion, indicating the two enzymes are functionally redundant to a degree and compensate for each other when only one is missing.¹⁷ That redundancy is one reason that blocking a single MMP has been hard to translate into effective cancer therapy.

Heart Disease and Plaque Instability

Metalloproteases also contribute to cardiovascular events. In atherosclerosis, plaques build up inside artery walls, and the risk of a heart attack depends heavily on whether a plaque is stable or prone to rupture. MMPs degrade the collagen and other structural proteins that hold the plaque’s fibrous cap together, and when MMP activity outpaces the body’s TIMPs, the cap weakens and the plaque becomes vulnerable.¹⁸ Studies of human coronary arteries have found more MMP-2 and MMP-9 staining in plaques from segments that had remodeled expansively, and significantly more active MMP-2 in those segments compared with those that had remodeled constrictively.¹⁹ The hypothesis that restraining MMP activity in plaques could prevent acute cardiovascular events has driven significant therapeutic interest.²⁰

Cartilage Destruction in Osteoarthritis

In joints, the critical metalloprotease culprits belong to the ADAMTS family rather than the MMPs. Cartilage gets its resilience from aggrecan, a large proteoglycan that traps water and resists compression. ADAMTS-4 and ADAMTS-5, sometimes called aggrecanase-1 and aggrecanase-2, cleave aggrecan at specific sites, and their activity increases sharply when inflammatory signals are present.²¹ Experiments in cartilage explants showed that depleting both ADAMTS-4 and ADAMTS-5 from the tissue reduced aggrecanase activity in the culture medium by about 90 percent, and a general metalloprotease inhibitor blocked aggrecan release while an MMP-specific inhibitor did not, confirming that the aggrecanases rather than the classic MMPs are driving cartilage breakdown.²²

Of the two enzymes, ADAMTS-5 appears to be the more important target. Data from knockout mouse studies and translational work with an anti-ADAMTS-5 antibody support the idea that ADAMTS-5 is the dominant aggrecanase in cartilage degradation, and the pharmacological evidence translates to human tissues and primate models.²³ This makes ADAMTS-5 one of the more promising metalloprotease drug targets currently under investigation for a disease that still has no approved disease-modifying therapy.

The Alzheimer’s Connection

ADAM10, a membrane-anchored metalloprotease, serves as the main constitutive alpha-secretase in the brain. Its job is to cut amyloid precursor protein (APP) right in the middle of the sequence that would otherwise become amyloid-beta, the peptide that aggregates in Alzheimer’s disease. By cutting there, ADAM10 generates a neuroprotective fragment called soluble APP-alpha and prevents amyloid-beta from forming.²⁴ Early work demonstrated that increasing ADAM10 expression and activity promoted this protective cleavage pathway, and the authors suggested that boosting ADAM10 might be beneficial for Alzheimer’s treatment.²⁵ More recently, a regulatory mechanism upstream of ADAM10 was uncovered: a cell-surface enzyme called GDE2 works through the protein RECK to control ADAM10’s alpha-secretase cleavage of APP, adding another potential point of therapeutic intervention.²⁶

Why Broad-Spectrum Inhibitors Failed and What Comes Next

Given that metalloproteases show up in cancer, arthritis, and cardiovascular disease, it might seem obvious to develop drugs that block them. Researchers did exactly that in the 1990s and early 2000s, and the results were largely disappointing. Broad-spectrum MMP inhibitors that targeted the zinc-binding site suppressed many enzymes at once, and the most common side effect was musculoskeletal syndrome: joint pain, stiffness, and inflammation. Later work with more selective compounds pointed to specific MMPs that are not involved in this side effect. Testing inhibitors that spare MMP-2, MMP-9, MMP-13, and MT1-MMP showed that those enzymes were not responsible for the musculoskeletal problems, helping clarify which targets are safe and which are not.²⁷

The newer generation of inhibitors takes a different tactic: instead of jamming the zinc site that all MMPs share, they target regions unique to individual enzymes. For MMP-13, which is implicated in both arthritis and certain cancers, researchers identified inhibitors that bind to sites away from the catalytic center, blocking the enzyme through an allosteric mechanism. These compounds showed high selectivity when tested against a panel of 30 different proteases and had a favorable safety profile in preliminary testing.²⁸ The same exosite-targeting strategy has shown promise in breast and colon cancer models, where blocking the hemopexin domain of MMP-13 selectively disrupted tumor-promoting activity.²⁹

Metalloproteases in Venom and Infection

Snake venom hemorrhagic toxins are metalloproteases. Called SVMPs (snake venom metalloproteinases), they target the basement membranes of capillaries, binding to and degrading collagen and other structural proteins that hold small blood vessels together. In normal blood flow, the weakened capillary walls distend and eventually burst, causing the massive internal bleeding that characterizes certain snakebites.³⁰ The most potent hemorrhagic SVMPs belong to the P-III class, which carry extra domains that help them bind to relevant targets in the vessel wall. Studies tracing the tissue distribution of these toxins found them accumulating specifically at basement membranes, where they hydrolyzed collagen at the sites of hemorrhagic lesions.³¹

Bacteria use metalloproteases as virulence tools, too. Bacterial metalloproteases degrade host extracellular matrix components like collagen, fibrin, and laminin, helping the pathogen penetrate tissue and establish infection. Some also cleave antibodies, antimicrobial peptides, and complement proteins, directly undermining the immune response.³² A specific example is ProA, a zinc metalloprotease from Legionella pneumophila, the bacterium that causes Legionnaires’ disease. ProA degrades flagellin monomers, the protein subunits of bacterial flagella, preventing them from being detected by the host’s TLR5 immune receptor and dampening the inflammatory alarm that would otherwise recruit immune cells.³³

Inside the Membrane and the Mitochondria

Not all metalloproteases operate outside the cell or on its surface. An unusual group, the site-2 proteases, actually cut their targets within the lipid membrane itself. The catalytic zinc sits roughly 14 angstroms below the membrane surface, and the enzyme uses a gating mechanism to open and close access to its active site. Crystal structures of a bacterial site-2 protease show two conformations: a closed form where the active site is sealed off, and an open form where two transmembrane helices swing apart to let the substrate in.³⁴ These enzymes are conserved from bacteria to humans, regulating signal transduction by releasing transcription factors or other signaling proteins from their membrane anchors.³⁵ In bacteria, site-2 proteases have been linked to stress responses, spore formation, and host-pathogen interactions.³⁶

Mitochondria harbor their own metalloprotease quality-control system. Two inner-membrane proteases, YME1L and OMA1, monitor protein health inside mitochondria and regulate how the organelle fuses and divides. Under normal conditions, YME1L degrades OMA1, keeping it in check. When stress depolarizes the mitochondrial membrane and depletes cellular energy, the relationship flips: OMA1 is stabilized and YME1L is degraded instead. This reciprocal degradation adjusts how they process OPA1, a protein that controls whether mitochondria stay elongated or fragment into smaller units, directly shaping the organelle’s ability to recover after an insult.³⁷ In yeast, losing OMA1 interferes with oxidative-stress responses and alters signaling through nutrient-sensing pathways, reinforcing the idea that these mitochondrial metalloproteases are integrated into the cell’s broader stress-management network.³⁸

Metalloproteases in Plants

Plants rely on metalloproteases for something essential to life on Earth: photosynthesis. The D1 protein at the heart of photosystem II is especially vulnerable to light damage, and unless the damaged copy is removed and replaced, the whole photosynthetic complex stalls. The enzyme responsible for this cleanup is FtsH, a zinc metalloprotease embedded in the thylakoid membrane of chloroplasts. FtsH degrades photodamaged D1 protein as part of the photosystem II repair cycle and also removes unassembled proteins from the thylakoid and surrounding compartment.³⁹ Beyond simple housekeeping, FtsH plays roles in the biogenesis of thylakoid membranes and the assembly of several protein complexes in the photosynthetic electron-transport chain.⁴⁰ Plants with defective FtsH grow poorly in bright light, underscoring just how critical this single metalloprotease is to photosynthetic fitness.

Metalloprotease Fragments as Disease Biomarkers

When metalloproteases cut collagen and other matrix proteins, they leave behind fragments with unique chemical signatures. Researchers have turned those fragments into biomarkers that can be measured in blood. One such fragment, CO3-610, is generated when MMP-9 cleaves type III collagen. In a rat model of liver fibrosis, CO3-610 levels correlated with the degree of fibrosis during disease progression and rose steeply alongside total collagen, suggesting it could serve as a marker that tracks worsening liver damage without needing a biopsy.⁴¹ A complementary marker, C1M, detects a type I collagen fragment generated by MMP-2, MMP-9, and MMP-13. It was significantly elevated in rat models of liver fibrosis and correlated with total liver collagen and collagen gene expression, giving it potential as a non-invasive way to assess liver damage in humans.⁴² The appeal of this approach is that it measures what the enzymes are actually doing to tissue in real time, rather than measuring the enzymes themselves, which can be present in inactive or inhibited forms.

1
PubMed Central. Catalytic Mechanism of Collagen Hydrolysis by Zinc(II)-Dependent Matrix Metalloproteinase-1 ↩︎
2
PubMed. Catalytic mechanism of matrix metalloproteinases: two-layered ONIOM study ↩︎
3
PubMed Central. Novel Insights into the Catalytic Mechanism of Collagenolysis by Zn(II)-Dependent Matrix Metalloproteinase-1 ↩︎
4
PubMed Central. The metzincins–topological and sequential relations between the astacins, adamalysins, serralysins, and matrixins (collagenases) define a superfamily of zinc-peptidases ↩︎
5
Current Opinion in Structural Biology. Structural features of a superfamily of zinc-endopeptidases: the metzincins ↩︎
6
PubMed Central. Distinct roles for ADAM10 and ADAM17 in ectodomain shedding of six EGFR ligands ↩︎
7
PubMed. Proteoglycan degradation by the ADAMTS family of proteinases ↩︎
8
PubMed Central. The cysteine switch: a principle of regulation of metalloproteinase activity with potential applicability to the entire matrix metalloproteinase gene family ↩︎
9
PubMed. Molecular structures and dynamics of the stepwise activation mechanism of a matrix metalloproteinase zymogen: challenging the cysteine switch dogma ↩︎
10
PubMed. Unconventional activation mechanisms of MMP-26, a human matrix metalloproteinase with a unique PHCGXXD cysteine-switch motif ↩︎
11
PubMed. Mechanism of inhibition of the human matrix metalloproteinase stromelysin-1 by TIMP-1 ↩︎
12
PubMed Central. The tissue inhibitors of metalloproteinases (TIMPs): an ancient family with structural and functional diversity ↩︎
13
PubMed Central. The Roles of Matrix Metalloproteinases and Their Inhibitors in Human Diseases ↩︎
14
PubMed. Extracellular matrix remodelling: the role of matrix metalloproteinases ↩︎
15
PubMed Central. Targeting Invasion: The Role of MMP-2 and MMP-9 Inhibition in Colorectal Cancer Therapy ↩︎
16
PubMed Central. The role of MMP-2 and MMP-9 in the metastasis and development of hypopharyngeal carcinoma ↩︎
17
PubMed Central. Contribution of host MMP-2 and MMP-9 to promote tumor vascularization and invasion of malignant keratinocytes ↩︎
18
PubMed Central. Multidimensional Contribution of Matrix Metalloproteinases to Atherosclerotic Plaque Vulnerability: Multiple Mechanisms of Inhibition to Promote Stability ↩︎
19
PubMed. Atherosclerotic arterial remodeling and the localization of macrophages and matrix metalloproteases 1, 2 and 9 in the human coronary artery ↩︎
20
Circulation Research. Matrix Metalloproteinases in Vascular Remodeling and Atherogenesis ↩︎
21
PubMed. ADAMTS-4 and ADAMTS-5: key enzymes in osteoarthritis ↩︎
22
PubMed. The role of ADAM-TS4 (aggrecanase-1) and ADAM-TS5 (aggrecanase-2) in a model of cartilage degradation ↩︎
23
PubMed Central. Translational development of an ADAMTS-5 antibody for osteoarthritis disease modification ↩︎
24
PubMed. The Role of ADAM10 in Alzheimer’s Disease ↩︎
25
PubMed. Constitutive and regulated alpha-secretase cleavage of Alzheimer’s amyloid precursor protein by a disintegrin metalloprotease ↩︎
26
PubMed Central. GDE2-RECK controls ADAM10 α-secretase-mediated cleavage of amyloid precursor protein ↩︎
27
PubMed Central. The Rebirth of Matrix Metalloproteinase Inhibitors: Moving Beyond the Dogma ↩︎
28
PubMed Central. Characterization of selective exosite-binding inhibitors of matrix metalloproteinase 13 that prevent articular cartilage degradation in vitro ↩︎
29
Bioscience, Biotechnology, and Biochemistry. Functional characterization of selective exosite-binding inhibitors of matrix metalloproteinase-13 (MMP-13) ↩︎
30
PubMed. Key events in microvascular damage induced by snake venom hemorrhagic metalloproteinases ↩︎
31
PubMed Central. Mechanisms of Vascular Damage by Hemorrhagic Snake Venom Metalloproteinases: Tissue Distribution and In Situ Hydrolysis ↩︎
32
PubMed Central. Bacterial Metalloproteases in Host-Pathogen Interactions ↩︎
33
PubMed Central. Zinc Metalloprotease ProA from Legionella pneumophila Inhibits the Pro-Inflammatory Host Response by Degradation of Bacterial Flagellin ↩︎
34
PubMed. Structure of a site-2 protease family intramembrane metalloprotease ↩︎
35
PubMed Central. Mechanistic insights into intramembrane proteolysis by E. coli site-2 protease homolog RseP ↩︎
36
PubMed. Site-2 proteases in prokaryotes: regulated intramembrane proteolysis expands to microbial pathogenesis ↩︎
37
PubMed Central. Reciprocal Degradation of YME1L and OMA1 Adapts Mitochondrial Proteolytic Activity during Stress ↩︎
38
PubMed Central. Oma1 Links Mitochondrial Protein Quality Control and TOR Signaling To Modulate Physiological Plasticity and Cellular Stress Responses ↩︎
39
Physiologia Plantarum. FtsH proteases in chloroplasts and cyanobacteria ↩︎
40
PubMed Central. FtsH Protease in the Thylakoid Membrane: Physiological Functions and the Regulation of Protease Activity ↩︎
41
PubMed Central. Measurement of CO3-610, a potential liver biomarker derived from matrix metalloproteinase-9 degradation of collagen type iii, in a rat model of reversible carbon-tetrachloride-induced fibrosis ↩︎
42
PubMed. A novel marker for assessment of liver matrix remodeling: an enzyme-linked immunosorbent assay (ELISA) detecting a MMP generated type I collagen neo-epitope (C1M) ↩︎