If you’ve run a western blot for MMP-9 (matrix metalloproteinase-9, gelatinase B) on tumor lysate, conditioned media, or inflamed tissue, you’ve almost certainly seen it: two prominent bands, one around 92 kDa and another around 82 kDa. The first time it happens, it looks like a mistake. It isn’t. The two bands correspond to two biologically distinct forms of the same enzyme, and knowing which is which is often the whole point of the experiment.
This post walks through what those two bands are, how to distinguish them cleanly, and when to reach for gelatin zymography instead of straight western blot.
The short answer
- ~92 kDa band — proMMP-9 (also called the zymogen or latent form). The enzyme with its prodomain still attached.
- ~82–83 kDa band — active MMP-9. The prodomain has been cleaved off, exposing the catalytic zinc.
Both are “real MMP-9.” The difference is enzymatically massive: proMMP-9 is inactive, active MMP-9 cleaves gelatin, collagen IV, elastin, and dozens of other substrates. Depending on your question, one, the other, or both may be your signal of interest.
Why MMP-9 is secreted as a zymogen
Matrix metalloproteinases are among the most tightly regulated proteolytic enzymes in biology. Cells that make them — neutrophils, macrophages, endothelial cells, invading tumor cells — can’t afford to release a fully active protease into their own extracellular matrix. So MMP-9 is synthesized with a 90-amino-acid prodomain (residues 20–107) whose cysteine residue coordinates the catalytic zinc in the active site. This “cysteine switch” keeps the enzyme latent as long as the prodomain is attached.
Activation happens by proteolytic cleavage — usually by MMP-3 (stromelysin-1), plasmin, or trypsin — which removes the prodomain and drops the apparent molecular weight from 92 kDa to about 82–83 kDa. In many samples both forms coexist because activation is a slow, regulated process; in fresh tumor lysate you might see 80% proMMP-9 and 20% active. In neutrophil-derived conditioned media stimulated with PMA, you can see the ratio flip within an hour.
How to tell the two bands apart on your blot
The molecular-weight difference is roughly 10 kDa — big enough to resolve cleanly on a standard 8–10% SDS-PAGE gel. A few practical checks:
1. Run a proper ladder alongside
Prestained standards are fine for most experiments, but if you’re trying to publish the identity of a specific band, run an unstained standard with well-characterized MW markers (75, 100 kDa) and compare band position quantitatively. A well-resolved gel can distinguish 82 kDa from 92 kDa unambiguously.
2. Use a recombinant positive control
The cleanest interpretation comes from running your sample alongside recombinant human MMP-9 protein at the two forms. Full-length proMMP-9 runs at 92 kDa; APMA-activated MMP-9 (aminophenylmercuric acetate treatment converts proMMP-9 to active MMP-9 in vitro) runs at 82 kDa. Loading both as controls next to your unknown lets you assign band identity by co-migration.
3. Confirm with an antibody that recognizes both forms
The TPB anti-MMP-9 rabbit polyclonal is raised against a C-terminal epitope conserved in both proMMP-9 and active MMP-9, so it detects both bands with roughly equal efficiency. Antibodies raised against the prodomain will preferentially detect the 92 kDa form only — useful for a specificity control but not for full quantitation.
Common experimental artifacts (that aren’t two forms)
Not every doublet is proMMP-9 / active MMP-9. A few artifacts to rule out first:
- Glycoform doublet. MMP-9 is N-glycosylated at Asn-38 and Asn-127. Under-glycosylated forms can run 3–5 kDa faster than the fully glycosylated 92 kDa form. If you see a “doublet” at 92/88 kDa (only 4 kDa apart), it’s more likely glycoform variation than activation.
- MMP-9 / lipocalin-2 (NGAL) complex. Neutrophils secrete MMP-9 covalently linked to NGAL. Under non-reducing conditions this complex runs at 130–135 kDa. Under reducing conditions the complex resolves back to 92 kDa proMMP-9. If your 92 kDa band is unexpectedly weak in reduced samples, check whether you’re running non-reducing conditions.
- Serum contamination. If you’re working with serum-containing samples, watch for cross-reactivity with alpha-2-macroglobulin (~180 kDa) or MMP-9 sequestered in it. Use serum-free conditioned media where possible.
- Non-specific bands. If your suspicious band shows up in negative controls (MMP-9-null tissue, MMP-9-blocked antibody), it’s not MMP-9. See the non-specific bands troubleshooting guide.
When you should use zymography instead
Western blot tells you MMP-9 is present. It doesn’t tell you whether MMP-9 is active. Even the 82 kDa band is not proof of enzymatic activity — a cleaved but denatured MMP-9 will still migrate at 82 kDa on an SDS-PAGE gel.
If your experimental question is “is MMP-9 activity elevated in this sample,” the right assay is gelatin zymography. In-gel gelatin substrate, non-reducing sample buffer, renaturing wash, and overnight development with calcium in the incubation buffer gives you clear zones of clearing that correspond to gelatinolytic activity. The 92 kDa (proMMP-9) band is often detectable on zymography because SDS partially activates the zymogen during electrophoresis, but the intensity ratio to the 82 kDa active form gives you a real-time snapshot of the biology.
Zymography and western blot are complementary: run both on the same samples to get activation state (zymography) alongside protein abundance (western blot). We cover the reagent selection for both in the metalloproteinase western blot protocol.
The role of TIMPs in what you see
MMP-9 is regulated post-translationally by tissue inhibitors of metalloproteinases — primarily TIMP-1. TIMP-1 binds proMMP-9 in a 1:1 stoichiometric complex and also inhibits active MMP-9. In many biological contexts (cancer stroma, coagulation, wound healing), the MMP-9 : TIMP-1 ratio matters more than either enzyme’s absolute abundance.
If you’re studying MMP-9 in a disease context, consider running TIMP-1 and TIMP-2 antibodies on parallel blots. A high MMP-9:TIMP-1 ratio predicts higher net proteolytic activity, and the ratio often changes long before absolute MMP-9 levels do. This is one of the key readouts in the cancer MMP/TIMP panel.
Positive controls we run in-house
Every TPB anti-MMP-9 lot is validated on:
- HT1080 human fibrosarcoma cell lysate (constitutively high MMP-9 expression)
- Conditioned media from PMA-stimulated THP-1 cells (macrophage-model activation)
- Recombinant proMMP-9 spike-in (loading confirmation)
- MMP-9 knockdown or knockout HT1080 lysate (specificity control) — see the CRISPR knockout protocol
Both the 92 kDa and 82 kDa bands are visible on all three positive controls. On the KO lysate, both bands disappear entirely — the specificity check that confirms the antibody isn’t recognizing a related MMP family member. See our validation approach for the full protocol.
Related targets you might also want to blot
MMP-9 is almost always studied alongside its close relative MMP-2 (gelatinase A, 72 kDa). If your samples are complex, consider the MMP-9 vs MMP-2 selection guide for antibody cross-reactivity notes, and browse the full metalloproteinase antibody catalog for related family members (MMP-1, MMP-3, MMP-13, ADAMs).
Bottom line
Two bands on an MMP-9 western blot is the normal, expected result: 92 kDa is the latent zymogen, 82 kDa is the active enzyme, and the ratio of the two tells you where the sample sits on the activation axis. Pair with zymography if activity matters, run a recombinant control alongside if band identity matters, and think about TIMP-1 if net proteolysis matters.
Questions about MMP-9 protocol design, band identity in a specific sample, or activation-state readouts? Our lab team is happy to help — reach out through the contact form or start a buying guide walkthrough.