Identity for GDF-8 rests on orthogonal evidence, not a single number. Intact mass by LC-MS fixes the molecular weight of the construct, peptide mapping confirms sequence, and non-reducing analysis shows the disulfide-linked dimer. Purity by RP-HPLC area percent describes how much of one thing is present, never which thing.
- GDF-8 is myostatin, encoded by MSTN as a 375-residue precursor whose C-terminal 109-residue mature domain forms a disulfide-linked homodimer.
- A purity percentage measures chromatographic homogeneity and carries no information about molecular identity.
- Intact mass by LC-MS should be reported under both reducing and non-reducing conditions, since the dimer and monomer answer different questions.
- Peptide mapping with tandem MS, reported with enzyme and sequence coverage, is the primary identity test for a protein of this size.
- Isobaric substitutions and scrambled disulfide connectivity are invisible to intact mass measurement.
- No recombinant GDF-8 product is approved as a medicine by any major regulator as of 26 August 2026.
First question: which molecule does the label mean
GDF-8 is myostatin, the product of the MSTN gene and a member of the TGF-β superfamily. In humans the gene encodes a 375-residue precursor. That precursor carries a signal peptide, a large N-terminal propeptide, and a C-terminal mature domain of 109 residues released by furin-type cleavage at a basic recognition site. The mature domain assembles as a disulfide-linked homodimer built around a cystine knot, and after cleavage the propeptide stays associated non-covalently, holding the dimer in a latent complex until that association is disrupted.
All of which means the string "GDF-8" on a vial label is under-specified. It could describe the mature homodimer, a reduced monomer, the propeptide alone, the latent complex, an N-terminally extended construct carrying an initiator methionine or an affinity tag, or a short synthetic fragment of the sequence that happens to share the name. These are different molecules with different masses, different chromatographic behaviour and different reference materials. An identity test that does not state which one it examined has not established identity.
Expression system matters for the same reason. Material produced in E. coli is non-glycosylated and has usually been refolded from inclusion bodies; material from CHO or HEK cells may carry glycans and will show mass heterogeneity that an E. coli product does not. A certificate that reports a single sharp intact mass for a mammalian-expressed glycoprotein is describing something other than what it claims to describe.
The recurring failure in this corner of the supply chain is a synthetic-peptide document set applied to a protein. RP-HPLC area percent plus a single ESI-MS spectrum is a reasonable package for a 30-residue synthetic sequence. For a 12 kDa disulfide-bonded dimer it leaves the folding state, the oligomeric state and the sequence itself unaddressed.
A purity percentage is not an identity claim
The number most buyers look at first is the least informative one on the page. "98.2% by HPLC" means that 98.2 percent of the ultraviolet absorbance integrated across the chromatogram fell under the main peak, at whatever wavelength, gradient, column and load the operator chose. It is a statement about homogeneity. It says nothing at all about what the main peak contains.
Several things follow from that. Detection at 214 nm responds to the peptide bond and sees essentially everything peptidic; detection at 280 nm responds to tryptophan and tyrosine and will under-report species poor in aromatics. Neither sees counterions, residual salts, water, or organic residues from synthesis or purification. A lyophilizate that is 98 percent pure by area can be a third trifluoroacetate and water by mass, and both figures are true simultaneously.
ICH Q6B treats identity and purity as separate specification categories for exactly this reason, and expects an identity test to be sufficiently specific to distinguish the product from closely related molecules. Area percent cannot do that. Two proteins of similar hydrophobicity co-elute routinely. A scrambled-disulfide isomer of the correct sequence may elute as a shoulder, or may not resolve at all under a standard gradient.
Treat the purity figure as one row of a table rather than the headline. The rows above it are the ones carrying the identity argument.
Intact mass by LC-MS: what it fixes and what it leaves open
Electrospray ionisation of an intact protein produces a charge envelope, a series of multiply protonated ions across a range of m/z values. Deconvolution converts that envelope into a neutral mass. For a 12 kDa protein on a modern high-resolution instrument, agreement within one or two daltons of the calculated average mass is a reasonable expectation; quoting parts-per-million accuracy figures borrowed from small-molecule work is optimistic at this size and with this kind of sample handling.
The reduced and non-reduced measurements answer different questions and both belong on the report. Reducing the sample collapses the dimer to monomer and adds two hydrogens per disulfide broken. Running the sample intact should show the covalent dimer.
The arithmetic for a mature human GDF-8 construct, with the caveat that the exact figure depends on the construct and should be recalculated from the supplier's stated sequence:
reduced monomer, calculated average mass ≈ 12,400 Da
dimer = (2 × 12,400) − (5 disulfides × 2 H) ≈ 24,790 Da
Four intrachain disulfides in the cystine knot plus one interchain bond gives five per dimer, so ten hydrogen atoms are lost relative to twice the reduced monomer. Seeing a mass near 24,790 under non-reducing conditions and near 12,400 after reduction is decent evidence that the covalent architecture is what it should be.
What intact mass cannot resolve is the part people forget. Leucine and isoleucine are isobaric. Aspartate and isoaspartate are isobaric. Any permutation of the same residues has the same mass, and so does any disulfide connectivity built from the same cysteines. A completely mis-paired, functionally dead cystine knot weighs exactly what the correct one weighs. Intact mass narrows the field. It does not close it.
Peptide mapping is where sequence is actually confirmed
Peptide mapping is the workhorse identity test for a protein of this size and is described as such in USP General Chapter <1055>. The protein is denatured, usually reduced and alkylated, then digested with a specific protease. Trypsin, cleaving C-terminal to lysine and arginine, is the default; Glu-C or chymotrypsin is used as a second enzyme where trypsin leaves gaps. The resulting fragments are separated by RP-HPLC and identified by tandem mass spectrometry.
A useful map report tells you four things: which enzyme or enzymes were used, the percentage sequence coverage achieved, whether both termini were observed, and whether the map was compared against a digest of a qualified reference standard run in the same sequence. Coverage in the high nineties with observed N- and C-terminal peptides is a strong identity result. Coverage of sixty percent with unobserved termini leaves the possibility of a truncation or an unexpected extension sitting exactly in the blind spot.
Terminal confirmation deserves separate attention for GDF-8 because the mature domain is defined by a proteolytic cleavage. If the construct was expressed as a precursor and processed, incomplete or ragged processing shifts the N-terminus. Edman degradation on the first five to ten residues is still used for this and gives an orthogonal read independent of the mass spectrometer. For the mature human sequence the first residues run Asp-Phe-Gly-Leu-Asp, which is a short and unambiguous fingerprint.
Disulfide mapping is the harder variant and is worth asking for when folding is in question. Digestion under non-reducing conditions, at a pH low enough to suppress thiol-disulfide exchange, preserves the bonds and lets the linked peptide pairs be identified by mass. This is the only routine method that distinguishes correct connectivity from scrambled connectivity. Most research-grade certificates do not include it.
Orthogonal methods and what each one is entitled to conclude
| Method | What it measures | Identity evidence? | Common limitation |
|---|---|---|---|
| RP-HPLC, area percent | Chromatographic homogeneity | No, on its own | Blind to co-eluting species and to non-absorbing content |
| RP-HPLC, retention time vs reference | Comparative behaviour | Weak, supporting only | Retention shifts with column age, temperature, gradient |
| Intact LC-MS, reduced | Monomer molecular weight | Yes, partial | Isobaric substitutions and permutations invisible |
| Intact LC-MS, non-reduced | Covalent dimer mass | Yes, partial | Says nothing about which cysteines pair |
| Peptide mapping with MS/MS | Sequence, coverage, termini | Yes, primary | Gaps in coverage hide local defects |
| Non-reducing SDS-PAGE | Apparent size of covalent species | Supporting | Cystine-knot proteins migrate anomalously |
| Size-exclusion HPLC | Dimer, monomer, aggregate distribution | Supporting | Column chemistry can adsorb or dissociate species |
| Cell-based reporter assay | Receptor engagement and downstream signalling | Functional, not structural | Cell-line and passage dependent; wide confidence intervals |
Two entries in that table deserve a note. SDS-PAGE apparent molecular weight is routinely off for cystine-knot proteins, and a band running at 26 or 28 kDa rather than 25 is not by itself a finding. It is a shape effect. Reading a gel as if it were a mass measurement is a habit worth breaking.
The bioassay is the only method listed that reports on folding. A reporter line carrying a SMAD-responsive luciferase construct responds to receptor engagement, and correctly folded material produces a dose-response curve while scrambled material does not. Mass spectrometry cannot make that distinction. If the intended work depends on the protein being active rather than merely present, a functional result on the specific lot is the evidence that matters, and its absence should be treated as an open question rather than an assumption.
Reading the certificate line by line
| Field | Adequate | Red flag |
|---|---|---|
| Product description | Species, residue range, expression host, tag status | "GDF-8 (Myostatin)" and nothing else |
| Lot number | Unique, appears on vial and document | Absent, or a generic catalogue number reused |
| Test dates | Each method dated, after the fill date | Undated, or predating manufacture |
| Intact mass | Observed and calculated, reduced and non-reduced | A single figure with no conditions stated |
| Peptide map | Enzyme, coverage percentage, termini observed | "Confirmed by MS" with no data |
| Purity | Method, wavelength, gradient, area percent | A bare percentage |
| Oligomeric state | SEC or non-reducing gel result | Not addressed |
| Endotoxin | EU/mg with method named | Omitted for a bacterially expressed protein |
| Residual moisture and counterion | Reported as percent by mass | Omitted, leaving net protein content unknown |
| Signature | Named analyst or QC release | Unsigned template |
Attached raw data separates a real certificate from a formatted assertion. Ask for the deconvoluted mass spectrum, the annotated peptide map and the chromatogram with its axes legible. A supplier who runs the tests has these files sitting in a folder. A supplier who buys documentation with the material will find reasons not to send them, and the reasons are usually about confidentiality.
One more field is worth requesting even though it rarely appears: net peptide or net protein content. Between counterion, residual water and any excipient, the mass in the vial is not the mass of protein in the vial. If a protocol depends on knowing the actual quantity, that number has to come from amino acid analysis or a quantitative UV measurement, not from the label weight. Our quality standard page sets out which of these tests we require per lot and which are run on a periodic basis.
Handling the material without invalidating the identity you paid for
A verified certificate describes the vial as it left the analytical laboratory. Everything after that is on the receiving lab. For a disulfide-bonded, refolded protein the two practical risks are aggregation and thiol-disulfide exchange, and both are accelerated by the same things: elevated temperature, agitation that whips air into solution, and alkaline pH in the reconstitution buffer.
Free thiols catalyse disulfide rearrangement, and a protein with five disulfide bonds and any residual free cysteine content has the substrate for it. This is one reason reducing agents do not belong anywhere near a working stock intended to stay in its native fold, and one reason buffers are chosen at neutral or slightly acidic pH rather than for convenience. Carrier protein, where a supplier includes it, is there to suppress surface adsorption at low concentrations, and its presence or absence changes how the material behaves in dilute solution.
Preparation arithmetic follows the same rules as any other lyophilizate. A 100 µg vial brought into 1 mL of diluent gives 100 µg ÷ 1 mL = 100 µg/mL, and the vial concentration calculator handles other combinations. The choice of diluent should be whatever the supplier's reconstitution note specifies; general background on aqueous diluents is in the bacteriostatic water guide. The broader storage argument, including why moisture rather than freezer temperature is the dominant variable for a dry solid, is set out in the storage and stability guide and applies here without modification.
Record the lot identifier against every prepared solution. When a result looks wrong, the first thing anyone will want to know is whether the material was what the protocol assumed, and a chain of records back to the certificate answers that faster than repeating the experiment.
Regulatory position
Myostatin was described by McPherron, Lawler and Lee in 1997 as a TGF-β superfamily member that negatively regulates skeletal muscle mass in mice. It has since been a target for pharmaceutical development programmes, several of which have been discontinued. No product consisting of recombinant GDF-8 protein is approved by the FDA, EMA or any comparable authority for human use, and material supplied for laboratory work carries no such authorisation.
Agents acting on this pathway also appear on the World Anti-Doping Agency Prohibited List under hormone and metabolic modulators, which is context for why the compound attracts interest outside research settings. That interest does not change what the material is or what its documentation covers.
Status verified 26 August 2026.
FOR LABORATORY AND IN-VITRO RESEARCH USE ONLY. NOT FOR HUMAN OR ANIMAL CONSUMPTION. NOT FOR PERSONAL, MEDICAL, DIAGNOSTIC, THERAPEUTIC, OR RECREATIONAL USE.
What is GDF-8 studied for?
Published research on GDF-8 investigates the areas below — which is a different question from what GDF-8 will do for anyone, a claim about a living system that nothing on this site is sold for.
What it is. Growth/differentiation factor 8, also called myostatin — a member of the TGF-beta superfamily.
What the research looks at. Extensively characterised as a target for inhibition in muscle-wasting conditions. Note the direction: the research interest is in blocking GDF-8, not in supplying it.
How it is thought to work. A secreted negative regulator of skeletal muscle growth, expressed specifically in developing and adult skeletal muscle, signalling through activin type II receptors.
What is not established. No approved product. Every therapeutic programme in this area has aimed at inhibiting this protein rather than administering it, and inhibitor programmes have repeatedly stopped for safety.
The full record, including the certificate for the lot in stock, is on the GDF-8 product page.
Common questions
Is an ESI-MS spectrum enough to confirm GDF-8 identity?
Why does the certificate list a mass around 25 kDa when the sequence is 109 residues?
What does 98% purity actually guarantee?
Can mass spectrometry tell me whether the protein is correctly folded?
Does expression host change what identity testing should look like?
What should I ask a supplier for beyond the certificate?
Sources
- McPherron, Lawler and Lee, Nature, 1997. Original description of GDF-8/myostatin as a TGF-β superfamily member regulating skeletal muscle mass in mice; supports the identification of the gene product and its family placement.
- UniProt Knowledgebase entry for human myostatin (MSTN). Curated precursor length, signal peptide and propeptide boundaries, furin processing site and mature domain sequence; supports the residue numbering and the mature N-terminal fingerprint.
- ICH Q6B, Specifications: Test Procedures and Acceptance Criteria for Biotechnological/Biological Products. Establishes identity and purity as separate specification categories and sets the expectation that an identity test be specific enough to distinguish the product from closely related molecules.
- USP General Chapter <1055>, Biotechnology-Derived Articles — Peptide Mapping. Describes peptide mapping as an identity test, including digestion conditions, separation and comparison against a reference digest.