How research peptides are actually made: solid-phase synthesis

Research peptides are built one amino acid at a time on an insoluble polymer bead, in a repeating cycle running from the sequence's C-terminal end toward its N-terminal end. The dominant method is solid phase peptide synthesis with Fmoc protecting-group chemistry — Fmoc SPPS. Each cycle strips a temporary protecting group off the growing chain's free end, couples the next residue, and washes away everything unbound. Nothing is purified until the sequence is finished and cut off the resin — which is why a certificate of analysis reads as it does: nearly every impurity class on a COA traces to a specific step in this cycle.

Why solid-phase synthesis replaced solution chemistry

Before 1963, peptides were assembled in solution, isolating and purifying an intermediate after every coupling; yields collapsed at each step and the work took months. Bruce Merrifield's contribution, published in the Journal of the American Chemical Society in 1963 (DOI: 10.1021/ja00897a025), was to covalently anchor the first amino acid to an insoluble bead, so excess reagents and soluble byproducts could be washed and filtered away rather than chromatographed. He demonstrated it on a tetrapeptide.

His original chemistry used Boc groups and hydrogen fluoride to release the finished peptide. Fmoc chemistry, developed through the 1970s and now the default for research and commercial synthesis alike, replaced that with a milder two-tier scheme (PMID: 29185032): the temporary Fmoc group comes off with a mild base, the permanent side-chain groups and resin linkage with acid. Because the two are orthogonal, the chain's reactive end can be opened hundreds of times without disturbing anything else.

The Fmoc SPPS cycle, step by step

A modern synthesizer runs this loop once per residue (PMID: 26785684).

1. Resin loading. The C-terminal amino acid is attached to a polystyrene or PEG-based resin through a cleavable linker, which determines the finished C-terminus — a Wang-type linker gives a free acid, a Rink amide linker an amide. Loading, measured in mmol per gram of resin, is a real variable: loading too heavily crowds the chains and encourages aggregation later.
2. Fmoc deprotection. Piperidine in DMF removes the Fmoc group from the chain's free amine, exposing it for the next coupling. The reaction releases dibenzofulvene, which the piperidine traps; the adduct absorbs UV near 300 nm, so instruments can monitor deprotection in real time and flag an incomplete cycle.
3. Coupling. The next Fmoc-protected amino acid is activated — commonly with a carbodiimide plus an additive such as Oxyma or HOBt, or with an onium salt such as HBTU or HATU — and its activated carboxyl is attacked by the free amine on the resin, forming the amide bond. Sterically hindered residues are frequently double-coupled to push the reaction toward completion.
4. Capping. Any amine that failed to couple is acetylated with acetic anhydride, permanently terminating that chain — a deliberate trade. A capped chain becomes a truncation: a short, acetylated fragment chemically very different from the target and easy to separate. An uncapped failed chain rejoins the next cycle and becomes a deletion sequence, nearly the same length and mass as the target, and far harder to remove.
5. Repeat. Steps 2 through 4 run once per residue, with washes between each.
6. Cleavage and global deprotection. Once the sequence is complete, a trifluoroacetic acid cocktail with scavengers such as triisopropylsilane and water severs the linker and strips every side-chain protecting group at once. The scavengers trap reactive carbocations that would otherwise re-attach to sensitive residues. The crude peptide is then precipitated out of the acid, typically into cold ether.

Why every added residue makes the synthesis harder

Yield compounds multiplicatively. If each coupling runs at 99.5% efficiency, an n-residue peptide needs n−1 couplings and theoretical crude purity is 0.995 raised to that power. The arithmetic is unforgiving.

Sequence length	Couplings	Crude purity at 99.5%/step	Crude purity at 99.9%/step
4 residues	3	98.5%	99.7%
7 residues	6	97.0%	99.4%
16 residues	15	92.8%	98.5%
30 residues	29	86.5%	97.1%
44 residues	43	80.6%	95.8%

Those figures are optimistic: they count only incomplete couplings. Real crude purity is lower and falls off faster, because coupling efficiency is not constant. Past roughly fifteen residues, chains on neighbouring resin sites begin forming β-sheet-like hydrogen-bonded structures that physically bury the reactive amine. These are the so-called difficult sequences, and mitigating them takes structural tricks — pseudoproline dipeptides, backbone amide protection with Dmb or Hmb groups, elevated-temperature coupling — rather than more reagent (PMID: 26785684).

Short sequences are therefore the easiest things to make well. Epitalon is a tetrapeptide, three couplings end to end; Semax and Selank are heptapeptides, six couplings each. No amount of downstream purification fully erases the gap between those and a 44-residue chain.

Every impurity class on a COA traces to a specific step

An HPLC chromatogram or mass spectrum records which steps ran cleanly and which did not.

Impurity class	Step it comes from	How it appears in analysis
Deletion sequence	Incomplete coupling, chain left uncapped and re-entering the next cycle	HPLC peak close to target; MS shows target mass minus one residue
Truncated sequence	Incomplete coupling followed by acetyl capping	Well-separated HPLC peak; MS shows a short fragment with an extra ~42 Da acetyl
Aspartimide and its isomers	Repeated piperidine exposure at an Asp-X site	A −18 Da species plus extra peaks at the exact target mass
Residual side-chain protection	Cleavage stopped early, or scavengers exhausted	MS shows target plus ~56 Da (tBu), 242 Da (Trt) or 252 Da (Pbf)
Oxidation products	Met, Cys or Trp exposed to air or oxidant	+16 Da per oxygen; usually an earlier-eluting shoulder on reversed-phase HPLC
Diastereomers (epimers)	Racemization during carboxyl activation, worst at Cys and His	Identical mass, shifted retention time — invisible to MS, resolved only by HPLC
Residual TFA, solvents, scavengers	Cleavage cocktail and precipitation	UV-silent; needs separate residual solvent and counterion testing

That table is also why a certificate has two halves. HPLC quantifies how much of the sample is target compound; mass spectrometry establishes what the molecule is. Neither alone is sufficient: the table contains failure modes invisible to each technique individually.

Aspartimide formation: the failure mode a mass spectrum can miss

Aspartimide is the clearest case of an impurity a purity number alone will not flag. Where a sequence contains aspartic acid, the protected side-chain carboxyl can be attacked by the backbone nitrogen of the next residue, closing into a five-membered succinimide ring and expelling water. The base used for Fmoc removal drives this, so risk accumulates with every remaining cycle — an Asp near the C-terminus of a long sequence sits in piperidine dozens of times.

The ring then reopens, and not cleanly. Hydrolysis produces a mixture of the normal α-aspartyl peptide and a rearranged β-aspartyl peptide, and ring-opening also racemizes the aspartyl centre, generating D-isomers; piperidine itself can open the ring to give piperidide adducts. The critical detail: those products are isobaric with the target — identical mass, different molecules. A mass spectrum cannot distinguish them. Only chromatography can, and only if the method was developed to resolve them. Susceptibility is sequence-dependent; Asp-Gly is the classic worst case, glycine offering no side chain to block the cyclization (PMID: 40857621). Epitalon's sequence is Ala-Glu-Asp-Gly, placing that motif at its C-terminal end; cyclic sequences containing aspartate, such as PT-141, carry the same liability alongside the extra chemistry their lactam bridge requires.

From cleavage to a finished vial

Crude peptide out of the cleavage cocktail is not a product. It is redissolved and run through preparative reversed-phase HPLC, collecting the target as a fraction and diverting everything in the impurity table to waste. Fractions are re-analysed, pooled, then frozen and dried under vacuum — lyophilization, the step that produces the dry cake in a vial. Analytical HPLC, mass spectrometry, residual solvent testing and water content are run on the finished lot, not the crude — the only order that describes what actually ships.

Why sequence length shows up in the price

Four costs stack as sequences get longer. More couplings mean more reagent and instrument time per gram. Difficult sequences require double couplings, specialty building blocks and slower cycles. Lower crude purity means preparative HPLC discards more material, so recovery falls. And a larger share of lots fails release specification and never ships — the economics behind a 99% purity floor rather than the more common 95%. Pricing that treats a tetrapeptide and a 44-residue chain alike says something about the process behind it.

Every step in this cycle has a known failure mode with a known analytical signature. The practical payoff: an impurity peak stops being an anonymous number and becomes evidence about how a specific lot was made. For research compounds, where identity and purity are the entire basis of an experiment, that distinction separates a documented lot from an assumed one. Merit publishes the full chromatogram and mass spectrum for every lot rather than a headline purity figure, so the certificate for any given lot can be checked against the failure modes its sequence is prone to.