How Peptides Are Made: Solid-Phase Synthesis, Cleavage, Purification, Lyophilisation and QC

Nearly every research peptide is made the same way: assembled on a solid resin by Fmoc chemistry, cut free with trifluoroacetic acid, purified by preparative HPLC, freeze-dried and tested. Each step leaves its own signature in the vial. Knowing the sequence of operations explains why peptides carry the impurities they do, why they usually arrive as trifluoroacetate salts, and why a 40-residue peptide costs more than a five-residue one.

Solid-phase peptide synthesis

Solid-phase peptide synthesis (SPPS) was introduced by Bruce Merrifield, and the field has since moved from the original Boc/benzyl protection strategy to the Fmoc/tert-butyl strategy that now dominates ^[1]. Fmoc SPPS is the method of choice today, and because therapeutic peptides are made by it at multi-tonne scale, high-quality Fmoc amino acid building blocks are cheap and widely available ^[2]. It is automatable and scalable, which is why many marketed peptide drugs are manufactured this way ^[3].

The principle is simple. The C-terminal amino acid is anchored through a linker to an insoluble polymer bead, and the chain is built towards the N-terminus. Because the growing peptide stays attached to the bead, excess reagents and by-products are removed by washing and filtration, so each step can be driven with a large excess of reagent. The linker decides the C-terminus: acid-type linkers give a free carboxylic acid, amide-type linkers a C-terminal amide.

Protection is orthogonal. Each incoming amino acid carries a temporary Fmoc group on its α-amine, removed by base before the next coupling. Reactive side chains carry permanent, acid-labile groups (tert-butyl ethers and esters, trityl, Boc, Pbf for arginine) that survive every cycle and come off only at the final acid cleavage. A widely used laboratory method of this kind, built on aminium-type coupling reagents and PEG-modified polystyrene resins, produced the 41-residue corticotropin-releasing factor in about 80 working hours ^[4].

The synthesis cycle

1. Deprotection. A base, classically piperidine in DMF, removes the Fmoc group and exposes the amine.
2. Wash.
3. Coupling. The next Fmoc amino acid is activated (with a carbodiimide plus an additive, or an aminium or phosphonium reagent) and forms the peptide bond.
4. Wash, and optionally cap. Acetylating any unreacted amines turns would-be deletion sequences into shorter, N-acetylated truncations that are easier to separate later.

The cycle repeats once per residue after the first. Incomplete steps accumulate multiplicatively, which is the arithmetic behind length-dependent cost and crude purity:

These figures are calculated, not measured, and ignore every other side reaction; they show why crude purity falls as sequences lengthen and why long peptides need more purification.

Side reactions during assembly

Most impurities in a finished peptide trace back to a specific chemical event during chain assembly.

• Deletions and insertions. Inefficient Fmoc removal leaves an amine blocked for a cycle, so a residue goes missing; excess amino acid reagents can add an extra one ^[5].
• Aggregation. As the chain grows, it can fold and stick to neighbouring chains on the resin, starving the N-terminus of reagent. This aggregation is the main cause of failed syntheses of "difficult sequences" ^[4]. Pseudoprolines, temporary ring-locked derivatives of serine, threonine or cysteine, disrupt the secondary structure responsible ^[6].
• Aspartimide formation. Base-mediated cyclisation at aspartic acid remains a persistent challenge in Fmoc chemistry, tackled with modified side-chain esters, non-ester masking groups, backbone protection or alternatives to Fmoc removable without base ^[7]. Adding HOBt to the deprotection solution or using piperazine instead of piperidine also reduced it ^[8].
• Racemisation. In microwave-assisted synthesis of a 20-residue model peptide, cysteine, histidine and aspartic acid were susceptible; lowering the coupling temperature from 80 °C to 50 °C limited histidine and cysteine racemisation, and the hindered base collidine minimised D-cysteine ^[8]. Each D-residue creates a diastereomer with exactly the target's mass.
• Diketopiperazine formation. In process work on tirzepatide, Eli Lilly chemists traced double-amino-acid deletion impurities to diketopiperazine formation during Fmoc removal and while Fmoc-Pro-Pro-Ser-resin intermediates stood after coupling; intermediates with a penultimate proline were especially prone, and Fmoc removal proceeded even in solvent alone ^[9]. Switching to Bsmoc-protected amino acids eliminated the by-product, while Oxyma additives and colder storage slowed it ^[9]. Proline-rich sequences such as BPC-157, whose Pro-Pro-Pro run produces several penultimate-proline intermediates during assembly, contain the same structural feature.

Cleavage and global deprotection

When the chain is complete, a concentrated TFA cocktail cuts the peptide from the linker and removes the side-chain protecting groups in one step. The protecting groups leave as reactive carbocations that can attach to electron-rich side chains. In a classic comparison of scavenger mixtures, Trp, Tyr, Met and Cys were the residues most affected, and Reagent K (82.5% TFA, 5% phenol, 5% water, 5% thioanisole, 2.5% ethanedithiol) suppressed the widest range of side reactions across ten peptides of 20–50 residues ^[10].

Methionine remains awkward. Oxidation to methionine sulfoxide and S-tert-butylation both occur during TFA cleavage; newer cocktails based on trimethylsilyl chloride and dimethyl sulfide eliminated the oxidation and reduced alkylation, and the alkylated sulfonium by-product could be reversed by warming in 5% acetic acid at 40 °C for 24 hours ^[11]. Protecting groups that survive cleavage leave adducts, one of the synthesis-related impurity classes catalogued for peptide medicines ^[5]. The crude peptide is then precipitated, typically with cold ether, and dried. It is now a trifluoroacetic acid salt.

Purification by preparative HPLC

No production method, chemical or biological, yields the target alone, and preparative liquid chromatography is the main way peptides are purified; conditions are chosen case by case, and continuous multicolumn processes are an emerging option ^[12]. In reversed-phase purification, fractions are collected across the main peak, analysed, and pooled if they meet specification. Tighter pooling gives higher purity at lower yield, which is one reason a 99% specification costs more than 95%.

Two consequences carry through to the vial. Impurities with near-identical sequences that were not resolved in the preparative run can co-elute in the analytical QC run as well ^[13], so the most persistent impurities are the hardest to measure. And because purification uses TFA as the ion-pairing reagent, the peptide is recovered as a TFA salt unless a counter-ion exchange follows ^[14]; TFA vs acetate peptide salts covers that step.

Lyophilisation

Pooled fractions are freeze-dried. The solution is frozen, ice is removed by sublimation under vacuum (primary drying), and bound water is driven off at slightly higher temperature (secondary drying). Successful freeze-drying depends on the thermochemical and thermomechanical properties of the amorphous solid that forms, which is why formulation and process design go together ^[15]. Residual water matters afterwards too: in freeze-dried amorphous peptides and proteins, deamidation rates rise with water content ^[16]. The lyophilisation guide and the lyophilisation glossary entry cover the physics and what a good cake looks like.

Quality control and release

The finished batch is tested before release: HPLC for purity and mass spectrometry for identity at a minimum, with amino acid analysis, counter-ion and water content where content matters. Process choices show up in the result. When analysts at the originator company compared liraglutide made synthetically or recombinantly by five suppliers, each had an impurity profile distinct from the originator's, and manufacturing scale and process strongly affected physical stability; trace metals increased high-molecular-weight species in liraglutide and semaglutide ^[17]. The authors' employer makes the originator products, but the point is general: the same sequence made by different routes is not the same material.

Beyond linear SPPS

Very long sequences are often made by joining shorter synthetic segments through chemoselective ligation, of which native chemical ligation is the best known ^[1]. Sustainability is the other frontier: current peptide manufacturing relies mainly on legacy technology that consumes large amounts of hazardous reagents and solvents, and in 2016 the American Chemical Society's pharmaceutical green chemistry roundtable named greener peptide processes a critical unmet need ^[18]. Industrial efforts now focus on substituting, recycling and reducing solvents ^[3].

What this means at the bench

• Expect a TFA salt unless the documentation says otherwise, and correct molar calculations for it; the molarity calculator does the arithmetic once net peptide content is known.
• Read the sequence for risk motifs. Asp–Gly or Asp–Asp pairs, proline runs, His, Cys and Met each point to impurity types that HPLC may not fully resolve; how to read peptide sequences explains the notation.
• Ask for MS data, not just purity. Isomeric impurities from racemisation and aspartimide chemistry have the target's mass, so identity and purity tests answer different questions; peptide purity explained sets out which impurities each test catches.
• Treat price as a clue to process. Length, modifications and purity target drive cost; the supplier evaluation guide covers what to ask when a price looks implausible.
• Store the lyophilised powder dry and cold. The chemistry that degrades peptides needs water; the peptide storage guide covers practice.