How a peptide is made: solid-phase synthesis explained for physicians

The problem Merrifield solved

A peptide is built by joining amino acids one at a time, protecting the groups that must not react and removing the protection just before the next coupling. The principle is old and simple; the difficulty is practical. In classical solution synthesis, the product must be separated from by-products and reagents after every step, and material is always lost. The Swedish Academy put it this way: if each step yields 90%, a very good yield, the overall yield after 100 steps would be 0.003%¹.

In 1963 Bruce Merrifield published the synthesis of a tetrapeptide on a solid support³. His idea was to anchor the first amino acid to an insoluble polymer in the form of small beads, add the next ones one after another and release the chain only at the end. Purification after each step is replaced by washing the resin, the yield of each step rises to 99.5% or better and the process can be automated². With this method his group synthesized an active enzyme, ribonuclease, with 124 amino acids².

The synthesis cycle, step by step

The chain grows from the carboxyl end toward the amino end¹⁷. Each cycle repeats the same sequence: remove the protecting group from the anchored amino acid, couple the next activated amino acid and wash. To keep chains from ending up incomplete, manufacturers repeat the coupling when needed and acetylate the sites that did not react (capping) so they stop growing⁷. Progress is monitored with tests such as the Kaiser test, a colorimetric ninhydrin assay that detects free primary amines⁷. At the end, an acid cocktail releases the peptide from the resin and removes the side-chain protecting groups⁷.

Fmoc vs Boc: two chemistries

There are two main strategies: Boc/benzyl and Fmoc/tert-butyl⁴. In Boc chemistry the protecting group is removed with TFA in every cycle; an optimized 1992 protocol used 100% TFA, 15- to 19-minute cycles and assembled 75 residues per day⁶. The final step of Boc chemistry requires hydrofluoric acid (HF)⁵, and the EMA asks for residual fluoride to be controlled when it is used⁷. In Fmoc chemistry the protecting group is removed with a base (piperidine) and the final cleavage uses TFA, without HF⁵⁹¹². Fmoc chemistry is now the method of choice: its building blocks are high quality and low cost thanks to multiton production of therapeutic peptides, and most Fmoc amino acids are available at HPLC purity above 99%⁵. Boc chemistry keeps its niches, for example "difficult" sequences that aggregate on the resin⁶.

What can go wrong: synthesis impurities

The impurities of a synthetic peptide are not mysterious; they are catalogued and each has a chemical origin⁷⁸:

• Deletion sequences: missing one or more amino acids because of incomplete coupling or deprotection⁷⁸.
• Truncated sequences: acetylated fragments produced by capping the chains that did not react⁷.
• Insertion sequences: an amino acid coupled twice, for example because of excess reagent⁷⁸.
• Epimerization (racemization): an L-to-D switch in one amino acid; in a 20-residue model peptide, cysteine, histidine and aspartic acid were the most susceptible, and lowering the coupling temperature from 80 °C to 50 °C limited the problem for histidine and cysteine⁹. Urethane-type protecting groups such as Fmoc resist the oxazolone formation that causes it⁷.
• Aspartimide: a side reaction of aspartic acid promoted by the base used in Fmoc chemistry; it can give up to nine by-products, some of which elute in the same peak as the target peptide⁵¹⁰.
• Adducts and oxidation: cleavage releases reactive cationic species that can attack tryptophan, tyrosine or methionine, so scavengers are added; side-chain oxidation, unremoved protecting groups and dimers or oligomers are also described⁷⁸.

The longer the chain, the more cycles accumulate and the more opportunities for these reactions¹²; in addition, sequences that aggregate on the resin are considered "difficult"⁶. Later degradation pathways (deamidation, hydrolysis, oxidation) are reviewed in Peptide stability.

Cleavage and counter-ions: the TFA case

TFA is used as the cleavage agent and as the ion-pairing reagent in reversed-phase HPLC purification, so peptides are usually obtained as TFA salts¹¹¹². This has consequences. In 1999, TFA at 10⁻⁸ to 10⁻⁷ M reduced osteoblast numbers in culture, and the TFA salts of amylin and calcitonin produced less proliferation than their hydrochloride salts; the authors recommended converting peptides to the hydrochloride or an equivalent salt before measuring biological effects¹¹. A mouse study published in 2026 reported that TFA, alone or as the counter-ion of several short peptides, lowered LDL cholesterol, triglycerides and atherosclerotic lesions in an atherosclerosis model, with mechanistic data pointing to binding to the PPAR-α receptor; the authors note it is not known whether these effects extend to humans¹³.

The fix is known. The EMA states that the type of counter-ion can affect the biological and physicochemical properties of the peptide, that acetate is usually used for synthetic peptides and that its amount should be controlled with a justified upper limit⁷. The exchange is also validated: an ETH Zurich group replaced TFA with chloride by freeze-drying in hydrochloric acid, with 10 mM as the optimal condition and no effect on purity¹².

Purification, freeze-drying and control

The crude peptide is a complex mixture. It is purified by preparative chromatography: fractions are collected, the purest are pooled and side fractions can be repurified under defined criteria⁷. It is then isolated, usually by freeze-drying (lyophilisation), which the EMA considers common practice; precipitation, crystallisation and spray drying are also used⁷. The resulting powder is often very hygroscopic⁷.

The specification of a pharmaceutical peptide separates concepts that are often confused: identity must unambiguously confirm the sequence; purity is measured by chromatography, with a 0.1% reporting threshold under the European Pharmacopoeia; content (assay) is expressed on the anhydrous, counter-ion-free substance; and water, counter-ion, residual TFA, solvents and endotoxins are controlled separately⁷. When two impurities elute together as a single peak, a 1.0% qualification threshold applies⁷. How to read these data on a certificate is covered in How to read a peptide certificate of analysis, and what independent laboratories have measured in What is really in the vial?.

Chemical synthesis or yeast: two routes to the same molecule

Not every peptide is made chemically. EMA product information states that reference semaglutide and liraglutide are produced in Saccharomyces cerevisiae by recombinant DNA technology¹⁴¹⁵. Synthetic versions of molecules that started as recombinant products now also exist. A study comparing batches from different manufacturers found that synthetic and recombinant liraglutide from five suppliers had impurity profiles distinct from the originator, that scale and process influenced physical stability and that trace metals increased high-molecular-weight species¹⁶. The route matters less than the control: both can give a quality product if the impurity profile is characterized and compared⁷¹⁶.

Scale also brings environmental challenges. Conventional solid-phase synthesis has extreme process mass intensity and poor atom economy because of the excess of protected amino acids and coupling reagents¹⁷, and it depends on solvents such as dimethylformamide (DMF). In a 2024 study, N-butylpyrrolidinone performed on par with DMF in the synthesis of five peptides, including thymosin α1¹⁸. Peptide chemistry keeps improving.

The limits

This chemistry is mature and precise, but it has limits worth knowing. Chromatographic purity does not detect what co-elutes with the main peak, and aspartimide can produce exactly that kind of impurity⁵⁷. The biological effects of TFA come from cells and mice; their relevance in humans is not established, and the 2026 study declares a patent on therapeutic uses of TFA¹¹¹³. The EMA guideline describes what is required of a medicine; research material is not necessarily made or tested to the same extent, so each batch's certificate remains the reference⁷. And the comparison of liraglutide manufacturers shows that two products with the same sequence can differ in impurities and stability¹⁶.