How Peptides Are Made: Synthesis, Purification, and Quality

Almost every peptide you encounter is manufactured by one of two fundamentally different routes, and knowing which one was used explains a great deal about its purity, cost, and impurity profile.

The first route is chemical synthesis. Here a chemist assembles the peptide bond by bond in a reactor, joining individual protected amino acids in a defined order. The overwhelmingly dominant version of this approach is solid-phase peptide synthesis (SPPS), covered in detail in the next section. Chemical synthesis is fast, highly automatable, and well suited to short and medium-length peptides. It also allows non-natural amino acids, modifications, and labels that biology cannot easily make.

The second route is recombinant production. Instead of building the molecule chemically, manufacturers insert the DNA sequence coding for the peptide into a host organism, most commonly the bacterium Escherichia coli, and let the cell's own machinery translate it. The cells are grown in fermenters, and the peptide is then extracted and purified from the biological material.

The practical dividing line is length and complexity. Published work notes that reliable chemical synthesis and purification of peptides longer than about 30 amino acids becomes difficult and highly sequence-dependent, while recombinant expression becomes increasingly attractive in the roughly 30 to 100 amino acid range and for peptides that are prone to degradation or hard to make chemically. Below that range, chemistry usually wins on speed and simplicity; above it, biology often wins on scalability and cost per milligram.

SPPS, introduced by Robert Bruce Merrifield in the 1960s, is the workhorse of peptide manufacturing. Its central idea is elegant: anchor the growing peptide to an insoluble polymer bead (the resin), so that after each chemical step the excess reagents and byproducts can simply be washed away while the peptide stays put.

The peptide is built from the C-terminus (the tail) toward the N-terminus (the head), the reverse of how ribosomes work in cells. Each amino acid is added in a repeating cycle:

• Deprotection: the temporary protecting group on the growing chain's exposed amino group is removed so it can react.
• Coupling: the next amino acid, with its own amino group protected and its carboxyl group chemically activated, is joined to the chain, forming a new peptide bond.
• Washing: excess reagents and soluble byproducts are rinsed away before the next cycle.

Two protecting-group strategies exist. The older Boc/benzyl approach uses acid to remove the temporary group but requires harsh, corrosive hydrogen fluoride at the end. The modern and far more common Fmoc/tert-butyl strategy uses a mild base (piperidine) to remove the temporary Fmoc group and a strong acid (trifluoroacetic acid, TFA) at the end. These two conditions are orthogonal, meaning each can be triggered without disturbing the other, which is why Fmoc chemistry dominates today.

Coupling relies on activating reagents such as carbodiimides (DIC), uronium salts (HATU, HBTU, TBTU), or phosphonium salts (PyBOP), usually with additives that suppress side reactions. When the full sequence is assembled, a final cleavage step with TFA releases the peptide from the resin and strips the side-chain protecting groups. Scavenger molecules are added at this stage to trap the reactive fragments generated during cleavage, preventing them from damaging the product.

For longer peptides and proteins, biology is often the better factory. In recombinant production, the gene encoding the desired peptide is cloned into an expression vector and introduced into a host such as E. coli, yeast, or mammalian cells. The host is cultured at scale, transcribing and translating the sequence into peptide chains using natural ribosomes.

This approach has real advantages. It can generate large quantities cost-effectively once a process is developed, it uses the cell's high-fidelity translation machinery, and it generates less chemical waste than large-scale synthesis. Reported figures suggest recombinant unit costs can fall sharply with scale, becoming competitive with chemical strategies for suitable peptides.

It also has characteristic challenges. Small peptides are often degraded by the host's own enzymes, so they are typically expressed as part of a larger fusion protein, joined to a carrier or tag that improves stability and solubility. After purification, that tag must be cleaved off, for example with a specific protease, to release the native peptide, and the cut product must then be separated again. Peptides that require disulfide bonds may need additional denaturation and refolding steps, and hydrophobic sequences can be difficult to keep in solution. Recombinant production also cannot easily incorporate the non-natural amino acids and chemical modifications that synthesis handles routinely.

Many well-known therapeutic peptides are made recombinantly, and the choice between synthesis and expression is ultimately an engineering decision driven by the specific sequence, required scale, and needed modifications.

A crucial point that non-specialists often miss: the peptide that comes off the synthesizer or out of the fermenter is never a single pure substance. It is a mixture, and the target molecule is only part of it.

In SPPS, the reason is statistical. Each coupling step is highly efficient but not perfect. If a single step proceeds at, say, 99% completion, then across dozens of steps the small failures compound. A chain that misses one amino acid becomes a deletion sequence; a chain whose synthesis stalls entirely becomes a truncated sequence. The longer the peptide, the more of these related impurities accumulate, which is a core reason long peptides are harder to make at high purity.

Specific sequences also invite specific chemical side reactions. One of the best-characterized is aspartimide formation: when aspartic acid sits next to certain residues (aspartate-glycine is the classic troublemaker), the repeated exposure to base during Fmoc removal can cyclize the side chain into a five-membered ring. That intermediate then breaks down into a family of byproducts, including racemized and rearranged (alpha- and beta-linked) peptides that are notoriously difficult to separate from the real product. Other recognized issues include incomplete removal of protecting groups and unwanted modifications introduced during cleavage.

Recombinant material carries its own impurity classes, including host-cell proteins, DNA, endotoxins, incompletely processed fusion protein, and misfolded variants. In both routes, the raw output is a population of molecules, and quality is determined by what happens next: purification.

Purification is the step that turns a crude mixture into a usable peptide, and for peptides the dominant technique is reverse-phase high-performance liquid chromatography (RP-HPLC), which has been the standard since the late 1970s.

The principle is separation by hydrophobicity. The crude peptide is dissolved and pumped through a column packed with a hydrophobic stationary phase, typically silica beads coated with C18 (octadecyl) chains. A mobile phase, usually water and acetonitrile with a small amount of an acidic modifier such as trifluoroacetic acid, is run through as a gradient that gradually becomes more organic. Peptides stick to the stationary phase according to how hydrophobic they are and release at different points in the gradient. Because truncated chains, deletion sequences, and chemical byproducts differ slightly from the target in size and hydrophobicity, they emerge at different times and can be collected separately.

Preparative HPLC does the bulk separation; analytical HPLC on a smaller scale is then used to measure the result. In practice, purification is where a great deal of the cost and skill lies. Impurities that are chemically very similar to the target, such as some aspartimide-derived byproducts or single-residue deletions, can co-elute, meaning they leave the column at almost the same time and are hard to remove completely. This is why some sequences can be purified to very high purity while others plateau lower despite the same effort, and why purification method development is a genuine specialty rather than a routine step.

Once purified, a peptide is usually in a dilute aqueous solution that is not stable for long. The standard way to preserve it is lyophilization, also called freeze-drying.

Lyophilization removes water by sublimation: the solution is frozen solid, then placed under deep vacuum so the ice converts directly from solid to vapor without passing through a liquid phase. What remains is a dry, amorphous powder, often the fluffy cake seen at the bottom of a peptide vial. Removing water dramatically slows the chemical and physical degradation pathways that would otherwise act on the molecule, which is why peptides are generally shipped and stored as lyophilized powder rather than as ready-made solution.

Formulation matters here. Freeze-dried preparations often include excipients: lyoprotectants or cryoprotectants such as trehalose or sucrose that form a protective glassy matrix and hydrogen-bond to the peptide in place of the removed water, and bulking agents such as mannitol or glycine that give the cake physical structure. These additives help the peptide survive both the freezing and drying stresses and the following months of storage.

Lyophilization is not a permanent guarantee of stability. Long-term integrity still depends on the amino acid sequence, residual moisture in the cake, storage temperature, exposure to oxygen and light, and the quality of the vial seal. Poorly dried or poorly sealed material can still degrade, which is one more reason batch testing matters rather than assuming a powder is stable by default.

Quality is not a single number, and it is not something you can judge by looking at a vial. It is the combination of several measurable properties, each answered by a different analytical test.

• Purity: what fraction of the material is the target molecule versus related impurities. This is most commonly reported as HPLC purity, typically measured by the area of the main peak in an analytical RP-HPLC run at a defined wavelength. A purity figure is only meaningful alongside the method used to obtain it.
• Identity: whether the main peak is actually the intended peptide and not merely a pure something-else. This is confirmed by mass spectrometry, which measures the molecular weight and can flag subtle modifications. Purity and identity are independent questions; a sample can be highly pure and still be the wrong molecule.
• Impurity profile: which specific impurities are present and at what levels. Regulators take this seriously. In the context of certain synthetic generic peptides, the FDA has articulated thresholds under which new peptide-related impurities must be characterized and justified, in part because some impurities can carry immunogenicity risk. That framework applies to approved drug products, not to research chemicals, but it illustrates why simply reporting a single purity percentage is considered insufficient for rigorous characterization.
• Content, moisture, and safety attributes: how much actual peptide is present (net peptide content, since salts and water add mass), plus residual solvents, water content, and, for injectable-grade material, endotoxin.

Because every batch is chemically unique, these attributes must be measured per batch. That is the entire point of a certificate of analysis (COA): a batch-specific document reporting the actual test results, ideally with the underlying HPLC and mass-spectrometry data, for the exact lot in hand. A generic spec sheet or a purity claim with no batch number and no chromatogram is a marketing statement, not evidence. Understanding how peptides are made makes clear why the COA, tied to a specific lot, is the only real proof of quality.