Solid-phase peptide synthesis and where impurities come from

Almost every figure on a peptide certificate is a statement about the outcome of a synthesis. The impurities enumerated on that certificate are not random contamination; each class of them is produced by an identifiable step of the assembly, and each is either easy or hard to see depending on how it differs from the intended molecule. This page sets out the chemistry in outline and then traces the characteristic impurities back to the stage that made them. It describes published synthetic chemistry, not a laboratory procedure.

The Merrifield principle

In 1963 Bruce Merrifield published the synthesis of a tetrapeptide assembled under heterogeneous conditions from the C-terminus to the N-terminus on a polymeric solid resin ¹. The first residue is attached through its carboxy terminus to the support, which protects the C-terminus against side reactions for the whole synthesis, and a short linker is interposed between residue and support ¹. The chain is then extended by repeating three operations: removal of the temporary Nα-protecting group, activation of the incoming residue's relatively inert carboxy group, and coupling ¹.

The advantage that made the method dominant is as much mechanical as chemical. Every reaction is carried out in a single vessel, and unreacted reagents and by-products are removed by washing rather than by isolating and purifying each intermediate, so large excesses of reagent can be used and high coupling yields obtained ¹. Those same excesses and washings are also the reason a crude product carries a characteristic population of closely related by-products rather than a single contaminant.

Fmoc/tBu and Boc/Bzl strategies

Merrifield's original chemistry used Boc as the temporary Nα group and benzyl-type groups for reactive side chains. Boc is removed with trifluoroacetic acid, whereas benzyl removal requires a strong acid such as hydrogen fluoride, so the scheme rests on graded acid lability rather than on two different mechanisms ¹. The introduction of the Fmoc group changed that: Fmoc is removed under basic conditions by β-elimination with piperidine, while the tert-butyl side-chain groups and the linkage to the resin are cleaved by acidolysis with trifluoroacetic acid ¹. The two removals therefore proceed by entirely different chemical mechanisms, acidolysis in one case and base-mediated elimination in the other ².

That orthogonality is the main argument for the Fmoc/tBu route, because it permits selective removal under milder overall conditions and supports more complex protecting-group schemes ¹. It also avoids the vessels required for corrosive and toxic hydrogen fluoride, and avoids the repeated trifluoroacetic acid treatments of the Boc route, which can affect sensitive peptide bonds and promote acid-catalysed side reactions ¹. The Boc/Bzl route retains a reputation for handling difficult sequences, where repeated acid treatment counteracts aggregation ¹.

Resin, linker and coupling chemistry

A support has to be insoluble in every solvent used, chemically and mechanically robust enough to survive filtration, and able to swell, since most of the chemistry happens inside the polymer matrix rather than on its surface ¹. Merrifield's first support was a copolymer of styrene and cross-linked divinylbenzene; the three classes in general use are traditional polystyrene, polyethylene glycol-functionalised polystyrene, and resins built from polyethylene glycol alone ¹.

The linker is the reversible connection between support and chain: it links the peptide to the solid support and protects the C-terminal carboxylic group ², and it determines the loading of the resin, the conditions under which the chain is released, and the C-terminal functionality of the product, most often an acid or an amide ¹. Some linkers, such as 2-chlorotrityl and Sieber amide, release partially or fully protected chains instead ¹.

Coupling requires the carboxy group to be converted into a more electrophilic species. Carbodiimides such as DCC and DIC served as the principal activators for decades, and their tendency to promote racemisation drove the development of suppressants such as HOBt, which traps the O-acylisourea intermediate responsible ¹. Uronium reagents such as HATU and TBTU and phosphonium reagents such as PyBOP followed, and Oxyma was introduced in 2009 as an additive for carbodiimide-mediated coupling in place of HOBt, whose explosive potential is inherent ¹. The choice between them is a trade-off: greater reactivity for hindered residues against a lower rate of epimerisation ¹.

Capping

A capping step is optional in the cycle and is commonly carried out with a solution of acetic anhydride, dimethylformamide and pyridine or a tertiary amine base ³. Its effect is to acetylate amino groups that failed to couple, so a chain that has missed a residue is terminated rather than carried forward; acetylation is reported to be particularly useful for reducing deletion products when long or difficult sequences are synthesised ³. Capping therefore converts a potential deletion impurity into a truncated one, which is a different and generally easier separation problem.

Cleavage and global deprotection

At the end of the assembly the temporary group is removed from the N-terminal residue, and conditions are then applied that strip the side-chain protecting groups and release the chain from the resin in the same operation ¹. That final step is carried out in the presence of scavengers, whose function is to trap the highly reactive carbocations generated as protecting groups depart and which would otherwise react with the peptide to form unwanted by-products ¹.

Where the characteristic impurities come from

Peptide-related impurities arising during synthesis have been summarised as insertions, deletions, substitutions, racemisation and β-alanine-containing contaminants ⁴. Traced to their origins:

Deletion sequences arise when a coupling does not go to completion and the unreacted chain is deprotected and extended in the next cycle, producing a chain missing one internal residue.

Truncated sequences arise when elongation stops altogether. The usual cause is aggregation on the resin, which renders the free N-terminus inaccessible to the next coupling and results in either a truncated synthesis or a very low crude recovery ⁵. Published syntheses of aggregation-prone sequences report crude material containing discrete truncated side products ⁵.

Racemisation and epimerisation occur where the base required for activation, or the base used for deprotection, allows a stereocentre to invert. The best-characterised route in Fmoc chemistry is aspartimide formation: deprotonation of the backbone amide next to an aspartic acid residue leads to nucleophilic attack on the side-chain carbonyl, and reversible rehydration of the resulting imide results in racemisation and a mixture of α- and β-peptides ⁶. The β-aspartyl products are typically formed in a three-to-one ratio relative to the α-aspartyl peptide ⁶. The Boc route produced little aspartimide because of its cyclohexyl side-chain protection, but the hydrogen fluoride used in its final cleavage promotes an acid-catalysed version of the same reaction ⁶.

Incomplete deprotection leaves temporary or side-chain protection in place. Partial removal of the temporary group is one of three problems that recur in difficult sequences, alongside incomplete couplings and difficulty loading sterically hindered residues ³, and complete removal of the temporary group matters precisely because it conditions the efficiency of the next coupling ⁵.

Side-chain adducts are the by-products of the final cleavage, formed when a carbocation released from a departing protecting group is captured by the peptide rather than by a scavenger ¹.

Why some impurities are harder to find than others

The detectability of each class follows from how far it differs from the target.

Deletion peptides differ by one residue and often co-elute with the target peptide ⁷. A co-eluting species is counted inside the main peak, because chromatographic purity is reported as the sum of impurities expressed as a share of total detected area ⁸. Their mass does differ from the calculated mass, so they can be found by mass spectrometry when they are looked for specifically.

Truncated sequences usually differ substantially in both mass and hydrophobicity, which makes them the most tractable class: the same published syntheses that report them also report separating them chromatographically ⁵.

Diastereomers produced by racemisation are the hardest case, because they share the molecular formula and therefore the intact mass of the intended peptide. They are visible only if the separation resolves them, and the side products of the aspartimide route are described as notoriously difficult to separate ⁶.

Adducts and incompletely deprotected species shift the mass by a defined increment and are consequently straightforward to assign once a mass spectrum is examined for them.

Purification by preparative reversed-phase chromatography

Once released, the crude material is separated from the resin and purified chromatographically, and the parallel development of separation techniques was a precondition for the success of the method, both analytically and preparatively ¹. Published preparative purifications of synthetic peptides use C18 or C4 reversed-phase columns with linear gradients of acetonitrile against water, both phases containing about 0.1% trifluoroacetic acid, with the collected fractions lyophilised ⁵.

The limits of the technique are the limits of the separation itself. Aggregating sequences chromatographed under conventional acidic conditions at room temperature give an asymmetric, broad and unresolved peak, which is itself evidence of aggregation rather than of composition ⁵. Anything that co-elutes with the target survives purification and is then counted as target by the analytical method that follows.