A deletion sequence is what you get when a coupling fails quietly and the chain keeps going. The unreacted α-amine is still there at the start of the next cycle, it accepts the next residue, and the chain finishes one residue short with the gap somewhere in the interior. Everything a purification exploits — hydrophobicity, net charge, hydrodynamic size, backbone conformation — is set by the residues that are present, so the impurity tracks the target through the gradient. Ball and Mascagni, developing a route to small proteins by stepwise solid-phase synthesis, named the bottleneck without hedging: classical stepwise SPPS had been used successfully for proteins up to 150 residues but usually with poor yields and homogeneity, and the major limitation was the inability to separate chromatographically similar deletion and truncated impurities from the target sequence [1].
This is not an occasional failure mode. Synthetic peptides always contain some level of deletion and truncated material as a consequence of how they are produced, and optimised synthesis protocols together with purification are what keep that level low enough that the impurities do not, for instance, drive false-positive readouts in the in vitro stimulation assays where such peptides are used [2].
What capping actually changes
Capping does not make a coupling work. It changes what the failure turns into. Acylating every free α-amine immediately after the coupling step converts the chains that failed into N-acylated species that cannot accept another residue, so a chain that stalls at cycle k stops at cycle k. Lindeberg and colleagues put the consequence in the form that matters for purification design: capping with acetic anhydride will, in principle, produce truncated peptides as the only side-products due to incomplete couplings, so a target built with an N-terminal metal-binding residue such as His, Trp or Cys is the only species carrying that affinity label, and most of the impurities can be removed by a single passage through an immobilised metal ion affinity column — shown on peptides of 30 to 40 residues [3]. That approach is stated to be useful especially when such residues are not present in other parts of the molecule [3], which is worth naming as a constraint: the handle has to be unique to the N-terminus, and that is a condition on the target sequence rather than on the capping step. Montanari and Kumar make the irreversibility explicit for a different cap: their fluoroalkylated amine is no longer able to couple in further peptide coupling steps and is stable to standard peptide synthesis conditions [4].
The gain is one of degree, but a large one. A truncation stopped at cycle k differs from the target by every residue after k; a deletion peptide differs by one. That is the premise capping rests on, although Ball and Mascagni describe deletion and truncated impurities alike as chromatographically similar to the target sequence [1], so how much a truncation actually gains in a gradient is not settled by chain length alone.
Acetic anhydride, and what the acetyl group does not do
Acetic anhydride is the reagent named for the step in the automated protocols cited here, where it is run as a discrete treatment after coupling. Zhang, Pokharel and Fang capped failure sequences with acetic anhydride during automated SPPS, then attached a polymerisable methacrylamide tag to the full-length sequences alone; purification was achieved by polymerising the full-length sequences, washing away impurities, and cleaving the peptide product from the polymer [5].
Notice what is doing the work in that scheme. The acetyl group separates nothing. What it does is make the target the only chain still carrying a reactive α-amine, and every non-chromatographic purification then grips that asymmetry from one side or the other. Handle on the target: the metal-binding N-terminus read by IMAC [3], the polymerisable tag [5], or the reversible Fmoc-based chromatographic probe derivatised with a lipophilic group that Ball and Mascagni added for a fast one-step reversed-phase purification [1]. Handle on the failures instead: Montanari and Kumar's trivalent iodonium salt, which reacts vigorously with free amines to deliver a long-chain fluoroalkyl group, after which deletion products are removed using flash fluorous chromatography [4]; or the reactive caps of Wunderlich and colleagues, two masked-thiol tags used as the capping reagent during SPPS, whose masked functionality is liberated on cleavage from the resin and then scavenged by chemoselective reaction with a maleimide-functionalised solid support, demonstrated on a 15-mer [6].
Two caveats travel with all of these schemes. The first is adoption. Wunderlich and colleagues, motivating their method, observe that HPLC purification is associated with high solvent consumption, relatively large effort and lack of efficient parallelisation, and that although many catch-and-release purification methods have been developed over the last decades, with one exception none of them has been widely established in industry and academia [6]. Preparative chromatography therefore remains the arrangement in general use [6], which is a reason to be clear-eyed about what plain acetylation does and does not buy on its own. The second caveat is orthogonality: a thiol-based cap collides with a target that contains cysteine, and it took a suitable protecting group strategy before that method could be applied to a thiol-containing peptide [6]. Whatever chemistry a cap introduces has to be absent from, or protectable in, the sequence being made.
Ball and Mascagni's optimised protocol — HBTU/HOBt activation, N-(2-chlorobenzyloxycarbonyloxy)succinimide as capping agent, and the reversible lipophilic probe — gave a 101-residue rat chaperonin 10 that ESI-MS indicated was virtually free of deletion impurities, at 9.6% overall yield. Its purity read 100% by RP-HPLC under four different conditions and 84% by capillary zone electrophoresis [1]. The distance between those two figures is the part worth carrying away: a reversed-phase method can score a preparation clean that an orthogonal separation does not. The analyte there was a 101-residue synthetic protein, not a short peptide. Across the three protocols they compared, 29%, 55% and 81% of correctly folded heptameric structure was obtained, the last from that optimised route; spontaneous folding was found to correlate well with the degree of purity as assessed by chromatography, ESI-MS and sequencing [1]. If the preparation is destined for a folding or binding experiment, deletion content is not a bookkeeping detail.
When capping is the wrong move
The reagent does not confine itself to the α-amines of failed chains. Ishiguro and Eguchi, applying an Ac2O capping procedure to the synthesis of a peptide containing His, observed a serious side reaction, chain-terminating and attributable to histidine and acetic anhydride, and studied its mechanism [7]. The chains that coupled correctly are exposed to the acylating agent at every capping step, so a capping protocol is a decision about the whole resin, not only about the fraction that failed.
Capping is also a commitment. An incomplete coupling that has not yet been capped can still be completed; recoupling and capping are alternative responses to the same risk rather than complementary ones, since the second forecloses the first. Once the amine is acylated the chain is dead and its material is lost yield [4]. Lucietto and colleagues took the other branch for Mycobacterium tuberculosis chaperonin 10 and its N-truncated fragments, synthesising by the solid-phase methodology using a double coupling protocol and without the aid of capping agents, then leaning on a charge-based separation instead: isoelectric focusing carried out in solution followed by semi-preparative RP-HPLC, a combination they found in general superior to ion exchange chromatography with RP-HPLC both in recovery of final material and in its purity [8].
And an acetyl group adds no functionality that a scavenger resin, an affinity column or a fluorous phase can grip. That is the plainest reading of the reagent development above: every purification cited here had to introduce a handle of its own, either on the target [1, 3, 5] or on the failures [4, 6]. If the separation plan depends on a property that acetylation barely perturbs, capping has moved the impurity from one hard peak to a slightly easier one and no further.
Capping selectively rather than routinely
Capping after every cycle spends reagent, cycle time and acylation exposure on cycles that did not fail. Capping after the couplings that are actually at risk requires knowing which those are, and the failures are not randomly distributed. Krchnak, Flegelova and Vagner established the shape of it: non-random incomplete aminoacylation of a pendent peptide chain on an insoluble polymeric support is sequence-dependent and is caused by aggregation of peptide chains, manifested by a decreased swelling capacity. They measured the volume of the swollen peptidyl-resin after each coupling across the syntheses of 87 sequence-unrelated peptides, derived for each amino acid an aggregation parameter reflecting the propensity of that swollen volume to decrease, and used those parameters to predict potentially difficult sequences [9].
The predictive picture has since been reweighted. Tamas and colleagues, building a model from existing peptide synthesis datasets, report composition-dependent aggregation as a stronger predictor than sequence-based patterns, and construct a composition vector representation from which an ensemble of trained models predicts aggregation behaviour and recommends the optimised use of aggregation-reducing tools [10]. If composition rather than local motif carries most of the signal, a list of notorious dipeptide motifs is a weaker basis for choosing which cycles to cap than an assessment of the whole sequence.
Measurement beats prediction where the instrument supports it. In automated fast-flow synthesis, Mohapatra and colleagues applied deep learning to ultraviolet-visible data collected from 35,427 individual Fmoc deprotection reactions; the integral, height and width of those time-resolved traces indirectly allow analysis of the iterative amide coupling cycles on resin, and the model predicts the outcome of deprotection reactions with less than 6% error [11]. That is a deprotection measurement, and the step from it to coupling completeness is indirect — but a per-cycle signal of that kind is what turns selective capping into an informed choice rather than a guess.
The first residue is a separate problem
The usual capping discussion concerns cycle n; the resin itself needs capping before cycle one. Rentsch and colleagues make the point for preloaded 4-alkoxybenzyl alcohol resins: a critical factor is the capping of remaining hydroxyl groups after loading with the first amino acid, since these free alcohols lead to truncated sequences during the following SPPS steps. Detection of hydroxyl groups by colour tests is difficult and unreliable, so capping efficiency is often controlled by time-consuming peptide test syntheses; they developed a two-dimensional high-resolution magic angle spinning NMR method for quantifying remaining 4-alkoxybenzyl alcohols in Fmoc-Xaa-Wang resins with a detection limit of 1 mol-% [12]. They validated it against samples of known ratio between Fmoc-Ala-Wang and 4-alkoxybenzyl alcohol resin, and application to preloaded Fmoc-Ala- and Fmoc-Thr(tBu)-Wang test resins demonstrated that the full range of essential amino acids can be quantified without further spectrometer calibration [12]. Residual alcohol on the support is a source of truncated material that the per-cycle amine capping step was never designed to address, and the colour tests used to police that step are an unreliable way to find it [12].
Kaiser, chloranil, and what a colourless bead proves
Two colour tests carry the routine monitoring load, and they interrogate different populations. The Kaiser test is a colour test for detection of free terminal amino groups in the solid-phase synthesis of peptides [13]; its quantitative form, from Sarin, Kent, Tam and Merrifield, turns the same ninhydrin reaction into quantitative monitoring of solid-phase peptide synthesis rather than a yes-or-no endpoint [14]. Vojkovsky's re-examined chloranil procedure is aimed at secondary amino groups on solid phase: about ten times more sensitive than the original method, it detected reliably all the secondary amino groups examined except N-(p-tolyl)glycine, in a substitution range down to 2.8 µeq/g, and the protected imidazole in His(Trt) did not interfere with the detection [15].
So the choice of test follows the residue exposed on the resin at that moment rather than the sequence as a whole. Where the exposed N-terminal amine is secondary, the chloranil procedure is the one with a published detection limit for that population [15]; the Kaiser test's stated remit is free terminal amino groups [13]. Running the wrong one and reading a negative is how a failed coupling gets recorded as complete.
Be precise about what a negative test licenses. A colourless bead after a capping step says the ninhydrin-reactive amine population has fallen below the test's threshold — that is, that acylation consumed the free amine. It says nothing about how much of the resin was capped rather than coupled, and that second number is the one that predicts how much truncated material the purification will have to remove. Getting it requires the quantitative form of the assay run before the capping step [14], or reconstruction afterwards from mass analysis of cleaved material. Colour tests also do not reach every relevant nucleophile: they are the wrong instrument for residual resin hydroxyls, which is why the preloaded-resin problem was taken to NMR [12].
Fitting the decision to the rest of the workflow
Three decisions travel together and are best made at once. Whether a given coupling is at elevated risk is an aggregation question, and belongs with the guide to on-resin aggregation and difficult sequences. Whether to answer that risk by recoupling or by capping is the trade set out above: capping converts an intractable separation into a tractable one at the cost of yield, of exposing every chain to an acylating agent, and of any chance to rescue the stalled chains. What the capped material then looks like in the crude belongs with the guide to synthesis impurities, where truncations sit alongside oxidation, incomplete deprotection and racemised species in the profile a purification has to resolve. The one thing capping cannot do is undo the coupling failure, which is why it is worth paying for at the cycles where failure is likely and hard to justify at the cycles where it is not.