Native chemical ligation asks one thing of the acyl segment: a thioester at its C-terminus. Two unprotected peptides meet in aqueous buffer, the N-terminal cysteine of the second attacks the thioester of the first, and the transient thioester-linked species rearranges spontaneously to give a native amide bond at the junction [1]. Almost everything else about the reaction is forgiving. The acyl donor is not, and building it is exactly where Fmoc synthesis lets you down.
The thioester bond has poor stability to the base used to remove the Nα-Fmoc group, which for years largely restricted thioester solid-phase synthesis to Boc/benzyl chemistry [2]. The workarounds all share a shape: carry something inert through chain assembly, then convert it to the acyl donor afterwards. A resin-bound aryl hydrazine linker is one — it is stable to Fmoc-SPPS conditions, and mild oxidation at the end of synthesis converts the acyl hydrazine into a highly reactive acyl diazene that reacts with an α-amino acid alkyl thioester to release the peptide α-thioester, a route used for thioesters, cyclic peptides and an SH3 domain [2]. On-resin N-acylurea is another: mildly activated N-acylurea peptides are formed on the support after chain assembly, which avoids over-activating the C-terminus [3]. These belong to the peptide thioester guide.
What the C-terminal hydrazide changes
The free C-terminal hydrazide is a different proposition from a linker. It is a functional group on the finished, cleaved, side-chain-deprotected peptide, and it is unmasked in the same tube where the ligation runs. Fang and co-workers put it forward as the reagent equivalent of a thioester synthon that is stable to the conditions of native chemical ligation, with the published summary framing the method's selectivity as pH-determined [4]. Peptide hydrazides are assembled by routine Fmoc chemistry [5, 6].
That stability is the entire design. A hydrazide sitting at the C-terminus of segment 2 is a spectator while segment 1's thioester is being consumed at ligation pH. It only becomes an acyl donor when you deliberately acidify, chill and oxidise. One functional group therefore covers two roles that would otherwise need two orthogonal protecting groups.
Nitrite oxidation: a deliberately narrow window
The reported activation conditions are tight. An unprotected peptide hydrazide, typically 0.5–2 mM in 6 M guanidinium chloride buffer, treated with 5–20 mM NaNO2 at pH 3.0–4.0 and −20 to −10 °C, is oxidised to the acyl azide within 30 min [7].
The temperature is not conservatism. Acyl azides are unstable at higher temperature and undergo Curtius rearrangement or degradation above 0 °C to give isocyanate-derived byproducts [7]. Those byproducts are not thioester precursors, so anything that rearranges is straight yield loss with no route back. In practice this means pre-chilling the buffer and the nitrite stock, running the oxidation in a cooled bath rather than on the bench, and treating the 30 min as a ceiling rather than a starting point to extend when the trace looks incomplete. An incomplete oxidation at −15 °C is a recoverable problem; a complete one at 4 °C may not be.
Thiolysis and the handoff to ligation
The acyl azide is not the ligation partner. Adding an external thiol such as 100 mM 4-mercaptophenylacetic acid (MPAA) and holding the mixture at pH 5.0–6.0 at room temperature for 10 min converts the azide quantitatively to the corresponding thioester [7]. Raising the pH then lets native chemical ligation proceed, and the oxidation, thioesterification and ligation are run in one pot [5].
MPAA is doing two jobs here, and it earned its place in the field doing only the second. Substituted thiophenols with pKa above 6 were found to best combine rapid, complete exchange with thioalkyl esters and effective leaving-group behaviour in the reaction with a cysteine peptide; MPAA gave an order-of-magnitude rate increase in model ligation studies and in the synthesis of the small protein OMTKY3 [8]. That comparison was between thiol catalysts in ligations of thioalkyl esters, not a measurement of acyl azide thiolysis. In the hydrazide protocol MPAA captures the azide and then stays in the tube as the ligation catalyst, which is convenient but means one reagent concentration governs two different steps.
Getting the hydrazide onto the resin
Three generations of hydrazine resin are in use. The first was made from a hydroxyl-functionalised resin — PAM for Boc chemistry, Wang for Fmoc — treated with p-nitrophenyl chloroformate and then hydrazine hydrate to give a hydrazine carbonyloxymethyl resin. A hydrazine 2-chlorotrityl chloride (2CTC) resin followed, with improved yield and purity of the crude peptide hydrazides, and then an Fmoc-hydrazine 2CTC resin with high loading capacity [7].
Storage is the practical failure mode. Trityl chloride resins loaded with free hydrazine are not shelf stable and are conventionally made at time of use; loading instead with 9-fluorenylmethyl carbazate gives resins whose loadings after storage under several common conditions were indistinguishable from day zero, while the chloride resin retained only 13.3% ± 0.3% of its initial loading, and the Fmoc group was shown to be required for that stability [6]. The same resins were used for the SPPS of a 40-mer [6]. If a hydrazide synthesis opens with poor first-residue loading, the resin's history is the first thing to check.
Hydrazides you did not make by SPPS
A C-terminal hydrazide does not have to come off a resin. Interrupting a thioester's acyl transfer with hydrazine as an external nucleophile generates a peptide hydrazide, and incorporating an α-hydroxy acid into the protein backbone followed by hydrazinolysis yields recombinant protein hydrazides; both approaches were used to prepare LC3 and HdeA [7]. The consequence is strategic rather than procedural: the same nitrite activation that serves a synthetic segment also serves an expressed one, which is what lets hydrazide chemistry support semisynthesis and not only total synthesis. The salicylaldehyde ester work below draws on hydrazides obtained either synthetically or recombinantly on exactly that basis [15].
Which residues can sit at the junction
The hydrazide-to-thioester conversion is described as fully compatible with all 20 amino acids located in the middle of the peptide sequence. The C-terminus is more restricted: Asp, Asn and Gln are not appropriate there, because they can give unproductive intramolecular cyclisation products [7]. Segment boundaries therefore need planning so that no segment ends in one of those three, and where a target forces it, the alternative acyl donors covered in the thioester guide are the fallback.
Stereochemistry at the activated carbon is reported to survive: the hydrazide-based transformation is described as racemisation-free [7], and the intramolecular version — lactamisation of fully unprotected peptides to give cyclic peptides and cyclic proteins under mild conditions — is reported to proceed in an epimerisation-free manner [9].
One pot, three or more segments
The assembly the hydrazide is built for runs N-to-C. Activate segment 1's hydrazide cold, thiolyse it, then add segment 2 carrying a free N-terminal cysteine and its own C-terminal hydrazide and raise the pH to ligate. The new C-terminal hydrazide is untouched by those conditions [4]. Re-acidify, re-chill, oxidise again, add segment 3. Each round consumes one masked thioester and regenerates the next.
Because each incoming segment is introduced only after the oxidation is over, this linear order does not require its N-terminal cysteine to be masked at all: the hydrazide is doing the job a thiazolidine would otherwise do. That is a property of the linear order, not of hydrazide chemistry in general — convergent assemblies, in which segments are ligated as blocks before being joined, do carry protecting groups [10].
Scaled up, this becomes a convergent strategy for assembling proteins from multiple peptide segments, demonstrated on the 142-residue model protein RpS25 built from six segments, with the hydrazide intermediates noted as cheap to make [10]. The published protocol for the approach walks through Fmoc solid-phase synthesis of the hydrazides and their ligation via in situ NaNO2 activation and thiolysis, and reports the total synthesis of 140-residue α-synuclein on milligram scale in roughly 30 working days; it does not involve hazardous HF, and is put forward as potentially helpful for targets that carry acid-sensitive modifications [5].
A one-pot variant that avoids isolating intermediate products employed a fast and selective arylboronate oxidation mediated by H2O2, and delivered five modified histone H3 and H4 analogues on multi-milligram scale with good homogeneity [11]. A more recent one-pot route replaces the thioester entirely: peptidyl N-acyl pyrazoles activated by azole reagents such as 3-methylpyrazole ligate directly with an N-terminal cysteine peptide without a thiol additive, which allowed an N-to-C sequential ligation combined with metal-free desulfurisation in a single pot for ubiquitin [12].
What the nitrite step actually damages
The residue that is reliably modified is cysteine. It is oxidised by nitrite during activation to S-nitroso cysteine, which is cleanly reduced back to cysteine in the presence of MPAA [7]. An S-nitroso adduct on a trace taken before thiol addition is therefore expected behaviour, not a failed run.
The real casualty is the protecting group, not the side chain. When ligations run C-to-N, the incoming segment's N-terminal cysteine is usually masked as a thiazolidine (Thz), and Thz is nitrosated during peptide hydrazide oxidation [13]. Two published fixes exist: Tbeoc protection of the Thz, which was found to overcome the instability [7]; and addition of L-thioproline as a protective agent that prevents nitrosation of the N-terminal thiazolidine during the oxidation, applied to the chemical synthesis of a 69-residue and a 113-residue protein [13].
The general caution is about oxidant excess. Introducing their non-oxidative route in 2018, Flood and co-workers stated that the oxidants these activations require are typically used in large excess and cannot be employed in concert with N-terminal thiazolidines, aryl amines, or other redox-sensitive residues; the Tbeoc and thioproline workarounds above are the counterweight to that generalisation. Their route instead uses stoichiometric acetylacetone to convert the C-terminal hydrazide to an acyl pyrazole, a mild acylating agent that exchanges efficiently with an aryl thiol and feeds directly into ligation, mild enough to be used through multiple sequential ligations without intermediate purification [14]. The reported conditions there are stoichiometric in the literal sense: the model hydrazide was exposed to acac at 2.5 equiv in 6 M guanidinium chloride with 2% v/v thiophenol, buffered at varying pH, with acidic conditions reported to markedly speed the reaction and pH 3 used as the working condition, and a subsequent round in a sequential assembly was run by acidifying the crude ligation mixture to pH 3–4, raising MPAA to 200 mM and adding 1 equiv of acac [14].
Be careful about how far you push that caution. The reviewed record does not report tryptophan or methionine damage under the standard window; the compatibility claim covers all 20 amino acids in mid-chain positions, with cysteine the residue that is modified and then restored [7]. What that licenses is narrower than a blanket guarantee for Trp- or Met-containing segments under any nitrite conditions, since the published tolerance was established inside a specific concentration, pH and temperature envelope. If oxidised side-chain adducts do appear on such a segment, the first hypothesis should be that the conditions drifted — a warm tube, a stale nitrite stock, an over-long oxidation — and the activation should be repeated before the second segment is spent. Where a sequence makes that risk unattractive, the non-oxidative routes are the alternative: the azole-activated acyl pyrazole work reports that sensitive residues including C-terminal Cys and Trp and internal Cys are well tolerated [12].
The same oxidation, a different ester
The acyl azide is not committed to thiols. Nitrite oxidation of peptide or protein C-terminal hydrazides — obtained either synthetically or recombinantly — followed by phenolysis with a dithiane-protected salicylaldehyde reagent gives SAL ester surrogates, which are then activated to the reactive SAL esters used in serine/threonine ligation and cysteine/penicillamine ligation. The scope was shown by N-terminal ubiquitination of Gadd45a, synthesis of ubiquitin-like protein 5 by a combined N-to-C NCL–STL route, and C-to-N semisynthesis of a myoglobin variant [15]. The hydrazide is better read as a branch point in the ligation toolbox than as a thioester stand-in.
What the read-outs license
An LC-MS trace after the cold oxidation tells you what the hydrazide has become at that moment, and the Curtius pathway gives byproducts chemically distinct from the acyl azide [7]. What it does not tell you is what the balance will be by the time the thiol goes in, which is why holding the temperature matters more than the timepoint you happened to sample.
The correct ligation product mass establishes that an amide formed at the junction. It does not establish that the stereocentre of the acyl segment's C-terminal residue survived, because epimers are isobaric. The racemisation-free and epimerisation-free descriptions of these transformations rest on dedicated studies [7, 9], not on the mass of the ligation product, and if epimerisation matters for your target it needs its own assay.
Finally, a one-pot three-segment yield is a single number covering six or more chemical steps. Nothing in the final trace apportions loss between an incomplete oxidation, a rearranged azide, an incomplete thiolysis and a slow ligation. Sampling after each oxidation and after each ligation is the only thing that turns that number into a diagnosis, and it is cheap compared with rebuilding a segment. The broader mechanistic account of what native chemical ligation and its extensions can and cannot do — thioester and selenoester surrogates, acyl-shift systems, auxiliaries and dechalcogenation — is set out at length elsewhere [16] and is worth reading alongside the native chemical ligation guide before committing to a segment plan.