A chain can come off the resin at the right mass, with an unremarkable coupling record, and still be useless because it will not dissolve. Solid-phase synthesis has been refined to a stage where efficient preparation of long and complex peptides is achievable, yet the post-synthesis handling of poorly soluble peptides often remains a significant hindrance to their purification and further use [1]. The scale is easy to underestimate until someone measures it: the crude of one SUMO-2[48–93] segment was barely soluble in HPLC solvents, at 0.05 mg/mL in 8:2:0.01 water/acetonitrile/TFA [2]. At that concentration the obstacle is not chromatographic selectivity. It is arithmetic on how much material you can inject.
A solubilising tag is the response that acts after the fact. You attach a short, highly polar sequence — oligolysine, oligoarginine, an oligo-acid or a short polyethylene glycol chain — through a linker chosen because it can be broken later under conditions the peptide will survive; you purify the tagged material as though it were an ordinary water-soluble peptide; you remove the tag at the end. That is a different intervention from the tools that suppress aggregation while the chain is being built. A pseudoproline inside the sequence disrupts the beta-sheet structures considered a source of intermolecular aggregation during chain elongation, and so increases solvation and coupling kinetics in peptide assembly [3]. A tag leaves the backbone alone and changes what the molecule does in water, at a position you choose.
Where the tag attaches, and why that decides what it is worth
The simplest attachment is at a terminus. A penta-lysine tag joined to the C-terminus through a base-labile linker solubilises the peptide for purification and is then removed by a brief base treatment to give the native sequence [1]. Terminal attachment is cheap: no residue in the target has to be a particular amino acid, and the linker chemistry is decoupled from the sequence. What it does not guarantee is proximity. The polar block sits at one end of a chain whose aggregating stretch may be far away, and that is the reason the field moved to side-chain and backbone anchors.
Side-chain attachment puts the tag on a residue you nominate. The Ddae linker (Fmoc-Ddae-OH) is incorporated on-resin at a lysine ε-amine, the solubilising sequence is elongated from it by ordinary Fmoc-SPPS, and the whole assembly is removed in solution after ligation-based assembly using mild aqueous hydrazine [4]. Its successor, Fmoc-Ddap-OH, is amine-reactive, has improved stability compared with its predecessor — which matters for multi-step syntheses with long handling times — and is cleaved by a hydroxylamine protocol that greatly accelerates removal [5]. Where no lysine is expendable, Fmoc-Glu(AlHx)-OH does the same job at a glutamate side chain, is stable to a wide range of chemical protein synthesis conditions, and is removed by palladium-catalysed transfer under aqueous conditions [6].
Cysteine is the other common anchor, and it is used from two directions. A tag on a Cys side chain can be taken off with palladium, and that removal can share a pot with the ligation and the desulfurisation, which is how histone protein H4 was prepared from two fragments [7]. The trityl tags reverse the order in the workflow: the Trt-K10 tag, an oligo-Lys sequence on a trityl anchor, is introduced from the trityl alcohol reagent Trt(OH)-K10 selectively to a Cys side-chain thiol of an already deprotected peptide under acidic conditions, so it goes on after global deprotection rather than during chain assembly [8]. A tag can also be hung on the N-terminal cysteine of a ligation partner as a mixed disulfide, using Boc-Cys(Npys)-OH to direct the disulfide formation; it is then cleaved in the course of the ligation itself, because standard native chemical ligation conditions include a reducing agent such as TCEP and a large excess of an aromatic thiol such as MPAA [2].
The most aggressive placement is the backbone. Removable backbone modification installs an activated O-to-N acyl transfer auxiliary bearing an Arg4 tag at a backbone amide, using 4-methoxy-5-nitrosalicylaldehyde by reductive amination during Fmoc-SPPS, and the group can be placed at any primary amino acid; the tagged membrane-spanning segments then behave like water-soluble peptides through purification, ligation and mass characterisation, and the tags come off at the end with TFA cocktails [9]. Because the position is unconstrained, the tag can be put where the trouble is: Arg4-tagged backbone modification groups were installed into the hydrophobic region of amyloid-β, improving the handling of purification, ligation and mass characterisation for the longer Aβ48 and Aβ49 chains [10].
Which polar sequence, and how much of it
Charge does most of the work, but not all of it, and the published comparisons are small. In one head-to-head on a hydrophobic model peptide, tags attached at an internal Cys through a disulfide linker changed the isolated yield of the peptide salicylaldehyde ester after reversed-phase purification from 5.8% with no tag to 24.1% with Ac-Lys10, 19.4% with Ac-Arg4, 14.6% with MeO-PEG3 and 20.8% with MeO-PEG7 [11]. Two things are worth reading out of that series. Every tag helped, including the uncharged PEG ones; and the ranking came from a single substrate in a single reaction, so it is not a general ordering.
A second comparison put Lys, Arg, Glu, Asn, Nε-trimethyl-Lys and Cys-sulfonate tags on a common trityl anchor and tested them on two hydrophobic model peptides; the trimethyl-lysine tag showed superior solubilising ability among those evaluated [12]. Trimethyl-lysine carries a permanent quaternary ammonium charge rather than a titratable one, which is the obvious reason to reach for it when the peptide has to be handled across a pH range — but again, two model peptides. The presence of Glu- and Asn-based tags in that panel is the practical point for anyone whose target is already strongly basic: the sign of the tag is a decision, not a default.
Polyethylene glycol is the option when charge itself is unwanted. A short, monodisperse oligoethylene glycol-containing photocleavable lysine tag was used to prepare a modified Aβ42 with increased solubility and decreased propensity to aggregate in aqueous media, and the tag was removed by UV irradiation to give Aβ42 in high purity and yield [13]. That construction also makes a point the yield tables do not: a tag can lower aggregation propensity, not merely improve handling.
Taking it off again
Cleavage chemistry is where a tag is chosen or rejected, because removal has to be orthogonal to everything the molecule still has to survive. The families in current use are few. Nucleophilic cleavage breaks the dimedone-type linkers: 1 M hydrazine at pH 7.5 removed the Lys6–Ddae group from a C20 model peptide with a half-life of about 1 h [4], and the hydroxylamine protocol on the Ddap linker was introduced to speed that step up [5]. Palladium removes the tag from a Glu side chain under aqueous conditions [6] and from a Cys side chain [7]. Acid removes the trityl anchors with a standard TFA cocktail [8] and the backbone modification groups with TFA cocktails as well [9]. Light removes the photocleavable linkers [13]. Reduction removes disulfide-linked tags: 0.2 M TCEP in pH ~7 phosphate buffer containing 6 M guanidinium chloride for 30 min to 1 h, or, alternatively, during the free-radical desulfurisation step [11]. In the neatest case the removal is not a separate operation at all, the mixed disulfide being cleaved during the ligation after the tag has already done its work in purification, characterisation and handling [2].
Traceless is a claim about the product, not about the effort. Ddap is a traceless cleavable linker, and the solubilising tag it carries is selectively removed to generate native peptide [5]; Glu(AlHx) is likewise a traceless linker, only temporarily attached to a Glu side chain [6]; removal of the backbone modification groups with TFA affords the native protein [9]. What is not traceless is the requirement that put the tag there. A Cys-anchored tag needs a cysteine at the attachment site, a Lys-anchored tag needs a spare lysine, the glutamate linker needs a glutamate. Where the native sequence offers none, a residue has to be engineered in, and the question becomes whether that substitution is erased at the desulfurisation step the one-pot route above already includes [7], or left in the product as a scar. Decide that before you decide the linker.
What the tag has to survive: acid, thiols and oxidation
Compatibility runs in two directions, and the sign is easy to get wrong. An acid-labile tag cannot be present during an acid step you still need. The trityl tags are detached by a standard TFA cocktail, which is precisely why they are installed on the deprotected peptide after resin cleavage instead of being carried through it [8]. The backbone modification groups resolve the same conflict by making acid stability switchable: the free phenol is acylated so the group survives the TFA deprotection step, the acyl phenyl ester is then quantitatively hydrolysed in neutral aqueous solution, and the liberated phenol raises the electron density of the ring enough to make the group labile to TFA on the next exposure [14]. The Ddae and Ddap linkers sidestep the conflict altogether by not being acid-cleaved: installed on resin, carried through the cleavage, removed later in solution by a nucleophile [4, 5].
Thiols and disulfides are the second axis, and here reducibility is a liability as often as a feature. A tag joined through a disulfide is detached by TCEP [11], so on the same reasoning it should not be expected to survive an oxidative folding step, and it cannot be removed without also reducing any native disulfide you intend to keep; the same property is what makes its removal free inside a ligation, where TCEP and MPAA are present anyway [2]. One non-obvious observation from that work is worth carrying forward: the disulfide linkage was stable under Cys/Pen ligation conditions and did not exchange with the N-terminal Cys residue, which the authors reported as surprising [11]. When the target does carry disulfides, the tag must be inert to the oxidation instead. A penta-lysine tag on a base-labile C-terminal linker was carried through regioselective disulfide bond formation between two separately synthesised chains, and only removed from the poorly soluble A-chain once that combination was complete; that route delivered the final detagged product in nearly 10% overall yield [1].
The purification window
Chromatography is the reason most of this machinery exists. Untagged, the SUMO-2 segment exhibited anomalous HPLC behaviour; the hexalysine-tagged version gave a clean and soluble crude product that was purified by standard HPLC [2]. A tag adds charge and hydrophilicity, so on a reversed-phase column expect the tagged species to elute earlier than its untagged counterpart, moving the target away from the late, shallow, low-recovery end of the gradient. Judge the outcome on recovery rather than on peak shape: the isolated-yield series quoted above [11] is a purification result, not a solubility measurement, and it is the quantity that tells you whether the tag paid for itself.
There is a second chromatographic use that is easy to miss. Because the trityl tag is added to a free thiol on a deprotected peptide, it can be applied directly to the reaction mixture of a thiol-additive-free ligation in a one-pot manner, for the sole purpose of making the ligation product isolable by HPLC [8]. Used that way the tag is not a synthesis aid; it is a handle for one purification.
A tag against prevention
The alternatives covered elsewhere in this library act during assembly rather than after it. Pseudoprolines are reversible protecting groups, and the structure-disrupting mechanism described at the top of this article operates while the chain is still on the resin [3]. The depsipeptide route swaps one backbone amide for an ester: a single isopeptide insertion into a 42-residue chain was, the authors suggest, enough to suppress the unfavourable nature of its difficult sequence, and the deprotected crude was easily purified by HPLC because it eluted narrowly with reasonable water solubility [15]. Backbone amide protection with the 2-hydroxy-4-methoxybenzyl group inhibits the aggregation caused by hydrogen-bonded interchain association, and in a side-by-side comparison on a difficult peptide, backbone protection and optimised Boc chemistry both delivered the product in excellent yield and purity [16].
Each of those makes a demand on the sequence and on the schedule. A pseudoproline needs a Ser, Thr or Cys to build the ring on [3]; the O-acyl isopeptide needs a side-chain hydroxyl to carry the ester; Hmb needs a backbone amide you are willing to protect [16]. All three must be decided before the first coupling. A tag need not be: the trityl approach is applied to material that already exists, after deprotection [8]. That asymmetry, rather than any yield comparison, is the honest case for keeping a tag in reserve even on a route that already uses the preventive tools — it is the only one of the four you can still choose once the crude is in front of you and insoluble.
What a rescue does and does not license
A tag that delivers pure tagged peptide has demonstrated one thing: that the tagged construct was tractable. The detagged peptide is a different molecule with the solubility it always had, and it still has to be recovered from the removal step and characterised on its own terms. Treat the published cleavage conditions accordingly. The timed hydrazinolysis quoted earlier was measured on one model peptide with one linker [4]; it is a starting point for your linker, not a specification for your peptide, and the useful experiment is a time course on your own material with the removal monitored to completion. For the same reason, the routes that fold the removal into a step already in the schedule — the one-pot sequence described above [7], or a tag that disappears during the ligation [2] — buy more than their chemistry suggests. They delete a purification at the point in the route where the material is least able to withstand one.