A Lys-Ala isopeptide bond links the carboxyl group of alanine to the side-chain nitrogen of lysine, instead of its main amino group. That makes a branched molecule. It has the same mass as ordinary Ala-Lys, so mass alone cannot tell the two apart.
- An isopeptide bond is ordinary amide chemistry attached to a side chain rather than to an α-amino or α-carboxyl group.
- The lysine side-chain nitrogen appears in the literature as ε, as N6 in IUPAC-IUB locant style, and as NZ in structural files.
- Nε-alanyl-lysine and the dipeptide Ala-Lys share a molecular formula and monoisotopic mass, so precursor mass alone cannot separate them.
- A reproducible tryptic missed cleavage at one lysine is the standard evidence that its side chain is acylated.
- The peptidoglycan crosslink formed between D-alanine and lysine is the isopeptide bond that β-lactam antibiotics prevent.
- Asparagine deamidation can produce an isoaspartyl backbone that carries no mass difference from aspartate and needs ETD or an isoAsp-specific assay to detect.
What this guide covers
A peptide bond in the ordinary sense joins the α-carboxyl of one residue to the α-amino group of the next. An isopeptide bond is the same amide chemistry hung somewhere else on the residue: a carboxyl condensed with an amine that is not the α-amino group, or an α-amino group condensed with a carboxyl that is not the α-carboxyl. The atoms sit in the same C=O and N-H arrangement as any amide. What changes is where the connection attaches, and therefore what shape the molecule takes and which enzymes will touch it.
Lysine is where this comes up most, because its side chain ends in a primary amine four carbons out from the α-carbon. Biochemical usage calls that nitrogen the ε-amino group. IUPAC-IUB nomenclature gives it the locant N6, counting from the carboxyl carbon. Structural files label the atom NZ, so crystallographers tend to say N-zeta. Three names for one nitrogen, and a document only needs to say which convention it is using.
A Lys-Ala isopeptide bond is the amide between the carboxyl carbon of alanine and that side-chain nitrogen. Spelled out, it is Nε-alanyl-lysine, or N6-alanyl-lysine in locant style. The ordinary dipeptide built from the same two residues, Ala-Lys, puts the alanine carboxyl on the lysine α-amino group instead. Same two amino acids, same molecular formula, same monoisotopic mass to five decimal places, different molecule. The free amine ends up somewhere else, and almost every practical difference follows from that.
This guide covers where the Lys-Ala linkage occurs in nature and in engineered proteins, how its bonding, charge and geometry differ from a backbone amide, how to tell the isopeptide from its α-linked isomer when the mass is identical, how to make one on purpose, which isopeptide bonds form in stored material without being asked, and what a certificate for a branched peptide should say. It lists the red flags and ends with a bottom line.
- Precursor mass: composition only, isomers identical
- Tryptic digest: a blocked lysine shows as a missed cleavage
- MS/MS or ETD: places the branch junction
- HMBC NMR: settles the bond outright, given milligram material
- Amino acid analysis: returns the same answer for both, so skip it
Where does a Lys-Ala isopeptide bond occur?
The most consequential example sits in the bacterial cell wall. Peptidoglycan stem peptides in many Gram-positive organisms read L-Ala, D-Glu, L-Lys, D-Ala, D-Ala, and a transpeptidase removes the terminal D-alanine and joins the carboxyl of the remaining D-Ala to the ε-nitrogen of the lysine on a neighboring strand. In Staphylococcus aureus that crosslink runs through a pentaglycine bridge rather than directly. In Escherichia coli, meso-diaminopimelate stands where lysine would be and the crosslink lands on its side-chain amine instead, which is the same isopeptide logic on a different residue. β-lactams acylate the active-site serine of that transpeptidase, which is why the crosslink is a drug target rather than a curiosity.
Two details in that stem peptide are easy to miss. The glutamate is already linked onward through its side-chain carboxyl, so the stem carries an isopeptide bond before any crosslinking happens. And the alanine donating the carboxyl is the D enantiomer, which contributes to the wall's resistance to ordinary proteases quite apart from the branch geometry.
| Linkage | System | Carboxyl donor | Amine acceptor | Note |
|---|---|---|---|---|
| D-Ala to L-Lys | Peptidoglycan crosslink | D-Ala α-carboxyl | Lys N6 | Direct or bridged; formed by the transpeptidase β-lactams inhibit |
| Gly to Lys | Ubiquitin and ubiquitin-like conjugation | C-terminal Gly carboxyl | Substrate Lys N6 | Reversible in vivo; cleaved by deubiquitinating enzymes |
| Gln to Lys | Transglutaminase crosslinks | Gln side-chain carboxamide | Lys N6 | Nε-(γ-glutamyl)lysine; off the backbone at both ends |
| Asn or Asp to Lys | Gram-positive pilin domains; engineered tag-catcher pairs | Side-chain carbonyl | Lys N6 | Forms spontaneously in a buried site; exceptionally stable once made |
| β-Asp (isoaspartate) | Deamidation product in stored material | Asp side-chain carboxyl | Following residue α-amino | Backbone rerouted through the side chain, with no enzyme involved |
The engineered tag-catcher pairs on that list are why the chemistry turns up in protein engineering catalogs: two domains that find their own lysine and asparagine and weld themselves together give a covalent handle no affinity pair can match.
How do bonding, charge and geometry differ from a backbone amide?
The electronics are unremarkable, and that is the point. Amide resonance flattens the C-N unit, restricts rotation about it, and makes the nitrogen a poor base. An isopeptide bond behaves the same way. It differs from a backbone amide in the flexibility around it: the lysine side chain contributes four rotatable bonds between the α-carbon and the amide nitrogen, so a branch point is conformationally looser than the same junction on a backbone. Modeling a branched peptide as a rigid tee gets the topology right and the dynamics wrong.
Charge matters more than it first appears. The lysine side-chain amine has a pKa near 10.5 as the free amino acid, so at neutral pH it is protonated and a weak nucleophile. Acylating it costs the molecule a positive charge at physiological pH and shifts the isoelectric point accordingly, which is why ion-exchange behavior often separates a branched species from its linear isomer when reversed-phase does not.
The protonation state also explains how biology gets the reaction to run at all. Enzymatic routes use an activated acyl donor: a thioester in the ubiquitin cascade, the D-Ala-D-Ala terminus in transpeptidation. The spontaneous pilin isopeptides do it differently, burying the lysine in a low-dielectric pocket beside an acidic residue that depresses its pKa and leaves a fraction of the amine deprotonated and reactive. Neither route is available to a peptide sitting in an open aqueous vial, which is the useful takeaway for anyone worrying about a stock solution rearranging itself.
How do you tell the isopeptide from the α-peptide isomer?
This is the part that costs laboratories real time. The two isomers are the same formula, so accurate-mass measurement establishes composition and stops there. Total acid hydrolysis followed by amino acid analysis is worse than useless for the question: it returns identical alanine and lysine, having destroyed the only structural information in the sample.
| Method | What it answers | What it does not answer |
|---|---|---|
| High-resolution MS, precursor only | Elemental composition | Nothing about connectivity; isomers are indistinguishable |
| MS/MS, collisional fragmentation | Where the linear series breaks down and which fragments carry the branch | Ambiguous when the branch is short and fragmentation is sparse |
| Electron-transfer dissociation | Junction placement with labile groups preserved | Needs multiply charged precursors and adequate signal |
| Tryptic digest pattern | Whether a specific lysine is blocked, seen as a reproducible missed cleavage | Which acyl group is on it, without MS/MS to follow |
| Edman sequencing or aminopeptidase ladder | Stalls or splits at a branch point, flagging it | Poor at resolving what the branch is |
| NMR, HMBC-type correlation | Unambiguous bond placement from side-chain protons to the acyl carbonyl | Needs milligram material and a clean sample |
| Reversed-phase HPLC | Often resolves the isomers; gives quantitation once identity is assigned | Co-elution happens, so a purity number is not an identity claim |
The tryptic signature is how the field maps isopeptides at scale. Trypsin cleaves after lysine and arginine, and a lysine whose side-chain amine is acylated is not cleaved. A missed cleavage that reproduces at one specific position, with a fragment carrying extra mass, is the standard evidence for a branch. Ubiquitination site mapping runs on that logic, reading the +114 Da diglycine remnant trypsin leaves on the modified lysine.
NMR settles it outright when there is enough material. Acylation shifts the adjacent methylene protons downfield, and a heteronuclear multiple-bond correlation from those protons to the alanine carbonyl carbon places the bond with no inference required. Slow, sample-hungry, and the method that ends arguments.
How to make a Lys-Ala isopeptide on purpose: step by step
Synthesizing a defined Lys-Ala isopeptide is a protecting-group problem before it is a coupling problem. Standard Fmoc chemistry protects the lysine side chain with Boc, which comes off in the same acid treatment that cleaves the peptide from resin, so it is no help when the side-chain amine has to be acylated mid-synthesis. The order of work is:
- Protect the lysine side chain with an orthogonal handle rather than Boc: Alloc removed with a palladium catalyst and a scavenger, ivDde or Dde removed with hydrazine, or Mtt or Mmt removed with very dilute acid that leaves tBu-class groups standing.
- Unmask that one nitrogen while every other protecting group stays in place.
- Couple the alanine building block onto it, using activation that will not racemize: carbodiimide with an additive such as HOBt or an oxime, or a preformed anhydride, is the conservative choice.
- Carry on with the chain.
- Read the crude for what it contains: the target, the α-linked regioisomer and the deletion product, and treat regiochemistry as a claim that needs its own analytical basis.
Two things go wrong in practice. Branch points are sterically crowded, so the coupling onto a freed side-chain amine is frequently incomplete, and the deletion sequence that results has a mass close enough to the target to hide under a shoulder. And activation chemistry that racemizes will hand back an epimer at the alanine that no amount of chromatography will separate cleanly, which is why the choice of activation in step 3 is not a detail.
The consequence for anyone buying a branched peptide rather than making one: the crude contains the target, the α-linked regioisomer, and the deletion product, and a single reversed-phase gradient may resolve two of the three. Regiochemistry is a claim that needs its own analytical basis, stated as such.
Which isopeptide bonds form without being asked?
Peptides also acquire isopeptide backbone by accident, which is the part of this topic that touches routine storage. Asparagine deamidates through a cyclic succinimide intermediate, and when that ring opens it can give either the ordinary α-linked aspartate or the β-linked isoaspartate, in which the side-chain carboxyl has been drafted into the backbone and the original α-carboxyl now hangs off it. Reported ratios favor the isoaspartyl product in flexible sequences, though the number is strongly sequence dependent and not worth quoting as a constant.
The analytical trap is sharp. Asn to Asp adds roughly one dalton, which a good instrument sees. Asp to isoAsp adds nothing at all, so the rearranged molecule is invisible to mass and often nearly invisible on a reversed-phase gradient. Electron-transfer dissociation resolves it, and the enzyme protein L-isoaspartyl methyltransferase recognizes isoaspartate specifically enough to serve as an assay. Neither is standard on a supplier's certificate.
Lysine plays a second role here. A side-chain amine can attack a succinimide or an activated ester on a neighboring molecule, giving a covalent dimer that reads as a high-mass shoulder on size-exclusion analysis rather than as anything the primary purity number would flag. Rates are low in a cold, dry, sealed solid and rise once water is present, which is the same argument the storage and stability guide makes about hydrolysis and deamidation generally. Diluent choice belongs in the same conversation, since a prepared solution is where all of this chemistry becomes possible; the diluent guide covers what the water itself contributes.
What should a certificate say about a branched peptide?
Read an ordinary certificate of analysis against the question this article started with and the gap is obvious. Identity is usually reported as an observed mass matching theory, which for isomers settles composition and no more. Purity is usually area percent from a reversed-phase run at 214 nm, which quantifies whatever eluted where the standard elutes and says nothing about which isomer that peak is. Both numbers can be entirely honest and still leave regiochemistry unestablished.
What closes the gap is an orthogonal identity test, which is the principle ICH Q6B states plainly for biotechnological products: identity should rest on a method that discriminates structure, not only mass or retention. For a branched sequence that means MS/MS with the junction assigned, a peptide map in the style of the USP general chapter on peptide mapping, or NMR on a reference lot. Ask which one was run, and on which lot. A certificate naming a method and a lot is a different document from one reporting two numbers with nothing behind them, and the difference is visible before any money moves. Our own quality standard sets out which tests we require and which we do not claim.
Where no compendial monograph exists for a given branched peptide, and for most of them none does, the honest position is that identity rests on the supplier's own analytical package. That is a reason to read the package, not a reason to treat the number as unknowable.
FOR LABORATORY AND IN-VITRO RESEARCH USE ONLY. NOT FOR HUMAN OR ANIMAL CONSUMPTION. NOT FOR PERSONAL, MEDICAL, DIAGNOSTIC, THERAPEUTIC, OR RECREATIONAL USE.
Red flags on a branched-peptide certificate
Any one of these is a question to put to the supplier before money moves.
- Identity reported only as an observed mass matching theory. For isomers that settles composition and no more.
- Purity reported only as area percent from a reversed-phase run, which says nothing about which isomer the peak is.
- No orthogonal identity method named: no MS/MS with the junction assigned, no peptide map, no NMR on a reference lot.
- A method named with no lot beside it.
- Amino acid analysis after total hydrolysis offered as proof of structure; it returns identical alanine and lysine from either isomer.
Bottom line
A Lys-Ala isopeptide bond is ordinary amide chemistry hung on the lysine side chain instead of the backbone, and the molecule it makes weighs exactly the same as the linear dipeptide built from the same two residues. Mass settles composition and stops. Connectivity takes MS/MS with the junction assigned, a peptide map, or NMR, and a certificate for a branched peptide should name which one was run and on which lot. Where no monograph exists, and for most branched peptides none does, that analytical package is the identity claim, so read it.
Common questions
What is a Lys-Ala isopeptide bond?
How is an isopeptide bond different from a normal peptide bond?
Can mass spectrometry distinguish the isopeptide from the α-peptide?
Where do Lys-Ala isopeptide bonds occur naturally?
Do isopeptide bonds form spontaneously in a stored peptide?
How is a defined isopeptide bond made synthetically?
Related guides: peptide chemistry basics
Sources
- IUPAC-IUB Joint Commission on Biochemical Nomenclature, recommendations on amino acid and peptide nomenclature. Basis for side-chain nitrogen locants (N6 for lysine) and for the definition of an isopeptide linkage as an amide outside the α-amino and α-carboxyl positions.
- Vollmer, Blanot and de Pedro, FEMS Microbiology Reviews, 2008. Peptidoglycan structure and architecture; supports the stem-peptide composition, the D-Ala to L-Lys crosslink, the pentaglycine bridge in Staphylococcus aureus, and the meso-diaminopimelate variant.
- Kang, Middleditch, Proft and Baker. Stabilizing isopeptide bonds revealed in Gram-positive bacterial pilus structure. Science 318(5856):1625-1628 . Basis for the isopeptide-bond description: the crystal structure that first showed spontaneous intramolecular lysine-asparagine isopeptide bonds in a Gram-positive pilin, and their stabilizing role.
- ICH Q6B and the USP general chapter on peptide mapping for biotechnology-derived articles. Supports the requirement that identity rest on a structurally discriminating method rather than on mass or retention time alone.
