The Peptide Bond and Its Formation

The chemistry of peptide bond formation is not a historical curiosity confined to textbooks — it is the working basis of major swaths of modern pharmaceutical and biotechnology industry. Therapeutic peptides such as insulin analogs, GLP-1 receptor agonists (semaglutide, used in diabetes and obesity treatment), and the antibiotic vancomycin are manufactured using the very activation chemistry derived here: carbodiimide- or active-ester-mediated coupling on solid-phase resins, exactly the DIC/HOBt logic of Example 2, scaled to industrial reactors. Antibody-drug conjugates and peptide-based diagnostics rely on the same nucleophilic acyl substitution mechanism to attach payloads via amide linkages to lysine side chains. Inside the cell, the ribosome's peptidyl transferase center is now understood, from high-resolution ribosome crystal structures, to be composed entirely of RNA — it is a ribozyme, not a protein enzyme — that positions the aminoacyl-tRNA's amine and the peptidyl-tRNA's ester carbonyl in precise proximity and orientation, dramatically raising the effective molarity of the reactants and providing modest general base catalysis, while the actual chemistry executed is exactly the tetrahedral-intermediate mechanism derived above. Antibiotics such as chloramphenicol and the macrolides (erythromycin and relatives) work by binding near this active site and blocking peptide bond formation or the nascent chain's exit, making this reaction mechanism a direct and validated drug target. On the reverse side, proteases — serine proteases like chymotrypsin and trypsin, cysteine proteases like the caspases central to programmed cell death, aspartic proteases like HIV protease, and metalloproteases like the matrix metalloproteinases involved in tissue remodeling and cancer metastasis — are all built to accelerate the thermodynamically favorable but kinetically slow hydrolysis derived in this article. HIV protease inhibitors, a mainstay of antiretroviral therapy, are designed as transition-state mimics of the very tetrahedral intermediate shown in the mechanism figures, binding far more tightly than either substrate or product because they resemble the fleeting high-energy species along the reaction coordinate. Peptide bond chemistry, in short, sits at the intersection of protein synthesis, protein degradation, antibiotic mechanism, and rational drug design — a single reaction mechanism recurring across an enormous span of biology and medicine.

Ask the tutor: How does the partial double-bond character of the peptide bond, once formed, restrict backbone rotation and give rise to the cis/*trans* isomer question at proline?

An amino acid in solution near physiological pH exists as a zwitterion: the α-carboxyl group is deprotonated (–COO⁻) and the α-amino group is protonated (–NH₃⁺). Neither of these two "obvious" forms can react to make an amide directly. A carboxylate anion is a terrible electrophile — its negative charge repels incoming nucleophiles, and there is no good leaving group attached. A protonated ammonium ion is not nucleophilic at all, because its nitrogen lone pair is tied up holding the extra proton. For a peptide bond to form, two things must happen: the amine must be un-protonated (free lone pair available), and the carbonyl carbon must bear a leaving group that can depart once the tetrahedral intermediate collapses. This is exactly why the reaction is called a condensation: two molecules join with loss of a small one, here water (or, in the activated pathway used by the cell, an alcohol or thiol leaving group derived from tRNA). Chemically it belongs to the broader family of nucleophilic acyl substitution reactions, the same mechanistic class as ester hydrolysis, transesterification, and amide hydrolysis, all of which proceed through a tetrahedral alkoxide intermediate at the carbonyl carbon. The product amide bond has unusual electronic structure. The nitrogen lone pair is delocalized into the carbonyl π system, giving the C–N bond roughly 40% double-bond character. This is only briefly previewed here — the geometric and energetic consequences (bond length ~1.32 Å, planarity of the six atoms around the bond, the cis/trans isomerism question) belong to their own treatment — but it is essential to keep in mind while working through the mechanism below, because that same resonance is why* the carbonyl carbon of an amide is a much weaker electrophile than that of an activated ester or anhydride. Nature exploits this asymmetry: it makes the bond from a reactive precursor, and once formed, the amide is comparatively inert, which is exactly the stability a structural polymer needs.

Once activation is achieved — whichever chemistry supplies it — the mechanism proceeds by the standard nucleophilic acyl substitution pathway: nucleophilic addition to form a tetrahedral intermediate, then elimination of the leaving group to re-form the carbonyl π bond, now as part of the new amide.

The essential lesson visible in both figures: the reaction's fate is decided entirely by what X is. If X = OH (free carboxylic acid), the intermediate has nowhere favorable to go and the forward reaction is slow and thermodynamically uninviting in water. If X is a good leaving group — an activated ester oxygen, an anhydride oxygen, or (biologically) the 3′-oxygen of tRNA in an aminoacyl-tRNA ester — the tetrahedral intermediate collapses readily, ejecting X⁻ and forming the amide.

The overall condensation reaction. For two amino acids (or an amino-acid residue and a growing chain) joining to form one peptide bond with loss of water:

where:

• R₁, R₂ are the side chains of the two residues,
• the –CO–NH– unit formed is the peptide (amide) bond,
• water is the condensation by-product (the reverse reaction, hydrolysis, consumes one water per bond broken).

Standard free energy of hydrolysis. The reaction is more conventionally tabulated in the hydrolytic (spontaneous) direction:

where:

• $R = 8.314\ \text{J mol}^{-1}\text{K}^{-1}$,
• $T$ is absolute temperature (K),
• $K_{\text{hyd}} = \dfrac{[\text{amino acid 1}][\text{amino acid 2}]}{[\text{dipeptide}]}$ at unit water activity (standard convention in dilute aqueous solution),
• $\Delta G^{\circ\prime}_{\text{hydrolysis}}$ is typically −10 to −21 kJ mol⁻¹ for peptide bonds (favoring hydrolysis), so the synthesis direction costs +10 to +21 kJ/mol.

Reading: because $\Delta G^{\circ\prime}_{\text{hydrolysis}}$ is negative, $K_{\text{hyd}} \gg 1$: at equilibrium in water, free amino acids are strongly favored over the intact dipeptide. This is the thermodynamic fact that makes protein synthesis a genuinely uphill process requiring metabolic energy input, and simultaneously the fact that makes proteases (which merely catalyze approach to this already-favorable equilibrium) thermodynamically "downhill" reactions needing no ATP.

Coupling to ATP: the activation step. The cell circumvents the unfavorable direct condensation by first activating the amino acid's carboxyl group as a mixed anhydride with AMP, then transferring it to tRNA:

where:

• the first step is catalyzed by aminoacyl-tRNA synthetase and uses ATP's β,γ-phosphoanhydride bond,
• pyrophosphate (PPi) is subsequently hydrolyzed by pyrophosphatase, pulling the whole activation essentially irreversible ($\Delta G^{\circ\prime}_{\text{PPi hydrolysis}} \approx -19\ \text{kJ/mol}$),
• the net cost is equivalent to two high-energy phosphoanhydride bonds per peptide bond formed (ATP → AMP + 2 Pi, not ATP → ADP + Pi).

Reading: the ribosome does not need to overcome the unfavorable +10 to +21 kJ/mol condensation free energy directly. Instead, the ester bond in aminoacyl-tRNA is itself a high-energy bond (its own hydrolysis is exergonic by roughly −25 to −35 kJ/mol), so that when the ribosomal peptidyl transferase center lets the peptidyl-tRNA's amine (on the incoming aminoacyl-tRNA) attack the ester carbonyl of the peptidyl-tRNA, the overall step is strongly exergonic even though a "plain" amide bond has now been formed. Energy paid at the activation step is spent to purchase a favorable, essentially irreversible bond-forming step at the ribosome.

Example 1 — Foundational: forming a specific dipeptide and naming the bond. Condense alanine (Ala) and serine (Ser), with alanine's carboxyl group reacting with serine's amino group, to form Ala-Ser. Alanine: $\text{H}_2\text{N–CH(CH}_3\text{)–COOH}$. Serine: $\text{H}_2\text{N–CH(CH}_2\text{OH)–COOH}$. The mechanism: serine's amino nitrogen lone pair attacks the carbonyl carbon of alanine's (activated, in vivo) carboxyl group. A tetrahedral intermediate forms at that carbon, bearing the incoming N, the original two oxygens, and the R-group carbon. Collapse expels the leaving group (in vivo, the tRNA oxygen; in a free hydrolysis-driven equilibrium picture, effectively OH⁻/H₂O), regenerating the carbonyl as an amide C=O. Product:

This is Ala-Ser: by convention the residue contributing the free α-amino group (Ala) is the N-terminal residue and is written and named first; the residue contributing the free α-carboxyl group (Ser) is the C-terminal residue, written last. One water molecule (H₂O) is released, mass balance: $\text{Ala (89.09 g/mol)} + \text{Ser (105.09 g/mol)} - \text{H}_2\text{O (18.02 g/mol)} = 176.16\ \text{g/mol}$ for Ala-Ser, matching the tabulated molecular weight of the dipeptide. The new C–N bond is the peptide bond; it is singular — Ser-Ala (serine's carboxyl condensing with alanine's amine) would be a different, distinct dipeptide, underscoring that peptide bond formation is directional, not symmetric.

Example 2 — Applied: solid-phase peptide synthesis, an activated-ester coupling with DIC/HOBt. In Fmoc solid-phase peptide synthesis (SPPS), a resin-bound amino acid's free α-amine must be coupled to the next Fmoc-protected amino acid's free carboxylic acid. Direct coupling is impractically slow (as shown in Derivation 2, a free carboxylate is a poor electrophile with a terrible leaving group), so the incoming amino acid's –COOH is first activated using a carbodiimide such N,N′-diisopropylcarbodiimide (DIC), typically together with an additive like HOBt (1-hydroxybenzotriazole) that forms a more selective active ester and suppresses racemization. Mechanistically: (i) the carboxylate oxygen attacks the electrophilic central carbon of the carbodiimide (R–N=C=N–R′), forming an O-acylisourea — this species has the original carbonyl carbon now attached to a genuinely excellent leaving group (the isourea oxygen); (ii) HOBt's nucleophilic nitrogen/oxygen displaces the isourea leaving group, forming the HOBt active ester (a much less basic, better-behaved leaving group than the isourea, pKa of HOBt-OH ≈ 4.6, an excellent leaving group by the analysis in Derivation 2); (iii) the resin-bound free amine then attacks this active ester's carbonyl carbon via the same tetrahedral-intermediate mechanism, expelling HOBt⁻ and forming the new peptide bond. Quantitatively, if a coupling is run with 4 equivalents of activated amino acid relative to 1 equivalent of resin-bound amine, and the reaction reaches 99.5% completion (typical SPPS target, since incomplete couplings compound multiplicatively over a synthesis), then for a 20-residue peptide requiring 19 couplings, overall yield is:

so about 90.9% of chains are full-length and error-free even at a per-step efficiency most chemists would call "excellent" — illustrating why each individual coupling must be driven essentially to completion (activation chemistry, not thermodynamics of a free acid + amine, is what makes 99.5%-per-step feasible at all).

Example 3 — Exam/competition level: equilibrium extent of tripeptide formation with a shared water pool. Consider assembling a tripeptide Gly-Gly-Gly from three glycine molecules via two peptide bonds, entirely by uncatalyzed condensation in aqueous solution (no ATP coupling), each bond having the same representative $\Delta G^{\circ\prime}_{\text{hydrolysis}} = -17\ \text{kJ/mol}$ used above, at $T = 310\ \text{K}$ (body temperature). Estimate the equilibrium fraction of glycine converted into triglycine if the total glycine concentration (summed over all species, on a monomer-unit basis) is 0.050 M.

Step 1. Recompute $K_{\text{hyd}}$ per bond at 310 K (not 298 K as before):

Step 2. For two peptide bonds broken sequentially (triglycine → diglycine + glycine → 3 glycine), treat each bond hydrolysis with the same $K_{\text{hyd}}$ (a simplifying assumption ignoring small differences between internal and terminal bonds). The overall hydrolysis of triglycine to 3 glycine plus incorporation of 2 waters has overall equilibrium constant approximately $K_{\text{hyd}}^2$ (since two independent bond-breaking equilibria compound):

This enormous number confirms that at equilibrium, essentially no triglycine can exist relative to free glycine — direct estimate:

With essentially all material as free glycine, $[\text{Gly}] \approx 0.050$ M, so

a vanishingly small equilibrium concentration — roughly one triglycine molecule per several billion glycine-equivalents. This confirms, quantitatively and at a temperature relevant to actual physiology, that no measurable protein synthesis could ever occur by uncatalyzed equilibrium condensation; the ribosome's strategy of expending ATP to make an irreversible, kinetically committed activation step (aminoacyl-tRNA formation, itself driven by pyrophosphate hydrolysis) is not a refinement but an absolute chemical necessity.

Limiting-case check: if $\Delta G^{\circ\prime}_{\text{hydrolysis}}$ were instead strongly negative for synthesis (i.e., if peptide bonds were thermodynamically stable relative to free amino acids, $K_{\text{hyd}} \ll 1$), the same algebra would predict $[\text{Gly-Gly-Gly}]$ close to the total available monomer — the formula correctly reverses sense as the sign of $\Delta G^{\circ\prime}$ flips, confirming it is not an artifact of the particular numbers chosen.

The mechanistic picture of amide bond formation as nucleophilic acyl substitution through a tetrahedral intermediate emerged from decades of physical organic chemistry on carboxylic acid derivatives, but its application to peptides is inseparable from the work of Emil Fischer, who first proved, at the turn of the twentieth century, that proteins are built from amino acids joined by what he termed the "peptide bond" (he coined the word Peptid). Fischer synthesized peptides of up to eighteen residues by classical solution chemistry — laborious, low-yielding, but decisive proof that the amide linkage he proposed was chemically correct and sufficient to explain protein composition. His painstaking stepwise couplings, protecting and deprotecting amino and carboxyl groups by hand, anticipated by half a century the logic that Bruce Merrifield would automate with solid-phase synthesis (for which Merrifield received the 1984 Nobel Prize in Chemistry), and both are direct chemical answers to the same problem this article derives: a free carboxylic acid will not react cleanly with a free amine, so it must be activated, protected, and controlled step by step.