Chemical Synthesis of Peptides

A single misplaced protecting group can turn a life-saving hormone into an inert powder. This is the central drama of peptide synthesis: building a polymer with perfect directionality and stereochemistry, one residue at a time. The chemical environment has no ribosome, no aminoacyl-tRNA synthetase, and no proofreading. Every amino acid carries an amine, a carboxylic acid, and often a reactive side chain. These are three functional groups all capable of reacting with each other and with their neighbors. Solid-phase peptide synthesis (SPPS), invented by Bruce Merrifield in 1963, solves this problem directly. It anchors the growing chain to an insoluble resin bead. Orthogonal protecting groups then control exactly which bond forms next. The problem matters far beyond the laboratory bench. Insulin, calcitonin, glucagon-like peptide-1 analogs (semaglutide), and countless research peptides are made this way at industrial scale. The chemistry is modular, automatable, and — crucially — purification happens by simple filtration. This avoids recrystallization at every step. Merrifield's insight was to invert the classical solution-phase logic. Instead of isolating a soluble intermediate after each reaction, he tethered the peptide to a solid support. Excess reagents and byproducts are then washed away with a Büchner funnel. The desired molecule stays put by covalent bond to the bead. The big picture is a repeating four-part cycle. First, deprotection of the N-terminal amine; then washing; then activation and coupling of the next protected amino acid; then washing again. This cycle is executed dozens of times to build chains of 10, 30, even 100+ residues. A final global deprotection and cleavage from the resin follows. Understanding this cycle mechanistically means understanding several things. These include carbodiimide- and uronium-based amide-forming chemistry, and the acidolysis and base-mediated elimination mechanisms that remove protecting groups. It also means understanding the thermodynamic penalty of forming an amide bond. A cell pays this penalty with ATP, but a chemist pays it with an activated ester. This article works through the full Merrifield strategy. It covers Boc versus Fmoc orthogonal protection schemes, and the carbodiimide and HBTU/HATU activation mechanisms with full arrow-pushing. It also covers the Fmoc β-elimination deprotection mechanism, resin chemistry, and linker design. Racemization pathways and how to suppress them are included as well. The article addresses practical difficulties too — aggregation, difficult sequences, and the epimerization risk during activation. These are the issues that separate a clean synthesis from a discouraging mess. It closes with the logic that connects synthesis back to Peptide Sequencing Strategies. Every synthetic peptide must ultimately be verified by the same tools used to read natural ones.

Key takeaways

SPPS anchors the C-terminal residue to an insoluble resin, then elongates the chain N-to-C-terminal-protected, one residue per cycle, by repeated deprotect–couple cycles.
Two orthogonal protection schemes dominate: Boc/Bzl (acid-labile temporary, HF-cleaved permanent) and Fmoc/tBu (base-labile temporary, acid-cleaved permanent) — orthogonality means one condition removes only one class.
Coupling requires activating the free carboxylic acid (via carbodiimide, e.g., DIC, or uronium/phosphonium salts, e.g., HBTU/HATU) because the carboxylate anion alone will not attack a free amine at useful rates.
Fmoc removal proceeds by an E1cb-like base-mediated elimination through the fluorenyl ring, generating dibenzofulvene, trapped by piperidine, not simple acid hydrolysis.
Racemization via oxazolone formation is the dominant side reaction during activation; HOBt/HOAt additives and low-temperature, in-situ activation suppress it.
After the final residue, a global deprotection/cleavage cocktail (TFA + scavengers for Fmoc; HF or TFMSA for Boc) removes side-chain groups and detaches the peptide from resin simultaneously.
Key idea: Solid-phase synthesis converts a multi-step purification problem into a filtration problem, using orthogonal protecting groups to guarantee that only one new peptide bond can form per cycle, in the correct direction.

Where this lives today

Solid-phase peptide synthesis is now the standard industrial route to therapeutic peptides, including GLP-1 receptor agonists such as semaglutide (Ozempic/Wegovy) and liraglutide, which combine SPPS-built peptide backbones with post-synthetic lipidation for extended half-life. Automated Fmoc synthesizers, running the deprotect-couple cycle under computer control with real-time UV monitoring of the dibenzofulvene-piperidine adduct (which absorbs strongly near 301 nm, allowing quantitative tracking of deprotection efficiency), routinely produce research-scale peptides of 10–50 residues within a day. Longer targets — full-length proteins, or peptides beyond ~50 residues where cumulative coupling-efficiency losses become prohibitive — are addressed by native chemical ligation, in which two or more SPPS-synthesized fragments (each independently manageable in length) are joined chemoselectively in solution via a thioester-mediated trans-thioesterification and spontaneous S-to-N acyl shift, extending the reach of chemical synthesis to genuinely protein-sized targets. HATU/HBTU-type activation chemistry, developed for peptide coupling, has also found use in general amide bond formation across medicinal chemistry and materials science wherever a mild, racemization-suppressing amidation is needed. On the analytical side, every synthetic peptide must be validated by mass spectrometry and often Edman degradation or LC-MS/MS fragmentation, connecting synthesis directly to the logic of Peptide Sequencing Strategies used to confirm sequence fidelity, detect deletion peptides (from incomplete couplings), and quantify residual racemization.

Ask the tutor: Why does the choice of resin linker (e.g., Wang vs. Rink amide) determine whether the final peptide has a free acid or an amide C-terminus, and how does that cleavage mechanism relate to the side-chain deprotection chemistry happening in the same step?

The physical picture — thorough prose

Think of the resin bead as a physical anchor and a chemical trap simultaneously. A polystyrene or PEG-based bead, cross-linked so it swells in solvent but does not dissolve, is functionalized with a linker — a small bifunctional molecule that will eventually bond to the first amino acid's carboxyl group and later be cleaved to release the finished peptide. Because the peptide is now covalently fixed to a solid the size of a grain of sand, any liquid-phase excess reagent, byproduct, or unreacted small molecule can be removed by simply washing the beads on a fritted vessel and flowing fresh solvent through. This is the entire reason SPPS is fast: there is no chromatography between steps, only filtration and washing, typically 5–10 times the resin's swollen volume, several times per cycle. The synthesis proceeds C-terminus to N-terminus — the opposite direction to ribosomal synthesis, which builds N-to-C. The first amino acid is attached through its carboxyl group to the linker, with its α-amino group temporarily protected. Every subsequent amino acid arrives from solution as a fully protected building block: α-amino group masked with a temporary protecting group (removed every cycle), side chain masked with a permanent (semi-permanent) protecting group (removed only once, at the end), and carboxyl group activated in situ just before coupling. Two orthogonal schemes have dominated: the original Merrifield Boc/Bzl strategy and the now near-universal Fmoc/tBu strategy. Orthogonality is the key design principle — the temporary group must come off under conditions that leave the permanent group and the resin linkage completely untouched, cycle after cycle, so that only the intended bond reacts at each step.

Figure

The Merrifield solid-phase cycle: deprotect, wash, couple, wash, repeat

The two central chemical events, deprotection and coupling, each have their own mechanism worth deriving in full. Deprotection of Fmoc is a base-mediated E1cb elimination through an aromatic system; coupling is nucleophilic acyl substitution on an activated carbonyl, and its efficiency and stereochemical fidelity depend on exactly how that activation is achieved.

Figure

Fmoc deprotection: E1cb elimination releasing dibenzofulvene

The core ideas — governing relationships

The overall bond-forming step is an amide (peptide) bond synthesis, thermodynamically uphill in water but driven forward by activation chemistry that replaces a poor leaving group (hydroxide) with a good one.

\text{R–COOH} + \text{HO–PG} \rightleftharpoons \text{R–CO–O–PG (active ester)}, \quad \Delta G^{\circ\prime}_{\text{hydrolysis}} \approx -20 \text{ to } -30\ \text{kJ/mol}

where:

R–COOH is the resin-bound (or incoming) amino acid carboxylic acid
PG-OH is the activating leaving group (e.g., HOBt, HOAt, or the carbodiimide-derived O-acylisourea oxygen)
$\Delta G^{\circ\prime}_{\text{hydrolysis}}$ is the free energy released when the active ester subsequently reacts with the amine, replacing a poor leaving group with a much better one and driving amide bond formation forward

Reading: unlike the direct condensation of a free carboxylic acid and amine (which is thermodynamically and kinetically unfavorable in solution because water is both product and solvent, shifting equilibrium back toward hydrolysis), activation first converts –COOH into a species with a far better leaving group. The subsequent aminolysis is then effectively irreversible under the reaction conditions, because the byproduct (HOBt⁻, urea, or similar) is a much weaker nucleophile/base than hydroxide and does not compete for the newly formed amide. The second governing relationship describes racemization risk during activation, which proceeds via oxazolone (azlactone) formation — an intramolecular cyclization competing with intermolecular coupling:

k_{\text{obs}} = k_{\text{coupling}}[\text{amine}] + k_{\text{oxazolone}}

where:

$k_{\text{obs}}$ is the overall rate of consumption of the activated ester intermediate
$k_{\text{coupling}}[\text{amine}]$ is the productive, bimolecular, stereoretentive pathway (rate depends on amine concentration)
$k_{\text{oxazolone}}$ is the unimolecular, epimerization-prone cyclization pathway (rate is independent of amine concentration)

Reading: racemization is suppressed by making $k_{\text{coupling}}[\text{amine}]$ dominate — i.e., using excess amine (the incoming resin-bound N-terminus, typically 2–5 fold excess of activated acid relative to resin loading in the other convention, but kinetically the amine nucleophile concentration matters), low temperature, and additives (HOBt, HOAt) that intercept the activated species as a less racemization-prone active ester before oxazolone can form.

Derivation ACarbodiimide-mediated coupling: DIC/HOBt mechanism, full arrow-pushing

Goal: show explicitly how a free carboxylic acid is converted to an amide bond via O-acylisourea and HOBt active ester intermediates, and why HOBt prevents racemization relative to the O-acylisourea alone.

Step 1 — carbodiimide activation. The carboxylate oxygen (deprotonated or as the neutral acid with general base assistance) attacks the central, sp-hybridized carbon of diisopropylcarbodiimide (DIC), N=C=N, in a nucleophilic addition. The carboxylate oxygen lone pair pushes into the carbodiimide carbon; simultaneously one of the C=N π bonds collapses onto the adjacent nitrogen, generating a negatively charged (or protonated, neutral) nitrogen. This produces the O-acylisourea: R–C(=O)–O–C(=N–iPr)(NH–iPr).

Step 2 — the danger of the O-acylisourea. This species is a highly activated ester, but it is also prone to two side paths: (a) direct intramolecular attack by the preceding amide carbonyl oxygen on the activated carbon, cyclizing to a 5-membered oxazolone (azlactone) ring with loss of the isourea leaving group, and (b) rearrangement to the unreactive, symmetric N-acylurea (a dead-end product formed by O→N acyl migration). Both are undesirable: the oxazolone route is the direct precursor to racemization (Step 4), and N-acylurea formation destroys the activated acid entirely.

Step 3 — HOBt intercepts the O-acylisourea. 1-Hydroxybenzotriazole (HOBt), added to the reaction, has a nucleophilic oxygen (as the anion, HOBt⁻, in equilibrium) that attacks the O-acylisourea carbonyl carbon before oxazolone cyclization or N-acylurea rearrangement can dominate. The tetrahedral intermediate collapses, ejecting the diisopropylurea leaving group (a highly stabilized, poor-nucleophile byproduct), and generating the HOBt active ester: R–C(=O)–O–Bt. This ester is activated (Bt⁻ is an excellent leaving group, resonance-stabilized across the benzotriazole ring and its N-oxide-like nitrogen) but far more resistant to oxazolone cyclization than the O-acylisourea, because the bulky, stabilized OBt leaving group departs faster than the competing intramolecular cyclization can occur.

Step 4 — oxazolone racemization pathway (minor, if HOBt is efficient). If oxazolone does form — from either the O-acylisourea or, at a slower background rate, from the OBt ester itself — the mechanism is: the backbone amide carbonyl oxygen (from the N-1 residue, i.e., the newly-formed peptide bond immediately preceding the activated carbon) attacks the activated carbonyl carbon intramolecularly, displacing the OBt or isourea leaving group and closing a 5-membered ring (2-oxazolin-5-one). The α-carbon of this oxazolone, now flanked by both the ring C=N and the ring carbonyl, bears an acidic proton (its pKa is lowered by conjugation into both). Base (excess amine, tertiary amine buffer, or trace hydroxide) removes this α-proton, generating a planar, resonance-stabilized enolate/aza-allyl anion. Reprotonation occurs from either face with comparable probability, scrambling the stereocenter — this is the chemical origin of epimerization during coupling.

Step 5 — productive aminolysis. The resin-bound (or free) amine nitrogen lone pair attacks the carbonyl carbon of the HOBt active ester, forming a tetrahedral oxyanion intermediate stabilized by the developing alkoxide charge. Collapse of this intermediate ejects HOBt⁻ (a stable, low-basicity leaving group, hence a fast, clean reaction) and forms the new amide C–N bond — the peptide bond. Because HOBt⁻ is far less basic and less nucleophilic than hydroxide or alkoxide, this step is effectively irreversible under SPPS conditions, and because it proceeds through a tetrahedral intermediate directly derived from the un-racemized OBt ester (rather than through the oxazolone), the stereocenter at Cα is retained.

Result: the productive pathway is

\text{R–COOH} \xrightarrow{\text{DIC}} \text{O-acylisourea} \xrightarrow{\text{HOBt}^-} \text{R–CO–OBt} \xrightarrow{\text{H}_2\text{N–resin}} \text{R–CO–NH–resin} + \text{HOBt}^-

and HOBt's kinetic role is to outcompete both intramolecular oxazolone cyclization and N-acylurea rearrangement, funneling nearly all activated acid through the stereoretentive OBt-ester intermediate.

Go deeper · Derivation B · 2 minUronium/phosphonium reagent activation: HBTU/HATU mechanism and the "uronium vs guanidinium" pitfall

Derivation BUronium/phosphonium reagent activation: HBTU/HATU mechanism and the "uronium vs guanidinium" pitfall

Goal: derive how HBTU activates a carboxylic acid directly (without a separate carbodiimide), and explain a well-known practical trap — pre-activation with excess HBTU on the free acid before adding base can form an unreactive guanidinium species.

Step 1 — deprotonation of the carboxylic acid. A tertiary amine base (DIEA, diisopropylethylamine, or NMM) removes the acidic proton from the carboxylic acid R–COOH, generating the carboxylate R–COO⁻. This step must occur essentially simultaneously with or before exposure to HBTU, because HBTU reacts with the carboxylate oxygen, not the neutral acid.

Step 2 — nucleophilic attack of carboxylate on the electrophilic uronium carbon. HBTU (2-(1H-benzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate) contains an electrophilic carbon at the center of the uronium group, flanked by two dimethylamino groups and an OBt-derived oxygen (in the O-form) or nitrogen (in the reactive N-form — HBTU is actually drawn/sold as the more stable N-isomer, but reacts via the O-form in solution). The carboxylate oxygen lone pair attacks this uronium carbon, and the OBt group departs as the resonance-stabilized benzotriazolyl anion (HOBt⁻ again, now released as free HOBt/HOBt⁻ in solution). This single step directly installs the OBt active ester, R–CO–OBt, plus tetramethylurea as byproduct — bypassing the O-acylisourea/N-acylurea branch point entirely, since there is no carbodiimide intermediate.

Step 3 — why this is fast and low-racemization. Because the mechanism proceeds directly to the OBt (or OAt, for HATU) active ester without ever forming a symmetric anhydride or O-acylisourea, the competing oxazolone pathway has less opportunity to intervene; HATU (with a 7-azabenzotriazole leaving group instead of benzotriazole) is even more effective, because the ring nitrogen adjacent to the leaving oxygen provides an additional neighboring-group stabilization of the transition state for aminolysis, further outcompeting oxazolone formation. Empirically, HATU gives lower racemization than HBTU, which gives lower racemization than DIC/HOBt alone, for difficult (sterically hindered or His/Cys-adjacent) couplings.

Step 4 — the guanidinium trap. If HBTU is mixed with the carboxylic acid before sufficient base is present (or in excess relative to base), the free carboxylate that does exist can indeed activate correctly, but any unreacted amine present in the pot — including, critically, the resin-bound N-terminal amine if it has not yet been fully separated from the activation step — can attack the uronium carbon directly instead of a preformed active ester attacking it. This generates a stable, unreactive tetramethylguanidinium adduct on the peptide's N-terminus, permanently capping the chain and terminating its further elongation — a classic, hard-to-diagnose synthesis failure. The fix is procedural, not just chemical: always pre-activate the carboxylic acid with HBTU/base for 1–2 minutes in a separate vessel before adding it to the deprotected resin-bound amine, ensuring the OBt active ester is fully formed before the nucleophile is introduced.

Result: the productive pathway is

\text{R–COO}^- + \text{HBTU} \rightarrow \text{R–CO–OBt} + \text{TMU} \xrightarrow{\text{H}_2\text{N–resin}} \text{R–CO–NH–resin} + \text{OBt}^-

and the guanidinium side reaction is

\text{H}_2\text{N–resin} + \text{HBTU} \rightarrow \text{resin–NH–C(=NMe}_2^+\text{)(NMe}_2\text{)} \quad (\text{dead end, chain-terminating})

Worked examples

Example 1 — Foundational: computing coupling efficiency and overall yield across a synthesis. A tripeptide is synthesized by SPPS with three coupling steps, each proceeding at 98.5% efficiency (a realistic value for a well-optimized Fmoc coupling using HBTU/DIEA on a straightforward sequence). What is the overall crude yield of the full-length, correctly coupled tripeptide, and how does this compare to a 30-mer synthesized under the same per-step efficiency? For $n$ sequential coupling steps each with efficiency $p$ , the fraction of chains that are full-length and correct after all couplings is

Y = p^n

For the tripeptide ( $n = 2$ couplings after the first residue is loaded — three residues means two peptide bonds formed by coupling, since the first residue is simply attached to resin):

Y_{\text{tri}} = (0.985)^2 = 0.9702, \text{ i.e., } 97.0\%

For a 30-residue peptide, $n = 29$ couplings:

Y_{30} = (0.985)^{29} = e^{29 \ln(0.985)} = e^{29 \times (-0.015113)} = e^{-0.4383} = 0.645, \text{ i.e., } 64.5\%

This demonstrates why per-step efficiency dominates the practicality of long syntheses: a seemingly excellent 98.5% coupling yield still leaves over a third of the crude product as truncated or deletion sequences by residue 30. Industrial and research syntheses of long peptides therefore push per-step yields above 99.5–99.8% (double coupling, extended reaction times, HATU for difficult junctions) specifically to keep $Y$ tolerable at $n \sim 50$ –100. As a limiting-case check: if $p \to 1$ (perfect coupling), $Y \to 1$ regardless of $n$ , correctly recovering ideal SPPS; if $p = 0.5$ (a badly failing coupling), even $n=2$ gives $Y = 0.25$ , showing the exponential penalty clearly.

Example 2 — Applied: choosing between Boc and Fmoc strategy for a Cys-containing therapeutic peptide with a C-terminal amide. A chemist must synthesize a 12-residue peptide containing an internal cysteine and requiring a C-terminal carboxamide (common in many neuropeptides and hormone analogs). Which protection strategy is preferable, and why, considering deprotection chemistry? Boc/Bzl strategy uses TFA (mild acid, ~25–50% in DCM) for the repetitive N-α deprotection every cycle, but requires strong anhydrous HF (or TFMSA) for final side-chain deprotection and resin cleavage. Cysteine thiols are notoriously prone to side reactions (disulfide scrambling, alkylation by cations generated during acidolysis) during the harsh HF step, and HF handling requires specialized apparatus (Kel-F or Teflon lines, HF-resistant reaction vessels) — a serious practical and safety burden. Fmoc/tBu strategy uses mild base (20% piperidine/DMF) for repetitive deprotection — leaving acid-labile side-chain groups (tBu ethers/esters, Trt on Cys/His/Gln/Asn, Boc on Lys) completely untouched every cycle — and only a single, milder TFA-based cocktail (typically TFA/water/triisopropylsilane/EDT, 92.5:2.5:2.5:2.5, or similar) for final global deprotection and cleavage. The Cys-Trt bond, in particular, is cleaved cleanly by TFA with a good carbocation scavenger (EDT or thiol-based scavengers) at the very end, in one step, minimizing the number of times the free thiol is exposed to reactive electrophiles. Given the internal Cys and amide C-terminus (achieved simply by using a Rink amide resin, whose linker releases a C-terminal primary amide upon TFA cleavage rather than a free acid), Fmoc/tBu is strongly preferred: it avoids HF entirely, minimizes thiol side reactions to a single final deprotection event, and directly delivers the amide C-terminus via linker choice rather than requiring a separate amidation step. This is why Fmoc/tBu has essentially superseded Boc/Bzl for routine synthesis, Boc chemistry being retained mainly for specialized cases (e.g., certain native chemical ligation building blocks, or when Fmoc's base sensitivity is itself a liability, such as very base-labile depsipeptide linkages).

Example 3 — Exam/competition level: quantifying racemization suppression and predicting optical purity. During Fmoc-SPPS coupling of Fmoc-Cys(Trt)-OH (a residue notoriously prone to racemization due to the electron-withdrawing, ring-stabilizing thiol substituent lowering the α-proton pKa in the oxazolone intermediate) to a resin-bound peptide, two activation protocols are compared: (a) DIC/HOBt at 25 °C, giving 3.8% D-Cys epimer in the product by chiral HPLC. (b) DIC/HOBt at 0 °C with pre-activation time reduced from 5 min to 30 s, giving 0.4% D-Cys epimer. Given that racemization proceeds through the oxazolone pathway with an approximate Arrhenius-type temperature dependence, and assuming the oxazolone-forming step has an apparent activation energy $E_a \approx 60$ kJ/mol relative to the competing coupling step, estimate whether temperature alone (25 °C → 0 °C) can plausibly account for the observed ~9.5-fold reduction in racemization (3.8%/0.4%), or whether the reduced pre-activation time must also be contributing. Using the ratio of rate constants at the two temperatures (Arrhenius form, with $R = 8.314\ \text{J mol}^{-1}\text{K}^{-1}$ , $T_1 = 298\ \text{K}$ , $T_2 = 273\ \text{K}$ ):

\frac{k(T_1)}{k(T_2)} = \exp\left[\frac{E_a}{R}\left(\frac{1}{T_2} - \frac{1}{T_1}\right)\right]

\frac{1}{T_2} - \frac{1}{T_1} = \frac{1}{273} - \frac{1}{298} = 0.0036630 - 0.0033557 = 3.073 \times 10^{-4}\ \text{K}^{-1}

\frac{E_a}{R}\left(\frac{1}{T_2}-\frac{1}{T_1}\right) = \frac{60000}{8.314} \times 3.073 \times 10^{-4} = 7217.2 \times 3.073\times10^{-4} = 2.218

\frac{k(T_1)}{k(T_2)} = e^{2.218} = 9.19

Temperature alone predicts roughly a 9.2-fold reduction in the oxazolone-formation rate constant on cooling from 25 °C to 0 °C, remarkably close to the observed 9.5-fold drop in D-epimer content. This shows that temperature is in fact sufficient, quantitatively, to explain almost all of the improvement — the shortened pre-activation time is a secondary, reinforcing factor (less time for the O-acylisourea/OBt-ester equilibrium population to reach steady-state oxazolone leakage) rather than the dominant cause. As a limiting-case sanity check: if $E_a \to 0$ (no temperature dependence to the racemization step itself), the ratio $\to 1$ regardless of $\Delta T$ , correctly predicting that cooling would do nothing — consistent with the general chemical principle that only kinetically (not thermodynamically) controlled side-reactions can be suppressed by temperature in this way.

Limiting cases, scope & common traps

Situation	Behavior	Why
Fmoc removal with 20% piperidine, RT, 2×5 min	Complete, clean deprotection	E1cb elimination is fast; dibenzofulvene is efficiently trapped as the piperidine adduct, preventing re-alkylation of the free amine
Boc removal with TFA/DCM	Complete N-α deprotection, side chains (Bzl-based) untouched	Boc carbamate cleaves via tert-butyl cation elimination (SN1-like) under acid; benzyl-type permanent groups are far more acid-stable
Coupling near a Pro or N-methyl residue	Slow, may need double coupling or HATU	Steric hindrance around the secondary amine nucleophile slows nucleophilic attack on the active ester
Long hydrophobic stretches (e.g., transmembrane-mimetic sequences)	"Difficult sequence" aggregation, incomplete deprotection/coupling	Growing chain undergoes intramolecular β-sheet-like H-bonding on-resin, physically occluding the N-terminus from reagents
Excess HBTU without adequate base present	N-terminal guanidinylation, chain-terminating	Free amine attacks uronium carbon directly instead of a preformed carboxylate
Cys, His, Trp, Met residues present	Elevated need for scavengers during final TFA cleavage	Carbocations generated from tBu/Trt/Boc groups during acidolysis alkylate electron-rich side chains unless scavenged (TIS, EDT, water, thioanisole)
$n \to 1$ (dipeptide)	$Y = p$ trivially; racemization/aggregation concerns minimal	No cumulative multiplication of per-step yield, and no long-range on-resin folding

The chemist behind it

Bruce Merrifield (1921–2006) developed solid-phase peptide synthesis at Rockefeller University beginning in 1959, publishing the foundational method in 1963 in the Journal of the American Chemical Society. Working initially with a simple insoluble, cross-linked polystyrene resin functionalized with a chloromethyl handle, he showed that a peptide chain could be built entirely while covalently tethered to an insoluble bead, with each cycle's excess reagents removed by filtration rather than recrystallization. He automated the process mechanically in the mid-1960s, building some of the first peptide synthesizer instruments, and used the method to synthesize ribonuclease A (124 residues) in the late 1960s, a landmark demonstration that solid-phase methods could reach full protein-scale targets. He received the Nobel Prize in Chemistry in 1984 "for his development of methodology for chemical synthesis on a solid matrix." Merrifield's original Boc/Bzl strategy has since been substantially supplanted for routine work by the milder Fmoc/tBu strategy developed by Louis Carpino and Colin Sheppard in the 1970s, but the core solid-phase logic — anchor, deprotect, activate, couple, wash, repeat — remains unchanged and is the basis of essentially all modern automated peptide synthesizers.

Origins · 1 minRobert Bruce Merrifield and how this was discovered

Robert Bruce Merrifield

1921–2006 · American biochemist, Rockefeller University; Nobel Prize in Chemistry, 1984

Merrifield trained as an organic chemist at UCLA before joining Rockefeller Institute (later Rockefeller University), where he worked initially on peptide growth factors and became frustrated by the slow, low-yielding purification demands of classical solution-phase peptide synthesis — each intermediate needed to be isolated, characterized, and purified before the next coupling. His 1963 paper describing solid-phase synthesis on an insoluble polystyrene support is now considered one of the most consequential methodological papers in the history of organic and biological chemistry, having enabled not only peptide but also, via analogous logic, oligonucleotide and oligosaccharide solid-phase synthesis and ultimately automated DNA synthesizers.

History

- 1901 — Emil Fischer and Ernest Fourneau report the first deliberate solution-phase peptide bond synthesis (a dipeptide, glycylglycine).

- 1932 — Max Bergmann and Leonidas Zervas introduce the carbobenzyloxy (Cbz/Z) protecting group, the first practical amine-protection strategy for peptide synthesis.

- 1963 — Bruce Merrifield publishes solid-phase peptide synthesis on a chloromethylated polystyrene resin (Boc/Bzl strategy).

- 1970 — Louis Carpino introduces the Fmoc protecting group as a base-labile alternative to Boc.

- 1984 — Merrifield awarded the Nobel Prize in Chemistry for solid-phase synthesis methodology.

Why does SPPS synthesize peptides C-to-N terminus while ribosomes synthesize N-to-C?

The direction is a consequence of which functional group is anchored first. In SPPS, the first (C-terminal) amino acid's carboxyl group is covalently attached to the resin linker, leaving its α-amine free to be deprotected and to react with the next incoming, C-activated amino acid — so the chain grows by adding new C-terminal-activated residues to a free N-terminus, meaning synthesis proceeds from C to N overall. Ribosomes work oppositely: the growing peptide's C-terminus is held on the P-site tRNA, and the incoming aminoacyl-tRNA's amine attacks that ester carbonyl, so the chain elongates by adding new residues to the C-terminus, growing N to C. Both strategies use the same underlying nucleophilic acyl substitution chemistry; they simply anchor and activate opposite ends.

What exactly is "orthogonality" and why can't you just use one protecting group for everything?

Orthogonality means the temporary (α-amino) protecting group and the permanent (side-chain) protecting groups are removed by mutually exclusive chemical conditions, so that removing one class every cycle never touches the other. If a single protecting group class were used everywhere, every deprotection cycle (needed dozens of times to expose the growing N-terminus) would also strip the side-chain protection, exposing reactive side chains (amines, thiols, imidazoles) to activation reagents throughout the synthesis, causing branching, capping, and side-chain-mediated side reactions. Fmoc (base-labile) paired with tBu/Trt/Boc (acid-labile) side-chain groups, or Boc (acid-labile) paired with Bzl-type (strong-acid/HF-labile) side-chain groups, achieves the necessary separation of conditions.

Why is HATU generally better than HBTU for difficult couplings?

HATU's leaving group is 7-aza-1-hydroxybenzotriazole (HOAt) rather than HOBt. The additional ring nitrogen, positioned adjacent to the leaving oxygen, provides intramolecular general-base or neighboring-group assistance during the aminolysis transition state, accelerating productive coupling relative to the competing oxazolone-formation pathway. This kinetic advantage becomes especially important for sterically hindered couplings (e.g., onto N-methylated residues, β-branched residues like Ile/Val/Thr, or Pro), where the productive aminolysis step is intrinsically slow and would otherwise leave more time for racemization or incomplete coupling to occur.

What actually happens during the final TFA cleavage cocktail step?

Two things happen essentially simultaneously: all remaining acid-labile side-chain protecting groups (tBu esters/ethers on Asp/Glu/Ser/Thr/Tyr, Trt on Cys/His/Asn/Gln, Boc on Lys/Trp) are cleaved by protonation and loss of a stabilized carbocation, and the peptide-resin linker bond is itself cleaved (also typically acid-labile, e.g., the Rink amide or Wang linker), releasing the free peptide into solution. Scavengers (triisopropylsilane, water, EDT, thioanisole, phenol depending on the sequence) are included specifically to intercept the transient carbocations generated during this step, preventing them from alkylating electron-rich side chains (Trp indole, Cys thiol, Tyr phenol, Met thioether) before they can diffuse away and be quenched.

Why does peptide synthesis become dramatically harder for longer sequences, beyond just multiplying per-step yields?

Beyond the straightforward multiplicative yield loss ( $Y = p^n$ ), long or hydrophobic sequences develop "difficult sequence" behavior: the growing, still-resin-bound chain can adopt transient β-sheet-like or aggregated secondary structure directly on the solid support, especially once ~6–10 hydrophobic residues have accumulated. This on-resin aggregation physically shields the reactive N-terminus from both piperidine (deprotection) and incoming activated esters (coupling), causing progressively worse deletion sequences at exactly the residues where aggregation is most severe — a problem addressed practically by elevated temperature synthesis, chaotropic/pseudo-proline building blocks that disrupt local β-sheet propensity, or switching to fragment condensation strategies for very long targets.

How is racemization actually measured and reported experimentally?

Racemization is quantified by incorporating the amino acid of interest, cleaving/deprotecting the peptide, hydrolyzing it to free amino acids, and separating the D- and L-enantiomers by chiral HPLC or GC (often after derivatization, e.g., with Marfey's reagent, which forms diastereomeric derivatives separable on an ordinary reversed-phase column). The reported number is typically the percentage of the D-epimer relative to total (D+L) for that residue, and acceptable levels for pharmaceutical peptides are usually below 0.5–1% per racemization-prone residue, driving the choice of low-racemization reagents (HATU, low temperature, minimal pre-activation time) discussed in Worked Example 3.

Does SPPS eventually connect back to how peptides are sequenced or characterized?

Yes — every synthetic batch must be validated against its intended sequence, typically by mass spectrometry (confirming the correct molecular mass, ruling out deletion or truncation sequences) and often by the same fragmentation-based logic used in natural Peptide Sequencing Strategies, such as tandem MS/MS or, historically, Edman degradation. This creates a tight feedback loop: synthetic chemistry produces the molecule, and sequencing/analytical chemistry confirms fidelity, quantifies impurities (racemized, truncated, or deletion peptides), and guides optimization of coupling conditions for future syntheses of the same or related sequences.