Solid phase peptide synthesis — SPPS — is the chemistry that made modern peptide drugs possible. Before 1963, making even a short peptide in a lab was a months-long ordeal of solution chemistry, with yields collapsing at every step. Bruce Merrifield's single insight — anchor the growing peptide to an insoluble resin bead so you can wash away excess reagents instead of purifying after each step — turned peptide synthesis from an art into an industrial process. It earned him the 1984 Nobel Prize in Chemistry.
Today, SPPS underwrites a multi-billion-dollar peptide therapeutics industry that is straining under demand. Every gram of semaglutide, every dose of oxytocin, every gram of a stapled-peptide clinical candidate starts with a resin bead in a reactor at Bachem, PolyPeptide, or a contract manufacturer. This article walks through how SPPS actually works, why the chemistry is elegant but unforgiving, and why the GLP-1 boom has exposed a real global manufacturing bottleneck.
What Is Solid-Phase Peptide Synthesis?
Solid phase peptide synthesis is a method for building peptides one amino acid at a time while the growing chain remains covalently attached to an insoluble polymer resin. The core trick, introduced by R. Bruce Merrifield in 1963 (J. Am. Chem. Soc., 85:2149), is simple: if your product is anchored to a bead, you can flood the vessel with excess reagents to drive each reaction to completion, then wash everything else away with solvent. No recrystallization, no column chromatography between steps, no heartbreaking yield losses.
A critical feature of SPPS is that peptides are built C-terminus to N-terminus — backward relative to ribosomal translation. The C-terminal amino acid is loaded onto the resin first, and new residues are added to its free N-terminus. This direction reflects the chemistry of amide bond formation: the free amine of the growing chain attacks the activated carboxyl of the incoming residue.
A brief history
- 1963 — Merrifield publishes the first SPPS paper; synthesizes bradykinin and later insulin B-chain.
- 1972 — Carpino introduces the Fmoc (9-fluorenylmethyloxycarbonyl) protecting group.
- 1984 — Merrifield wins the Nobel Prize in Chemistry.
- 1990s — Fmoc chemistry becomes dominant over Merrifield's original Boc (tert-butyloxycarbonyl) chemistry because it avoids HF cleavage.
- 2000s–present — Automation, continuous-flow SPPS, green solvents, and hybrid chemical-enzymatic strategies.
Mechanism: The Coupling–Deprotection Cycle
At its core, SPPS is a repeating cycle of four steps per amino acid:
- Deprotection — remove the temporary N-terminal protecting group to expose a free amine.
- Coupling — activate the next amino acid's carboxyl group and form a new amide bond.
- Wash — flush excess reagents and byproducts.
- Repeat — for every residue in the sequence.
At the very end, a final cleavage step releases the full peptide from the resin and simultaneously removes the side-chain protecting groups.
Fmoc vs Boc
The two dominant SPPS chemistries differ in how the N-terminal protecting group is removed between coupling cycles.
| Feature | Fmoc chemistry | Boc chemistry |
|---|---|---|
| Introduced | Carpino, 1972 | Merrifield, 1963 |
| N-terminal PG | Fmoc (base-labile) | Boc (acid-labile) |
| Deprotection reagent | Piperidine (20% in DMF) | TFA (50% in DCM) |
| Final cleavage | TFA (strong acid) | HF (dangerous, specialized equipment) |
| Side-chain PGs | Acid-labile (tBu, Trt, Pbf) | Benzyl-type, stable to TFA |
| Dominant today? | Yes (>90% of industry) | Niche, still used for difficult sequences |
Fmoc chemistry uses orthogonal protection: the Fmoc group falls off under mild base, while the side-chain protecting groups only come off under strong acid (TFA) during the final cleavage. This orthogonality is why Fmoc won — Boc chemistry required handling liquid hydrofluoric acid for final cleavage, which is spectacularly hazardous.
Coupling reagents
Activating the incoming amino acid's carboxyl is its own chemistry. Historical reagents like DCC have been largely replaced by uronium salts (HBTU, HATU, COMU) and phosphonium salts (PyBOP, PyAOP), usually with DIPEA or NMM as base. HATU and COMU give higher yields on difficult couplings but cost more.
Protecting groups for side chains
Many amino acid side chains are nucleophilic or reactive and must be masked during chain assembly. A partial list:
- Lys — Boc on the ε-amine
- Arg — Pbf on the guanidine
- Cys — Trt or Acm on the thiol
- Ser/Thr/Tyr — tBu on hydroxyls
- Asp/Glu — OtBu on the side-chain carboxyl
- His — Trt
- Trp — Boc (protects the indole)
Get one of these wrong and you get branching, racemization, or chain termination.
Clinical and Experimental Evidence: What SPPS Enables
Merrifield's original paper synthesized a tetrapeptide in 1963. By 1971 he had synthesized ribonuclease A — 124 residues — as a proof that a biologically active protein could be made chemically. Today, SPPS routinely delivers peptides up to 40–50 residues at high purity and multi-kilogram scale. A non-exhaustive list of SPPS-manufactured blockbusters:
- Oxytocin (9 aa) — first peptide drug, now manufactured by SPPS
- Leuprolide (9 aa) — prostate cancer
- Octreotide (8 aa, cyclic) — acromegaly, carcinoid
- Desmopressin (9 aa) — diabetes insipidus
- Liraglutide (31 aa, lipidated) — diabetes
- Semaglutide (31 aa, lipidated) — diabetes, obesity
- Tirzepatide (39 aa, lipidated) — diabetes, obesity
- Ziconotide (25 aa, cyclic) — chronic pain
- Bivalirudin (20 aa) — anticoagulation
- Teriparatide (34 aa of PTH) — osteoporosis
Where SPPS runs out of steam
SPPS scales poorly past ~50 residues. Each coupling has some probability of failure (~99.5% per step with optimized conditions), and those errors compound: for a 50-mer, (0.995)⁵⁰ ≈ 78% of chains are full-length. For a 100-mer, it's ~61%. For a 200-mer, it's <37%, and the purification burden becomes prohibitive.
Solutions for longer targets:
- Fragment condensation — synthesize overlapping fragments by SPPS, then ligate them in solution.
- Native chemical ligation (NCL) — Kent et al., 1994 (Science); a thioester + N-terminal cysteine react to form a native amide bond, enabling ligation of unprotected peptide fragments.
- Expressed protein ligation (EPL) — couples recombinantly expressed protein thioesters with synthetic peptides.
- Recombinant expression — E. coli, yeast, or mammalian cells. This is how insulin is made; chemical SPPS for insulin is possible but not economic.
Applications and Use Cases
| Target length | Dominant method | Example |
|---|---|---|
| 2–15 aa | SPPS | Oxytocin, octreotide |
| 15–50 aa | SPPS | Semaglutide, tirzepatide, teriparatide |
| 50–100 aa | SPPS + fragment ligation or NCL | Small proteins, stapled designs |
| 100+ aa | Recombinant expression | Insulin, mAbs, erythropoietin |
SPPS is also the workhorse of peptide R&D — combinatorial libraries, stapled peptide screens, cyclic peptide drug discovery, and the synthesis of chemical probes. Every major pharma and biotech has in-house SPPS capacity for medicinal chemistry campaigns.
The GLP-1 manufacturing bottleneck
The demand for semaglutide and tirzepatide has outstripped global peptide manufacturing capacity. Both drugs are 31- and 39-residue lipidated peptides, each requiring dozens of couplings, careful handling of the lipidation step, and exhaustive purification by reverse-phase HPLC. Novo Nordisk acquired Catalent in 2024 in large part to secure fill-finish capacity. Eli Lilly has invested billions in new peptide plants in Indiana, Ireland, and Germany. Specialty CDMOs like Bachem (Switzerland), PolyPeptide (Sweden/US), and CordenPharma have multi-year backlogs. Bulk peptide API prices have spiked. This is not a temporary glitch — it is a structural mismatch between a chemistry designed for gram-scale research and a drug class that now needs multi-ton production.
Connection to Gene Editing
At first glance, SPPS and gene editing look like opposite paradigms: one is pure chemistry on a bead, the other is enzymatic editing of a living genome. But they connect in three important ways.
First, guide RNAs and CRISPR components. The synthetic guide RNAs used in therapeutic CRISPR editors are made by solid-phase oligonucleotide synthesis — essentially the same Merrifield trick with nucleotides instead of amino acids, using phosphoramidite chemistry. Every dose of Casgevy, every therapeutic sgRNA in clinical trials, exists because Merrifield's solid-phase idea generalized to other biopolymers. See our articles on CRISPR and base editing.
Second, peptide-based editors and reprogramming. When you deliver Yamanaka factors or CRISPR components as proteins — rather than as mRNA or plasmids — you need those proteins at high purity and scale. SPPS cannot build full-length Cas9 (1,368 aa), but it can make cell-penetrating peptide tags and short functional domains that are fused to recombinantly expressed editors. Hybrid chemical-biological manufacturing is an active area.
Third, both fields share a delivery ceiling. Whether you're dosing semaglutide or an in vivo CRISPR editor, you're fighting the same biology — proteases, membrane barriers, clearance, immunogenicity. Solving one tends to unlock solutions for the other.
Limitations and Open Questions
- Solvent use. Classical Fmoc SPPS consumes huge volumes of DMF, DCM, and piperidine. "Green SPPS" using safer solvents like 2-MeTHF, NBP, and γ-valerolactone is advancing but not yet industry default.
- Long peptides remain hard. Native chemical ligation is powerful but operationally complex.
- Racemization at epimerization-prone residues (Cys, His) during coupling is a persistent worry.
- Analytical characterization. Modern peptide drugs, especially lipidated ones, strain HPLC-MS methods. Regulatory expectations for impurity profiles keep rising.
- Supply chain geopolitics. Peptide CDMO capacity is concentrated in Switzerland, Sweden, Italy, China, and India. Shortages are a structural risk for patients.
Frequently Asked Questions
Why is peptide synthesis done backward (C-to-N)?
Because SPPS anchors the C-terminus to the resin first, and new residues attack the growing chain's free amine. The N-terminus of the incoming amino acid is protected (Fmoc or Boc) to prevent it from reacting with itself. After coupling, you remove the protection and repeat.
What's the difference between Fmoc and Boc SPPS?
Fmoc uses a base-labile N-terminal protecting group (removed by piperidine) and acid-labile side-chain protecting groups. Boc uses an acid-labile N-terminal group and requires HF for final cleavage. Fmoc is safer, more versatile, and now dominant; Boc is reserved for specific difficult sequences.
How long can you make a peptide by SPPS?
Reliably up to 40–50 residues. Beyond that, yields and purity drop. Longer targets use fragment condensation, native chemical ligation, or recombinant expression.
Who invented SPPS?
Bruce Merrifield at Rockefeller University, published in 1963 and awarded the Nobel Prize in Chemistry in 1984. The original paper is one of the most cited in organic chemistry.
Why are semaglutide and tirzepatide in shortage?
Because peptide manufacturing capacity — especially for lipidated 30–40-mers — is globally constrained. SPPS was designed for research and specialty drug volumes, not for drugs consumed by tens of millions of patients. Novo Nordisk, Eli Lilly, Bachem, and PolyPeptide are building capacity, but lead times for new plants are 3–5 years.
Is SPPS used for anything besides peptides?
Yes. The same solid-phase principle powers phosphoramidite oligonucleotide synthesis — the basis of synthetic DNA, RNA, and therapeutic oligos including antisense drugs and CRISPR guide RNAs. Merrifield's insight generalized to every major class of biopolymer.
