Peptide synthesis: how peptides are made, from the first resin bead to a purified kilogram

Peptide synthesis is the controlled assembly of amino acids into a defined sequence. In practice that means repeating one short chemical cycle dozens of times with near-perfect efficiency, then separating the intended molecule from a crowd of closely related failures. The chemistry has been settled since the 1960s. What separates a research-scale synthesis from a kilogram of active substance is process control, purification and documentation.

This guide walks through every route used in practice: the Fmoc solid-phase cycle step by step, liquid-phase synthesis, fragment condensation, recombinant expression and enzymatic ligation. It then covers scale-up, purification, what actually drives the price of a peptide, and the material footprint of the whole operation. For the mechanics of the amide bond itself, see what is a peptide bond. For a route-by-route comparison aimed at buyers, see SPPS vs LPPS vs hybrid.

The four routes to a peptide

Route	How the chain is built	Typical fit	Main limitation
Solid-phase (SPPS)	On an insoluble resin, C-terminus first, reagents washed away each cycle [6]	The predominant platform for peptide manufacture [1]	Large solvent and reagent excesses; efficiency falls with length [1]
Liquid-phase (LPPS)	In solution, intermediates isolated	Shorter sequences, including at large scale [1]	More process development, so higher initial cost and longer timelines [1]
Fragment condensation (hybrid)	SPPS fragments joined in solution	Longer peptides and large volumes [1]	More steps; fragment coupling adds its own impurities [9]
Recombinant or semi-synthetic	Expressed in a host organism, optionally modified chemically	Long sequences of natural L-amino acids [1]	Established platforms assemble only natural L-amino acids [1]

A 2024 analysis by fourteen member companies of the ACS Green Chemistry Institute Pharmaceutical Roundtable, covering 40 synthetic peptide processes, states that peptide manufacturing “is often limited to solid-phase peptide synthesis (SPPS), liquid phase peptide synthesis (LPPS), and to a lesser extent hybrid SPPS/LPPS, with SPPS emerging as a predominant platform technology” [1]. More than 80 peptide drugs have reached the market since insulin [17], and a 2022 review counts 33 non-insulin peptide drug approvals worldwide since 2000 [16].

Solid-phase peptide synthesis, step by step

Bruce Merrifield published the solid-phase synthesis of a tetrapeptide in 1963 [2] and received the 1984 Nobel Prize in Chemistry “for his development of methodology for chemical synthesis on a solid matrix” [3]. The insight was to anchor the growing chain to an insoluble bead. Because the peptide cannot escape, a large excess of reagent can be used to drive each reaction to completion and then simply rinsed away, with no isolation step between residues [6].

Step 1: choose the resin and the linker

The solid support is a polymer bead that swells in organic solvent, sometimes to about six times its dry volume [6]. Three families are in common use: polystyrene, polyethylene glycol-grafted polystyrene, and fully PEG-based resins [6]. Between the bead and the peptide sits a linker, and that choice is not cosmetic. The linker “determines the loading of the resin, the distance between resin and peptide, chemical conditions for coupling and release and most importantly, the C-terminal functionality of the synthetic peptide” [6]. A standard acid-labile linker releases a C-terminal acid or amide, the two naturally occurring terminals, while a 2-chlorotrityl resin releases a fully protected fragment suitable for later condensation [6].

Loading is expressed as the millimoles of amino acid attached per gram of resin [6]. It sets how much peptide a given reactor volume can hold, which makes it one of the levers on process economics. EMA treats a resin preloaded with the first residue as a starting material, while the unloaded support is not, because it is not incorporated into the active substance [9].

Step 2: remove the Fmoc group

Each cycle begins by unmasking the alpha-amino group of the residue at the top of the chain. In Fmoc chemistry that is done with piperidine, classically at 20% in dimethylformamide [1]. The released dibenzofulvene is reactive, but on a solid support it and its adducts can simply be washed away [4]. The fluorene group also absorbs strongly in the ultraviolet, which gives a free in-line indicator of whether the deprotection worked [4].

Fmoc is now the dominant strategy. A 2016 review states that “today, Fmoc SPPS is the method of choice”, in large part because it avoids the anhydrous hydrogen fluoride that Boc chemistry needs for final cleavage and because it gives a genuinely orthogonal combination of temporary and permanent protecting groups [4]. Boc chemistry always carried the problem that the conditions for removing the temporary group and the semi-permanent side-chain protection were not fully differentiated [4].

Step 3: couple the next amino acid

The incoming Fmoc amino acid arrives with its side chain protected and its carboxyl group activated by a coupling reagent. Carbodiimides such as DCC and DIC were the main activators for decades; uronium salts such as HATU and TBTU and phosphonium salts such as PyBOP are now routine, and Oxyma was introduced as a safer additive in place of benzotriazole-based HOBt [6]. Side chains carry their own protection throughout: Pbf on arginine, trityl on asparagine, glutamine, cysteine and histidine, OtBu on aspartate and glutamate, tert-butyl on serine, threonine and tyrosine [4].

Two practical notes from the literature. First, acetic acid contamination in commercial Fmoc amino acid derivatives is a serious problem, because it “cannot be detected by RP-HPLC and causes permanent capping”; suppliers offer gas chromatography methods with a detection limit of 0.2% or a semi-quantitative ninhydrin assay instead [4]. Second, the cost of the building blocks themselves is no longer the obstacle it once was: very high-quality Fmoc derivatives “are available at low cost because of the economies of scale arising from current multiton production of therapeutic peptides by Fmoc SPPS” [4].

Step 4: wash, test, and cap if needed

Washing is where most of the solvent goes. The ACS analysis is blunt about it: “Organic solvents and water constitute the greatest proportion of the waste generated as a large excess of solvent is routinely used to rinse the resin after each coupling and deprotection operation” [1]. In the synthesis stage of the processes surveyed, dimethylformamide alone accounted for 89.2% of solvent use [1].

Completeness is checked on a resin sample. The classic method is the colour test for free terminal amino groups published by Kaiser and colleagues in 1970 [7]. If a coupling has not gone to completion, the chemist either repeats it or caps the remaining free amines so they cannot react in later cycles. Published Fmoc/tBu protocols set out each of these operations in order, from resin handling and anchoring through coupling, capping and Fmoc deprotection to final cleavage and disulfide bridge formation [5]. EMA describes SPPS exactly in these terms: the sequence is built up “by repeated cycles of deprotection, washing, coupling and capping steps” [9]. Capping changes which impurity you get: EMA describes systematic acetylation of unreacted coupling sites as a way of preventing deletion sequences, which differ from the product by one or more residues, and notes that the acetylation procedure yields acetylated peptide fragments, that is truncated sequences, instead [9].

Step 5: cleave and deprotect

When the sequence is complete, the peptide is released from the resin and the side-chain protecting groups are removed. In Fmoc chemistry this is a single acidolytic step with a trifluoroacetic acid cocktail. The cocktail contains scavengers, because cleavage liberates highly reactive carbocations that will otherwise attack sensitive residues such as tryptophan, tyrosine and methionine [6] [9]. Removal of the Pbf group from arginine is usually complete within one to two hours, but peptides with several arginines still need extended cleavage times [4].

The crude peptide is then precipitated, usually into cold ether, and taken forward to purification. In the ACS data set, isolation used solvent for little more than a single precipitation of the crude material plus the final lyophilization [1].

Step 6: difficult sequences

Some sequences fight back. The main cause of failure is aggregation of the growing chain on the resin, driven by intermolecular hydrogen bonding between protected peptide chains [8] [4]. The standard tools are pseudoproline dipeptides, which kink the chain, break its structure and make the peptide more soluble, and which need no special coupling conditions; backbone protection such as Dmb-glycine, where a substituent on the amide bond stops it taking part in interchain hydrogen bonding; and depsipeptide units [4] [8]. The Nature Protocols account of Fmoc SPPS reports making the 41-residue peptide corticotropin-releasing factor in about 80 working hours by standard protocol, and reaching an extremely difficult WW-domain analogue only with depsipeptide and pseudoproline units [8].

Aspartic acid deserves a separate mention. Aspartimide formation is described as the most serious side reaction in Fmoc chemistry; it can generate nine different by-products, some co-eluting with the target, and Asp-Gly is the worst case [4]. These are the impurity classes a specification has to control, which is covered in peptide API impurities and specifications.

Automation, microwaves and faster cycles

Automated synthesizers differ in how they transfer and mix solutions, their scale, and extras such as in-line monitoring by chromophore or conductivity, inert atmosphere and automated cleavage [6]. Microwave-assisted instruments were developed to push difficult couplings: microwave energy heats any molecule with a dipole moment directly, which raises reaction rates and has been shown to allow synthesis of difficult and rigid sequences [6]. The trade-off is side reactions. Racemisation is “particularly high with microwave heating and preactivation”, and in one model study cysteine protected with the standard trityl group racemised to 8.0% under conventional conditions but 26.6% at 80 degrees Celsius [4].

Cycle times have fallen a long way. A 2014 report on high-efficiency SPPS describes a four-minute standard cycle and a 90% reduction in total waste, with the (65-74) acyl carrier protein fragment and the 42-residue amyloid-beta peptide obtained at 93% and 72% purity by UPLC-MS in 44 and 229 minutes respectively [10].

Liquid-phase synthesis and fragment condensation

Liquid-phase synthesis builds the chain in solution and isolates each intermediate, which allows impurities to be removed early instead of carried to the end. It is more work to develop: “typically, a greater effort is required to develop an LPPS-based manufacturing process compared to SPPS, resulting in higher initial costs and longer process development” [1]. The pay-off shows in the material numbers. The four LPPS processes in the ACS data set, sequences of 10 to 20 residues, averaged a process mass intensity of 3,984 against about 13,000 for SPPS, with 263 kg of input per kilogram per amino acid [1].

Tag-assisted variants attach the chain to a soluble tag that keeps it in the organic layer, so by-products and excess reagents can be removed by extraction or precipitation rather than filtration; the technology is in use at multikilogram scale for short sequences [1].

Fragment condensation combines the two. Fragments are made by SPPS, often on 2-chlorotrityl resin so they come off protected [6], then coupled in solution. The ACS analysis notes that the fragment approach “may be more efficient than linear SPPS”, because higher initial resin substitution may be achievable, difficult couplings can be circumvented, the resulting impurity profile may make chromatography easier, and smaller excesses of amino acid derivatives may suffice [1]. Enfuvirtide, a 36-residue peptide, was the early production-scale example; its bulk was made by solution-phase condensation of three side-chain-protected fragments [12] [6]. Tirzepatide is a recent one [1].

Regulators treat fragments carefully. EMA accepts protected natural L-amino acid derivatives as starting materials in general, allows short peptide fragments in justified cases, and treats modified fragments, including those carrying PEG chains or cyclisation, as intermediates rather than starting materials [9].

Recombinant, semi-synthetic and enzymatic routes

Expression in a host organism avoids stepwise chemistry altogether. Its reach is limited: “currently established recombinant platforms can only assemble peptides containing the natural L-AAs”, and in general development is faster and cheaper by chemical synthesis, so recombinant routes are concentrated on long sequences of natural amino acids [1]. Where they fit, the economics can be compelling, since large-scale recombinant synthesis “has the advantage of potentially driving cost down many folds” [1].

There is a regulatory consequence that is easy to miss. In the USA, “peptides produced recombinantly are filed as Biologics License Applications (BLA), whereas synthetic peptides are filed as New Drug Applications (NDA) if their primary sequence is 40 amino acids or fewer” [1].

Semi-synthesis splits the difference. The prescribing information for the reference semaglutide product states that “the peptide backbone is produced by yeast fermentation” and that the protraction mechanism is albumin binding, “facilitated by modification of position 26 lysine with a hydrophilic spacer and a C18 fatty di-acid” [19]. The design account explains why the route works: substituting the only other lysine, at position 34, with arginine left a single site for acylation and so enabled “a semi-recombinant process, where the peptide backbone was produced recombinantly and the fatty acid attached afterwards by a simple chemical reaction” [18]. The label records a matching minor modification at position 34 “to ensure the attachment of only one fatty di-acid” [19].

Chemical and enzymatic ligation

Joining two unprotected fragments selectively is its own field. Native chemical ligation, reported in 1994, reacts two unprotected peptide segments chemoselectively to give a thioester-linked intermediate that rearranges spontaneously to a native peptide bond at the junction [14]. Later chemoselective methods include alpha-ketoacid-hydroxylamine ligation and serine/threonine ligation [13].

Enzymes can do the same job in water. Engineered subtilisin variants, from subtiligase through peptiligase to omniligase-1, couple an activated acyl donor peptide to an unprotected acyl acceptor. Subtiligase gave average ligation yields of 60 to 70% and needed a tenfold excess of acceptor, “which is not good enough for industrial applicability”; peptiligase more than doubled the synthesis-to-hydrolysis ratio, and omniligase-1 variants coupled more than 250 peptides out of a 400-member acceptor library [15]. Enzymatic assembly of deprotected fragments is described in the ACS analysis as an orthogonal approach to chemical synthesis [1].

Recombinant expression combined with genetic code expansion extends the reach further, allowing unnatural amino acids to be introduced biologically, though chemical synthesis remains the preferred industrial method because it accepts arbitrary building blocks [16].

Scale-up: from a vial to a kilogram

Scale-up in peptide synthesis is mostly a question of reactors, chromatography columns and yield. Production-scale SPPS reactors run “from hundreds of liters such as the automatic 250 mL x 6 SPPS reactor up to thousands of liters, as illustrated in the 6000 L SPPS reactor system”, and production-scale manufacture of peptides of around 70 amino acids by SPPS has “proven feasible in principle” [1]. Chemical routes to relatively complex peptides have been developed on scales up to metric tons per year [11].

The selection criteria are explicit in the ACS analysis: “material demand, required batch size, available manufacturing capability and capacity, complexity, scalability, potential for automation, availability and costs of starting materials, risk (process, timeline, costs), and regulatory control strategy”, and the choice should fit the stage of the product’s life cycle [1]. Early material is usually made by SPPS because lab synthesizers are available and less process development is needed before scale-up [1].

Stage	What changes at scale	Evidence
Synthesis	Reactor volume, resin loading, solvent handling, in-line monitoring	[1] [6]
Cleavage	Larger acid volumes, scavenger ratios, precipitation and filtration behaviour	[1] [6]
Purification	Two to three orthogonal chromatographic steps, fraction pooling rules, repurification of side fractions	[9] [20]
Counter-ion	A final ion-exchange step to set the salt form before isolation	[1]
Isolation	Tangential flow filtration or evaporation to remove cosolvent, then lyophilization; tray lyophilizers handle batches below 50 kg	[1]
Documentation	Named sites, defined starting materials, in-process controls, reference standards	[9]

Purification is half the process

Purification is where most of the material and much of the cost goes. The ACS analysis states that “downstream processes account directly for approximately 50% of the Process Mass Intensity (PMI), due to the use of dilute product solutions and long chromatographic gradients”, and that purification yield has a second, compounding effect, because losses late in the process have to be paid for with extra upstream material [1].

Most peptide active substances are purified by preparative reversed-phase chromatography, often with an orthogonal ion-exchange step, because related impurities are too similar to the product for anything less selective [1] [20]. Water and acetonitrile dominate: in the purification stage of the surveyed processes, water was 77.3% and acetonitrile 18.4% of solvent use [1]. Continuous multicolumn approaches attack the yield-purity trade-off directly; the original multicolumn countercurrent solvent gradient purification paper reported a yield close to 100% for calcitonin at the required purity, against a maximum of 66% from single-column batch chromatography [21].

Because reversed-phase purification uses trifluoroacetic acid, cationic peptides come out of it as trifluoroacetate salts [23], which is why a final ion-exchange step is often needed “to ensure the peptide is associated with the right counterion before going to the final isolation step” [1]. Our companion guides cover preparative purification and counter-ions, TFA versus acetate in detail.

Isolation is almost always freeze-drying. EMA states plainly that “lyophilisation of synthetic peptides is considered common practice”, with precipitation, crystallisation and spray drying as alternatives [9], and the ACS analysis notes that lyophilization is energy-intensive but gives high yield compared with those alternatives [1]. The cycle itself is covered in our peptide lyophilization guide.

What drives the cost of a peptide

Searchers ask how much it costs to synthesize a peptide. There is no single number, because cost tracks the amount of work and the amount of material consumed per unit of product rather than the mass delivered. The published evidence points at six drivers.

Driver	Why it moves the cost	Evidence
Chain length	Every residue adds a deprotection, a coupling and several washes. PMI per amino acid averaged 874.5 for commercial and Phase 3 SPPS processes and 1,464 for Phase 1 and preclinical ones	[1]
Purification yield	Downstream processing is about half of total material use, and late losses must be offset upstream	[1]
Crude purity	A cleaner crude means a shorter gradient, fewer repurifications of side fractions and a wider collection window	[9] [20]
Difficult sequences	Aggregation-prone stretches need double couplings, pseudoproline or backbone-protected building blocks, and give lower yields	[4] [8]
Modifications	Fatty acid or PEG attachment, cyclisation and unnatural residues add steps, and large PEG units are generally introduced manually in solution	[6]
Quality grade	A GMP active substance brings defined starting materials, in-process controls, validated methods, reference standards and a dossier; research-grade material does not	[9]

Two further observations help set expectations. Standard Fmoc building blocks are comparatively cheap because of multiton demand [4], so for a routine sequence the reagents are rarely the dominant line. And process maturity matters: PMI per amino acid falls as a process moves from preclinical to commercial, which is the material signature of development work paying off [1].

Green chemistry and process mass intensity

Process mass intensity is the total mass of input per unit mass of product, covering synthesis, purification and isolation [1]. On that measure peptides are expensive to make:

Modality	Process mass intensity (kg input per kg product)
SPPS overall, as the analysis states	About 13,000 [1]
SPPS, commercial and Phase 3 (14 processes, 6 to 43 residues)	1,683 to 34,585, average 13,063 [1]
SPPS, Phase 1 and preclinical (14 processes)	Average 33,141 [1]
LPPS (4 processes, 10 to 20 residues)	Average 3,984 [1]
Oligonucleotides, for comparison	3,035 to 7,023, average 4,299 [1]
Biopharmaceuticals, for comparison	About 8,300 [1]
Small molecules, for comparison	Median 168 to 308 [1]

The ACS Green Chemistry Institute Pharmaceutical Roundtable identified the development of greener processes for peptide active substances as a critical unmet need in 2016, and the 2019 perspective reporting that work noted the state of the art relied primarily on legacy technologies using large amounts of highly hazardous reagents and solvents [22]. The 2024 analysis sets out what is being done about it: N-butylpyrrolidinone performs better than dimethylformamide on racemisation and aspartimide formation, though its viscosity gives inferior resin swelling; binary mixtures such as dimethyl sulfoxide with dioxolane, 2-methyltetrahydrofuran or ethyl acetate have been screened for the deprotection step; reducing piperidine concentration showed no significant difference in SPPS performance between 5%, 10% and 20% in dimethylformamide; and flow washing after Fmoc removal, rather than batch washing, cut solvent consumption by more than half [1]. One reported recycling scheme recovered 86% of ethyl acetate, 70% of dimethyl sulfoxide and 38% of Oxyma for reuse [1].

Safety sits alongside sustainability. The same analysis flags hydrogen cyanide formation from DIC and Oxyma, sensitising coupling agents, highly corrosive trifluoroacetic acid, and solvents such as dichloromethane and the ethers, and states that “moving forward, alternatives to TFA are recommended from a safety/sustainability standpoint” [1]. Piperidine is also a controlled chemical on the US Drug Enforcement Administration’s List I, which is a supply-chain constraint as much as a chemical one [1].

What to ask a peptide synthesis supplier

1. Which route, and why? SPPS, LPPS, hybrid, recombinant or semi-synthetic, with the reasoning tied to length, scale and modifications [1].
2. What are the starting materials? Protected amino acid derivatives and preloaded resins are generally acceptable as starting materials; longer or modified fragments usually are not [9].
3. What is the crude purity before chromatography? It predicts purification yield, gradient length and how much repurification will be needed [9] [20].
4. How many orthogonal purification steps? Two to three combined reversed-phase and ion-exchange steps are the norm for an active substance [1].
5. Which counter-ion is delivered, and how is it set? Ask whether a final ion-exchange step is part of the process [1] [23].
6. How is the material isolated and what is the residual moisture? Lyophilization is standard; the alternatives change the solid-state properties [9].
7. Which impurities are reported, identified and qualified? EMA points to the European Pharmacopoeia thresholds for peptide-related impurities: reported above 0.1%, identified above 0.5% and qualified above 1.0%. Deletion, truncation, insertion, diastereomeric and cleavage-derived impurities are route-specific [9].

How Winhycare Pharma makes peptides

Winhycare Pharma supplies research peptides made by solid-phase synthesis as research-use-only material, at purity of at least 99% as stated in our product brochure, and offers custom peptide synthesis for sequences outside the catalogue. Semaglutide, tirzepatide, retatrutide, cagrilintide and BPC-157 are also supplied as API to licensed pharmaceutical manufacturers through our peptide API hub; US-DMF and CTD Module 3 support is available for semaglutide only. Process and analytical development, scale-up and technology transfer are offered as peptide CRDMO services from Hyderabad, India, and client-supplied peptide can be taken into vials through our lyophilisation service. To discuss a sequence, use the custom synthesis enquiry form.

For laboratory research use only. Not for human or veterinary use.