Tiny Peptide Catalysts Bring Enantioselective Photochemistry Under 750 Daltons

Chemists have long dreamed of building chiral molecules with nothing more than light and a cleverly designed catalyst. A team at the Technical University of Munich has now taken a striking step toward that goal by showing that a peptide of just four amino acids, weighing less than 750 daltons, can harvest visible light and hand its energy to a substrate with remarkable precision. The work, published in Nature Synthesis, demonstrates that these miniature photocatalysts can drive two of the most sought-after reactions in modern asymmetric synthesis: the photochemical deracemization of chiral allenes and olefins, and the enantioselective [2 + 2] photocycloaddition of quinolones. In both cases, the tiny catalysts delivered products in yields reaching quantitative conversion and enantiomeric ratios as high as 99:1, performance that rivals far larger and more elaborate catalytic systems.

The challenge that the Munich group, led by Thorsten Bach, set out to solve is rooted in a fundamental tension in photochemistry. When a molecule absorbs a photon, it is promoted from its ground state to an excited state carrying more than 200 kilojoules per mole of energy. Conventional asymmetric catalysis works by lowering activation barriers along a reaction coordinate, but many photochemical reactions proceed almost barrierlessly once the excited state is populated. A chiral catalyst therefore cannot simply stabilize a transition state after the fact; it must engage the substrate before excitation occurs. Triplet energy transfer, or sensitization, offers one elegant solution. The catalyst absorbs light, undergoes rapid intersystem crossing to a high-lying triplet state, and then transfers that energy directly to a substrate held in its chiral pocket, promoting the substrate to its reactive first excited triplet state while it remains under the catalyst’s stereochemical control.

The idea of asymmetric induction through sensitization dates back to the 1970s, but it was not until 2009 that a catalyst, a chiral xanthone with a lactam binding motif, achieved both high activity and high enantioselectivity. Since then, the field has split into two camps. Small molecular photocatalysts can be exquisitely selective, but each structural modification demands extensive synthetic effort, making optimization slow and substrate-specific. Photoenzymes, by contrast, can be tuned rapidly by directed evolution, yet they require protein frameworks exceeding 20 kilodaltons to position a single photocatalytic chromophore, which limits their quantum efficiency per unit mass and complicates their use in organic solvents. The Munich team asked whether small peptides could bridge this gap, combining the modularity of proteins with the low molecular weight of small molecules.

Their design is deceptively simple. At the N terminus of the tetrapeptide sits pyroglutamic acid, chosen because its lactam ring can recognize and bind lactam-containing substrates through hydrogen bonding. At the C terminus they installed a thioxanthone-based amino acid, abbreviated pTxa, which absorbs visible light and acts as the sensitizing chromophore. Between these two functional ends, just two amino acids proved sufficient: proline and the achiral aminoisobutyric acid, a pair previously shown by the Miller group to induce tight beta-turn structures in short peptides. The thioxanthone amino acid was prepared from (S)-serine, and with three stereogenic centers present, the team synthesized and evaluated all four diastereoisomers of the resulting tetrapeptide in a test deracemization of a chiral allene.

The screening results revealed how critically the peptide backbone matters. Under irradiation at 425 nanometers in dichloromethane, the best diastereoisomer, designated 6a, showed almost no enantiopreference at room temperature, but as the temperature dropped the selectivity climbed steadily, reaching an enantiomeric ratio of 94.5:5.5 at minus 70 degrees Celsius. A preparative run under optimized conditions delivered the allene in quantitative yield with a 96:4 enantiomeric ratio. Strikingly, the three other diastereoisomers performed poorly, underscoring that the three-dimensional arrangement of the peptide backbone, not merely the presence of the chromophore and binding site, governs the stereochemical outcome. Photophysical measurements confirmed that the embedded thioxanthone behaves much like parent thioxanthen-9-one, with a longest-wavelength absorption at 383 nanometers and a triplet energy of 269 kilojoules per mole, high enough to sensitize the target substrates.

To understand why one diastereoisomer succeeded where its mirror-related siblings failed, the team turned to nuclear magnetic resonance spectroscopy and molecular dynamics simulations. At 233 kelvin, NMR experiments yielded 33 distance restraints that defined the free peptide as a well-ordered type II’ beta-turn, with a heavy-atom root mean square deviation of just 0.22 angstroms across the structural ensemble. A hydrogen bond between the amide proton of the thioxanthone amino acid and the backbone carbonyl of pyroglutamic acid stabilizes the turn, and the photoactive thioxanthone ring sits tilted above this fold. When the substrate enantiomer ent-5a was added, strong chemical shift perturbations appeared within the pyroglutamate ring, and the structure of the resulting complex placed the allene roughly 4 angstroms from the thioxanthone ring, approximately the sum of their van der Waals radii and ideal for efficient Dexter-type energy transfer.

The molecular dynamics simulations told an equally compelling story about the losing enantiomer. When the opposite allene enantiomer was modeled in the same binding pose, the smallest achievable allene-chromophore distance exceeded 5 angstroms without breaking the stabilizing hydrogen bonds, and in every simulation the contact was lost within nanoseconds, with the substrate frequently dissociating entirely. The same rapid dissociation occurred for the three alternative peptide diastereoisomers. In other words, only one specific combination of peptide chirality and substrate enantiomer allows simultaneous hydrogen-bonded binding and close chromophore contact. That geometric exclusivity is precisely what deracemization requires: the favored enantiomer is repeatedly excited through energy transfer, racemizes via an achiral intermediate, and the disfavored enantiomer, which binds weakly and fleetingly, accumulates in the photostationary state.

The synthetic scope of the catalysts proved impressively broad. For allene substrates bearing bulky tertiary substituents, enantiomeric ratios remained at 95:5 or better, with aryl-substituted allenes reacting in yields from 82 percent to quantitative and selectivities between 95:5 and 98:2, tolerating halogen, trifluoromethyl, and methoxy groups. When smaller substituents eroded the selectivity, a difluorinated analogue of the catalyst, designed to strengthen the hydrogen-bonding interaction, restored enantiomeric ratios to 94:6 or higher, even for the methyl-substituted allene. A second catalyst variant proved consistently superior for deracemizing tetrasubstituted alkylidenecycloalkanes, where energy transfer to the olefin triggers rotation around the former double bond and racemization through an achiral triplet state. The team also showed that the catalyst can be recovered: in a 0.24-millimole experiment, 70 percent of the peptide was re-isolated after column chromatography without loss of performance.

The final demonstration extended the platform to enantioselective [2 + 2] photocycloadditions of quinolones, a reaction class previously tackled with both molecular thioxanthone catalysts and engineered photoenzymes. The tetrapeptide matched them, converting quinolones to cyclobutane products in yields from 77 percent to quantitative with enantiomeric ratios spanning 89:11 to 99:1. Because the quinolones do not absorb at 425 nanometers, no racemic background reaction competes with the catalyzed pathway. The products were not fully stereospecific, consistent with a triplet 1,4-diradical intermediate that permits bond rotation before ring closure, and the team proposes that the thioxanthone both sensitizes the quinolone and shields one of its enantiotopic faces, directing olefin attack at carbon C4 from a single trajectory.

What makes this work resonate beyond its immediate results is the design philosophy it validates. Peptide catalysts are assembled modularly from the vast pool of amino acids, so tuning a catalyst for a new substrate class becomes a matter of editing a sequence rather than redesigning a synthesis or running an enzyme evolution campaign. Their small size promises better mass-normalized quantum efficiency than photoenzymes, and their solubility in organic solvents sidesteps the handling difficulties that plague protein catalysts in preparative photochemistry. The Munich team is careful to frame the tetrapeptides as a proof of concept, but the combination of structural characterization by NMR, mechanistic validation by simulation, and broad synthetic utility suggests that sub-kilodalton peptide photocatalysts could become a genuine third pillar of enantioselective photochemistry, standing alongside small molecular sensitizers and photoenzymes rather than merely imitating either.

Subject of Research: Small peptide photocatalysts for enantioselective visible-light-driven deracemization and [2 + 2] photocycloaddition reactions via triplet energy transfer

Article Title: Photocatalytic tetrapeptides for enantioselective visible light-induced transformations

Article References: Irle, R. M., Lazzarini, I., Hagn, F., Zacharias, M., & Bach, T. (2026). Photocatalytic tetrapeptides for enantioselective visible light-induced transformations. Nature Synthesis. https://doi.org/10.1038/s44160-026-01173-7

Image Credits: AI Generated

DOI: 10.1038/s44160-026-01173-7

Keywords: photocatalysis, tetrapeptides, enantioselectivity, triplet energy transfer, deracemization, [2 + 2] photocycloaddition, thioxanthone, pyroglutamic acid, NMR spectroscopy, molecular dynamics, asymmetric synthesis, photochemistry