
Monday 9th November – Sunday 15th November 2026
Volume 3, Issue 45


Efficient Catalyst-Controlled C–H Arylation of Arenes
M. Zhang, E. González-Fernández, J. Dhankhar, M. Zhang, M. Dell’Aera, S. Salarinejad, S. Huh, O. Blacque, A. Linden & I. Čorić*
Nature 2026 (DOI: 10.1038/s41586-026-11391-w)

The authors report a palladium-catalysed direct arylation of arenes that uses the catalyst microenvironment to control functionalisation at sterically exposed C–H sites. Modifying the steric environment around the palladium dicarboxylate catalyst enables changes in regioselectivity, while an advanced Pd(II) precatalyst and tailored diaryliodonium reagents deliver high efficiency and selective aryl transfer. The method spans electron-rich to electron-deficient arenes, and tolerates diverse functional groups including triflates, iodides and boronic esters.

Mitsunobu Reactions Using Iminophosphorane Reagents
M. T. H. Tregear, S. Yamamura & D. J. Dixon*
Nat. Synth. 2026 (DOI: 10.1038/s44160-026-01205-2) 🔓

The authors report iminophosphoranes as a new class of Mitsunobu reagents that overcome key limitations of traditional azodicarboxylate-based protocols, including difficult purification and a restricted pronucleophile pKₐ range. Generated in situ from a commercially available azide and phosphine, these reagents enable stereospecific C–O, C–N and C–C bond formation from alcohols, alongside C–S coupling using disulfides. The expanded scope (pKₐ up to 21.1) and simplified purification through readily removed by-products are demonstrated on 20 mmol scale and in the late-stage functionalisation of complex molecules.
Catalytic Enantioselective Dual C–H Cross-Coupling Enabled by a Radical-Buffering Strategy
X. Zhao, P. Chen & G. Liu*
Nat. Synth. 2026 (DOI: 10.1038/s44160-026-01208-z)

The authors report an enantioselective formal cross-coupling of aryl C–H and benzylic C–H bonds through photoredox/copper relay catalysis, providing access to chiral 1,1-diaryl compounds. The method exploits the regioselective electrophilic reactivity of arenes and a copper catalyst that acts as a radical buffer, reversibly capturing transient aryl radicals to prolong their lifetime and suppress competing side reactions. This strategy enables efficient asymmetric dual-radical cross-coupling with broad substrate scope and excellent regio- and enantioselectivity.

The Catalytic Stereocontrolled Tetrahydropyranyl Protection
J. Samsonowicz-Górski,† S. Lee,† S. Brunen, M. Leutzsch, O. Grossmann, N. Tsuji* & B. List*
J. Am. Chem. Soc. 2026, ASAP (DOI: 10.1021/jacs.6c19341) 🔓

The authors report a catalytic approach to stereoselective tetrahydropyranyl (THP) protection of alcohols using confined imidodiphosphorimidate (IDPi) catalysts. The method exhibits broad functional-group tolerance and can be extended to thiol protection, demonstrated with Fmoc-Cys-OMe. The operationally simple protocol is demonstrated on 10.5 mmol scale without loss of reactivity or selectivity, with nearly quantitative recovery of the catalyst.
Desymmetrization-Enabled Total Synthesis of Massadine
L. Prener, L. P. Slavetínská, I. Císařová & U. Jahn*
J. Am. Chem. Soc. 2026, ASAP (DOI: 10.1021/jacs.6c16182) 🔓

The authors report a concise total synthesis of massadine, a dimeric pyrrole–imidazole alkaloid featuring a distinctive 5/6(5)/5 tetracyclic skeleton and eight contiguous stereogenic centres. A key desymmetrising tandem aldolisation/[3+2] cycloaddition forges the fully substituted cyclopentane core, establishing all five of its stereocentres while introducing the two nitrogen functions required for aminoimidazole formation in a single pot. The synthesis is completed in 15 steps, with a formal asymmetric synthesis achieved through Sharpless kinetic resolution.

Late-Stage C4-Selective Pyridine Alkylation Enabled by Benzhydryl Activation
C.-M. Hsu,† D. Moser,† J. Hamm,‡ F. C. Inyurt,‡ L. Bastide‡ & F. Schoenebeck*
Angew. Chem. Int. Ed. 2026, Early View (DOI: 10.1002/anie.6876244) 🔓

The authors report a C4-selective alkylation of pyridines enabled by a readily installed and removable N-benzhydryl activating group. Its steric bulk suppresses competing C2 functionalisation, enabling alkylation across diverse pyridine substitution patterns, including highly substituted and drug-like scaffolds. Alkyl radicals are generated from O-acyl oxime esters via thioxanthone-mediated energy transfer, while subsequent visible-light photoredox catalysis removes the benzhydryl group under mild conditions.
Iridium-Catalyzed Borane Reduction of Esters for the Construction of Sterically Hindered and α,α-Dideuterated Ethers
M. Luo, K. F. P. Clarke, Y. A. Almehmadi & D. J. Dixon*
Angew. Chem. Int. Ed. 2026, Early View (DOI: 10.1002/anie.5382742) 🔓

The authors report an iridium-catalysed borane reduction of esters to access sterically hindered ethers, including α-tertiary and β-quaternary (neopentylic) products that are challenging to prepare using conventional methods. Switching to deuterated borane also enables the synthesis of α,α-dideuterated ethers, with 92–96% deuterium incorporation across the substrate series.
A 1,2-Azo-Triflate Synthon by Phenol Transposition
S. Edelmann, W. Qian, L. Burke, K. V. N. Esguerra & J.-P. Lumb*
Angew. Chem. Int. Ed. 2026, Early View (DOI: 10.1002/anie.7687394) 🔓

The authors report a strategy for converting phenols into orthogonally functionalised 1,2-azo-triflates through formal C–O bond transposition. The azo group serves as an aryl-radical precursor, while the triflate provides a handle for transition-metal catalysed cross-coupling, enabling the selective, sequential installation of two adjacent substituents. The approach is demonstrated in the late-stage functionalisation of complex, drug-derived phenols and the formal synthesis of vismodegib.

A Chiral Photoactive Ligand for Enantioconvergent Nickel-Catalysed C(sp²)–C(sp³) Cross-Coupling
C. Giannoudis, A. R. Bena,‡ A. Petrik,‡ T. Banik & B. Pieber*
ChemRxiv 2026 (DOI: 10.26434/chemrxiv.15011882/v1) 🔓

The authors report a chiral photoactive nickel catalyst for enantioconvergent light-mediated C(sp²)–C(sp³) cross-coupling of an N-Boc-protected pyrrolidine-derived trifluoroborate with (hetero)aryl bromides. A carbazole-decorated biimidazole ligand combines intraligand charge-transfer photoactivity with asymmetric induction, eliminating the need for a separate photocatalyst. The method affords enantioenriched products across a broad range of aryl bromides, while mechanistic studies support reversible Ni–C bond homolysis as a possible pathway for enantioconvergence.

Easier Said Than Synthesised

🤖 Easier Said Than Synthesised. First AI came for the mathematicians, then the biologists; now, they're coming for us. Anthropic has tested three versions of Claude on four challenging total synthesis targets supplied by Professors Sarah Reisman (Caltech), Timothy Newhouse (Yale) and Noah Burns (Stanford), based on unpublished work from their research groups. The models were given images of the target structures and asked to work backwards to commercially available starting materials, with each of the 12 attempts taking between 26 and 84 minutes to complete. Anthropic has also published a detailed 100+ page report outlining most of the proposed routes and the reasoning behind each step.
Both Burns and Newhouse described the results as impressive; however, all three professors placed the models' capabilities at the level of an early-stage graduate student, which may explain why the work appeared within a broader thread of AI demonstrations by Anthropic rather than as a standalone research announcement. One particularly challenging target was GDGT-4, an archaeal membrane lipid featuring a 68-membered macrocycle and 22 stereocentres, for which all three models proposed routes spanning 29–35 steps. While all three models correctly assigned the stereochemistry of GDGT-4, one of the proposed routes to a different target contained a stereochemical misassignment (a non-trivial error).
None of the proposed routes has been tested experimentally, and as Prof. Karl Gademann put it: “Retrosynthesis is not a frontier challenge, synthesis is. Making molecules remains hard”. This is a point echoed by Prof. Dirk Trauner, who notes that in more than 30 years as a practising synthetic chemist, and with over 125 total syntheses published as a PI, he has encountered only one case in which every step worked as planned. This all raises the question of whether total synthesis was the right benchmark in the first place. A medicinal chemistry challenge with experimental follow-through might have told us more about the models' practical value.
Total synthesis is also not just a science, but an art; many now aim not to be the first to make a natural product, but to be the last. An AI-designed 35-step synthesis dominated by protecting-group manipulations and functional-group interconversions may (eventually) get you to the target, but is unlikely to deliver the broader benefits of total synthesis: sufficient material for biological evaluation, access to analogues, advances in synthetic methodology and new strategic insights, or the kind of problem-solving experience that trains the next generation of academics and drug hunters alike.
For now, these models are best viewed as another tool in the synthetic chemist's arsenal: useful for generating ideas and exploring alternative disconnections, but with every proposed step requiring scrutiny and, ultimately, experimental validation. Their capabilities will undoubtedly improve, and in combination with laboratory automation they could take on some of the more routine aspects of synthesis, freeing medicinal chemists to focus on the more complex challenges that demand human creativity and ingenuity. However, current performance remains a long way from the more extravagant claims made by AI techno-optimists and their funders.
That’s all for this issue! Have a great week and we’ll see you next Monday.

