Decades of enzymological research framed active sites as fussy and narrowly tuned, built to recognize one substrate and do one job. New research is overturning that dogma, showing that many well-characterized enzymes harbor the capacity for radically new-to-nature chemistry, and drug discovery is just starting to exploit it.
How glycolysis shaped our assumptions about enzymes
For much of the 20th century, enzymology was shaped by a particular class of enzymes: soluble proteins acting on small, soluble metabolites. Glycolytic enzymes like hexokinase, aldolase, and lactate dehydrogenase became the workhorses of early biochemistry. These enzymes were easy to purify, easy to assay, and great for building foundational concepts in kinetics, the lock-and-key model, and later, induced fit.
Those systems taught invaluable lessons, but they also left a lasting imprint on the field. Enzymes came to be viewed as highly specific, tightly constrained catalysts, evolved to recognize a single substrate (or a narrow family of analogues) and convert it along a single, well-defined pathway. The canonical depiction of a precisely shaped active-site pocket accommodating its ideal substrate derives largely from studies of enzymes in central metabolism.
However, that worldview is increasingly strained when applied to large, structurally complex substrates such as peptides, proteins, or densely functionalized natural products. Here, enzymatic catalysis shifts from processing a metabolite to altering the architecture of entire molecules.
Over the past decade, studies of natural product biosynthesis, ribosomally synthesized and post-translationally modified peptides (RiPPs), and both natural and engineered metalloenzymes have provided compelling evidence that the rules and assumptions derived from glycolysis are far too narrow.
Enzymes that build architecture, not just bonds
Natural product pathways are expanding our view of what enzymes can do. In many RiPP systems, a short precursor peptide is transformed by multiple enzymes into a highly structured, often extensively modified final product. These enzymes do more than modify side chains: they can fundamentally reshape the peptide’s backbone and overall topology.
A few examples to show how far this can go:
- Prolyl oligopeptidases as macrocyclases. In amanitin-producing mushrooms, a specialized prolyl oligopeptidase cleaves a precursor peptide and catalyzes its macrocyclization to yield the cyclic octamer alpha-amanitin. The same enzyme that acts as a standard protease in one context can function as a macrocyclase in another.
- Asparaginyl endopeptidases (AEPs) that ligate instead of cleave. Initially categorized as seed storage proteases, certain plant AEPs primarily act as transpeptidases, excising and cyclizing peptides. These enzymes can recognize relatively large peptide substrates and perform cleavage-coupled macrocyclization with impressive efficiency.
- Backbone-N-methylated macrocycles. Fungal RiPP pathways, such as the omphalotin system, can introduce multiple N-methylations along the peptide backbone before macrocyclizing the core peptide. The result is a highly modified macrocycle with drug-like pharmacokinetics, assembled entirely by enzymes acting on a simple, initially unremarkable peptide.
Across bacteria, fungi, and plants, genome mining continues to uncover new enzyme families capable of backbone rearrangements, crosslinks, prenylations, and macrocyclizations. In many cases, the chemistry these enzymes achieve is extremely challenging — or nearly impossible — to reproduce through synthetic chemistry.
Old protein scaffolds, entirely new chemistries
In parallel, a different set of observations is emerging in metalloenzymes such as heme proteins, including cytochrome P450s. These enzymes evolved as oxygenases, inserting oxygen into C–H bonds across a broad range of substrates. However, their active sites are far more versatile than early studies suggested.
By applying directed evolution to natural hemoproteins, researchers have generated enzymes capable of carbene and nitrene transfer reactions. Engineered P450 variants can now catalyze a wide range of transformations, including enantioselective cyclopropanation of alkenes, C–H insertions to form C–N and C–C bonds, cyclopropenation of alkynes, and alkene aminohydroxylation and primary amination.
Many of these reactions rival or exceed the best small-molecule catalysts in terms of selectivity and turnover, and some are unprecedented in traditional organometallic chemistry. Similar observations are now appearing in other enzyme classes. Fe/alpha-ketoglutarate oxygenases have been retooled for non-natural C–H functionalization, dehydrogenases have successfully installed challenging stereocenters in drug intermediates, and artificial metalloenzymes have been built by embedding transition-metal catalysts into protein scaffolds.
The distinction between natural function and new-to-nature chemistry is often a matter of context and engineering, not a hard limit imposed by the protein. Enzymes evolved to handle certain metabolites or substrates, but their active sites and dynamics frequently contain latent catalytic potential that can be uncovered and directed.
From catalysts to modular construction tools
Reframing enzymes from single-reaction specialists to modular construction tools enables new biotechnological applications to appear.
- Programmable macrocycle foundries. Macrocyclic peptides and peptide-like scaffolds are attractive therapeutic modalities as they can be conformationally rigid, resistant to proteolysis, and well-suited for modulating protein–protein interactions. AEPs, macrocyclases, and other RiPP maturation enzymes can be repurposed as general cyclization catalysts for the production of designed peptide libraries.
- Chemoenzymatic “late-stage” diversification. Engineered P450s and other metalloenzymes can be integrated into synthetic routes to introduce highly selective C–H functionalizations or ring-forming steps late in a synthesis, condensing multi-step sequences into a single biocatalytic transformation. This strategy enables focused structure–activity relationship studies on complex scaffolds that are too fragile or elaborate for traditional catalytic methods.
- Plug-and-play RiPP machinery. Because RiPP pathways separate information (precursor peptides) from machinery (enzymes), they naturally lend themselves to modular engineering. Swapping leaders, shuffling core sequences, or recombining enzyme sets allows researchers to design hybrid pathways that install chemistries from multiple natural product families into a single peptide backbone.
In each of these cases, enzymes are being redeployed as generalizable tools for molecular construction rather than as catalysts confined to specific biosynthetic pathways.
Enzymology is still in its infancy
Classical enzymology textbooks largely describe the solutions nature evolved to meet its own metabolic needs, shaped by evolutionary constraints that are largely orthogonal to the goals of drug discovery, materials science, or synthetic biology. Shifting the question from what an enzyme does to what it could do — with a different substrate, cofactor, or a small number of mutations — substantially broadens the landscape of accessible chemistry.
Enzymes capable of macrocyclizing long, N-methylated peptides, P450s that forge C–C and C–N bonds beyond those found in nature, and RiPP enzymes that install chemistries rivaling total synthesis are early signals of what is possible, not the final word.
We are moving from a metabolism-centric view of enzymes as housekeeping catalysts to a broader view of enzymes as programmable molecular machines, capable of architecting complex scaffolds and accessing reactivity that stretches well beyond biology’s historical repertoire
If that shift continues, the most transformative enzymatic functions of the coming decades may not be the ones nature evolved on its own, but the ones we uncover and engineer by challenging the assumptions that glycolysis taught us.
In that sense, enzymes from nature and evolution are mature, but from the human-influence standpoint, enzymology is barely out of infancy. The tools to redesign, repurpose, and recombine enzymes are finally in place. There is nearly unimaginable power in how we choose to deploy the next steps. We will gain a better understanding of biology and be able to reshape it in useful and previously unthinkable ways.
