The enzyme in question is D-apionate oxidoisomerase, abbreviated ApnO, a protein first discovered in 2018 that belongs to a very small group of enzymes that cannot be clearly assigned to a single Enzyme Commission class. Alongside ketol-acid reductoisomerase, known as KARI, ApnO stands out because its single active site supports two distinct reactions. Until now, however, ApnO had never undergone detailed functional and structural analysis, leaving open the question of how, at the molecular level, one catalytic pocket manages to accommodate chemistry that classification rules say should require separate enzymes.

To answer that question, researchers at Ehime University turned to X-ray crystallography, the technique that remains the gold standard for resolving the three-dimensional arrangement of atoms within a protein. By crystallizing ApnO and diffracting X-rays through the crystals, the team determined its structure at a level of detail sufficient to identify not only the fold of the protein but also the positions and coordination geometries of the metal ions sitting at the heart of its active site. Those metal ions turned out to hold the key to understanding how this unusual enzyme works.

The first surprise came from a comparison of overall architecture. Despite showing no significant sequence homology with KARI, meaning the two proteins share little detectable similarity in their linear chains of amino acids, ApnO turned out to closely resemble KARI in its overall three-dimensional structure. This is a classic example of structural convergence or distant evolutionary relationship hidden at the sequence level: two proteins can arrive at similar folds while their primary sequences drift beyond recognition. For enzymologists, the resemblance was immediately suggestive, because KARI is the other famous member of the small set of enzymes that defy single-class classification.

KARI’s catalytic mechanism depends on two magnesium ions, designated Mg(I) and Mg(II), which sit at two distinct binding positions, site I and site II, within the active site. In KARI, both metal ions adopt octahedral coordination geometry, a highly symmetric arrangement in which six ligands surround the metal at the vertices of an octahedron. This divalent metal pair helps stabilize the charged intermediates that arise during the reductoisomerase reaction. The expectation going into the study was that ApnO, with its structurally similar fold, might employ a comparable metal arrangement.

Instead, the crystal structure revealed something strikingly different. At the position equivalent to KARI’s Mg(II) site, ApnO binds zinc or manganese ions, and the coordination geometry there is tetrahedral rather than octahedral, meaning four ligands surround the metal in a pyramid-like arrangement. Meanwhile, at the position equivalent to Mg(I), a manganese ion adopts the expected octahedral geometry. The result is a hybrid active site in which one metal position mirrors the canonical arrangement seen in KARI while the other departs from it in both metal identity and geometry, a combination not previously observed in this structural family.

Distinguishing which specific metal ions occupy which sites in a crystal structure is not trivial, because metals of similar size can be difficult to tell apart by electron density alone. The research team overcame this challenge by exploiting the anomalous dispersion effect, a physical phenomenon that occurs when X-ray wavelengths are tuned near the absorption edges of particular elements. At those wavelengths, atoms of a given element scatter X-rays in a characteristic way, allowing zinc and manganese to be identified and distinguished unambiguously within the crystal. This careful metal identification gave the researchers confidence in their assignment of Zn(II) or Mn(II) to site II and Mn(II) to site I.

The detailed comparison between ApnO and KARI extended beyond the metals themselves to the protein residues that surround them. The zinc or manganese ion at site II in ApnO and the magnesium ion at the corresponding position in KARI differ not only in the type of metal ion and in coordination geometry but also show no conservation of the surrounding active-site residues. In other words, the amino acids that grip the metal in ApnO are not the same, nor even recognizably related, to those that grip the metal in KARI. This strongly suggested that the site II metal-binding capability was acquired independently in the two enzymes, an example of convergent evolution at the level of a single catalytic position.

Site I tells a different evolutionary story. Because the metal ion at that position is coordinated in a comparable octahedral fashion in both enzymes, the researchers concluded that this metal-binding site likely shared a common evolutionary origin, passed down from a shared ancestor even as the rest of the proteins diverged. The picture that emerges is therefore a mosaic: one metal position inherited from a common ancestor, the other invented separately in each enzyme lineage, together producing an active site capable of catalyzing two distinct reactions within a single structural framework. Such mosaic evolution may be a general feature of the rare enzymes that blur the boundaries of the Enzyme Commission classification.

The study also carries a broader message about structural biology in the era of accurate computational prediction. With modern machine-learning tools now able to predict protein structures with remarkable precision, one might ask why laborious experimental structure determination remains necessary. The ApnO work provides a clear answer: the direct identification of metal ions and the discovery of novel binding modes at active sites cannot be reliably revealed by predicted models alone. Metal identity, coordination geometry, and anomalous scattering behavior are experimental observables, and in metalloenzymes they can make the difference between a plausible guess and a verified mechanism. By combining crystallography with anomalous dispersion analysis, the Ehime University team has not only explained how ApnO manages its dual chemistry but also offered new insight into the molecular evolution of enzymes that defy the general paradigm in which a single enzyme’s activity is classified under a single reaction mode.

Subject of Research: Structural basis of dual-reaction catalysis by the rare enzyme D-apionate oxidoisomerase at a single active site

Article Title: Structural basis for enzyme catalyzing two distinct reactions at a single active site

Article References: Structural basis for enzyme catalyzing two distinct reactions at a single active site. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: ApnO, D-apionate oxidoisomerase, ketol-acid reductoisomerase, X-ray crystallography, enzyme classification, metal coordination, anomalous dispersion, molecular evolution, active site, zinc, manganese, Journal of Biological Chemistry

The central idea behind the study is elegantly simple. Instead of training a neural network from scratch on raw amino acid sequences, a process that demands enormous computational resources and vast amounts of labeled data, the researchers started from pre-trained protein embeddings derived from UniProt, the widely used open-access protein database. These embeddings represent each protein as a numerical vector, in this case a string of 1,024 numbers, that encodes structural and functional information the underlying language model learned from hundreds of millions of sequences across all kingdoms of life. In effect, the model arrives at the plant classification problem already fluent in the language of proteins, and it only needs to learn the narrower task of deciding whether a given plant protein is an enzyme. This approach dramatically reduces the data and compute required, making sophisticated protein function prediction accessible to research groups without supercomputing budgets.

To test the framework rigorously, the team assembled a curated dataset of 22,267 unique protein sequences drawn from four major plant species that span enormous evolutionary and agricultural ground: Arabidopsis thaliana, the small weed that serves as the workhorse of plant genetics; Brassica species, which include oilseed rape and many vegetables; Oryza sativa, rice, the staple crop feeding half the world; and Triticum aestivum, wheat, whose large and complex genome has long challenged biologists. Each sequence in the dataset was converted into its 1,024-dimensional embedding vector, creating a numerical portrait of every protein that a classifier could then learn from.

One of the most serious pitfalls in machine learning applied to biological sequences is data leakage. Proteins that share recent common ancestry often retain similar sequences, and if close relatives of the same protein family end up in both the training set and the test set, a model can appear far more accurate than it really is by essentially recognizing family members rather than learning genuine functional signals. The researchers confronted this problem head-on using a homology-aware data splitting strategy. They applied CD-HIT, a standard clustering tool, at a sequence identity threshold of 60 percent, grouping the 22,267 proteins into 13,784 clusters. Entire clusters were then assigned to either training or test partitions, ensuring that no pair of proteins sharing more than 60 percent sequence identity could straddle the boundary. This design choice makes the reported performance figures considerably more trustworthy than those of many earlier studies that ignored sequence similarity when splitting data.

With a clean evaluation pipeline in place, the team systematically benchmarked ten different models, ranging from simple baselines such as logistic regression to baseline deep neural networks and their centerpiece, an attention-enhanced deep neural network. Attention mechanisms, the innovation behind much of the recent progress in artificial intelligence, allow a network to weigh the relative importance of different parts of its input. In this context, attention lets the classifier focus on the dimensions of the embedding vector that carry the most informative signals about enzymatic function, rather than treating all 1,024 features as equally relevant. The attention-enhanced network consistently outperformed the alternatives, achieving a test area under the ROC curve of 0.987 plus or minus 0.001, an accuracy of 0.955 plus or minus 0.004, an F1-score of 0.939 plus or minus 0.005, and a Matthews correlation coefficient of 0.903 plus or minus 0.008. For a binary classification task of this difficulty, an MCC above 0.9 represents a very strong balance of sensitivity and specificity.

Impressive as those numbers are, the authors went one step further to answer a question that matters enormously for real-world applications: does a model trained on some plant species work on species it has never seen? They evaluated this using a Leave-One-Species-Out protocol, in which the model is trained on three of the four plant species and tested entirely on the fourth. The framework held up strikingly well under this stress test, with an average LOSO AUC of 0.985 plus or minus 0.008. This robust cross-species generalization suggests that the embeddings capture features of enzymatic function that transcend species boundaries, meaning a classifier trained largely on well-annotated Arabidopsis proteins could plausibly be deployed on crops whose protein functions are far less characterized.

The statistical rigor of the study also deserves attention. Rather than relying on a single favorable run, the authors benchmarked across many hyperparameter configurations and applied the Friedman test, a non-parametric statistical test for comparing multiple methods across repeated evaluations. The test confirmed highly significant differences among the ten models, with a p-value below 0.001, and the proposed attention-enhanced framework won 67 percent of the hyperparameter configurations examined. The authors are careful and honest about the magnitude of the attention mechanism’s contribution, describing it as modest yet statistically significant. That kind of measured claim, backed by formal statistical testing rather than cherry-picked results, is exactly what distinguishes credible machine learning work in bioinformatics from the hype that sometimes surrounds the field.

The practical implications reach well beyond the leaderboard. Accurate enzyme classification underpins genome annotation, metabolic pathway reconstruction, and the identification of enzymes that could be engineered for crop improvement or industrial biotechnology. For wheat and Brassica crops in particular, where functional annotation lags behind genome sequencing, a fast and reliable computational filter for enzymatic function could accelerate the discovery of enzymes involved in yield traits, disease resistance, or stress tolerance. Because the method relies on pre-trained embeddings rather than raw sequence training, it can be run on modest hardware, lowering the barrier for laboratories in developing countries and for plant breeding programs that lack dedicated machine learning infrastructure. The study received no specific grant funding, and the authors declare no competing interests, with the work published open access under a Creative Commons license.

The work also fits into a broader and rapidly accelerating trend. Protein language models trained on UniProt and similar databases have already reshaped protein structure prediction and function annotation across biology, and this study demonstrates that the same transfer learning paradigm delivers state-of-the-art results in the specific and agriculturally important domain of plant enzymes. The combination of homology-aware evaluation, cross-species testing, and systematic statistical comparison sets a methodological standard that future studies in computational proteomics would do well to follow. As plant genomes continue to be sequenced at a pace far outstripping experimental characterization, tools like this framework will become indispensable for translating raw sequence data into biological knowledge.

What makes this research genuinely exciting is the convergence it represents. A public database built by a global community, a general-purpose AI technique borrowed from natural language processing, and a carefully curated plant protein dataset come together to solve a problem central to food security and biotechnology. The result is not a speculative prototype but a validated, statistically scrutinized classifier that performs near flawlessly on held-out data and generalizes across species as different as a mustard relative and a cereal grain. If the pattern holds as embeddings improve and datasets grow, the era in which every newly sequenced plant genome arrives pre-annotated with reliable enzyme predictions may be closer than anyone expected.

Subject of Research: Transfer learning with UniProt protein embeddings for plant enzyme classification

Article Title: Transfer learning with attention-enhanced deep neural networks for plant enzyme classification using UniProt protein embeddings

Article References: Shahzad, A., Akram, S., & Mehmood, T. (2026). Transfer learning with attention-enhanced deep neural networks for plant enzyme classification using UniProt protein embeddings. BMC Bioinformatics. https://doi.org/10.1186/s12859-026-06623-9

Image Credits: AI Generated

DOI: 10.1186/s12859-026-06623-9

Keywords: deep learning, transfer learning, enzyme classification, plant proteomics, UniProt embeddings, attention mechanism, bioinformatics, protein function prediction, Arabidopsis thaliana, Oryza sativa, Triticum aestivum, cross-species generalization