TL;DR: Researchers reported a computationally guided workflow called DeepPCD that lifted the hydrolytic efficiency of a newly identified nitrilase by 1.3 × 10⁵-fold after changing only seven candidate residues.
A new nitrilase meets a new design workflow
A research team reported a novel nitrilase, Bh NIT, identified from Bradyrhizobium huanghuaihaiense through a computationally assisted screening strategy, according to the abstract of Designing a DeepPCD Strategy for Efficient Engineering Nitrilases toward Industrial Bioconversion | ACS Synthetic Biology | ACS Publications.
The authors listed for the study are Xiao-Ting Zhou, Qian Guo, Yao-Fang Ding, Qi Shen, Ya-Ping Xue and Yu-Guo Zheng, with Ya-Ping Xue identified as the corresponding author through the contact address [email protected].
The publisher is the American Chemical Society, and the article is identified as a Research Article in ACS Synthetic Biology.
The publication history states that the manuscript was received on January 15, 2026, that a revision was received on September 22, 2026, that acceptance followed on September 23, 2026, and that online publication occurred on October 03, 2026.
The authors declared no competing financial interest.
Support acknowledged in the publication record includes the Taishan Industrial Experts Program, the Zhejiang Provincial Science and Technology Plan Project under award identifier 2024C03011, and the National Natural Science Foundation of China under award identifier 32571726.
The central claim is that integration of artificial intelligence with protein engineering has created opportunities for enzyme discovery and improvement of overall catalytic performance, and the team presents its work as an experimentally validated case study for computationally guided nitrilase engineering.
What DeepPCD organizes
The team named its approach DeepPCD and described it as an integrated computationally guided workflow.
According to the abstract of Designing a DeepPCD Strategy for Efficient Engineering Nitrilases toward Industrial Bioconversion | ACS Synthetic Biology | ACS Publications, DeepPCD organizes sequence-design, structure-filtering, stability-assessment, evolutionary-compatibility and molecular dynamics analyses.
The stated purpose of that organization is to prioritize experimentally testable mutations.
The comparison offered in the abstract is with approaches that require extensive experimental screening of large mutant libraries.
Relative to those extensive-screening approaches, the authors state that the DeepPCD workflow facilitates efficient enzyme engineering modifications.
The target protein for engineering in the reported work was Bh NIT itself.
The supporting materials described for the study list computational prediction of nitrilase thermostability and catalytic parameters, sequence alignment, sequence similarity analysis and phylogenetic analysis of candidate nitrilases.
Those supporting materials also list sensitivity analysis of cutoff values used during candidate nitrilase screening, as well as pH and temperature profiles of Bh NIT.
The supporting description further lists HPLC calibration curves and FT-IR characterization of 3,5-DCPN, 3,5-DCPA and 3,5-DCPM.
Secondary structure analysis, AlphaFold3 structural modeling and model quality assessment of Bh NIT are also listed among the supporting contents.
Seven residues, large measured changes
The engineering result emphasized by the authors involved modification of only seven candidate residues of Bh NIT.
From that limited set of changes, the team reported obtaining a mutant exhibiting a 1.3 × 10⁵-fold improvement in hydrolytic efficiency.
That figure represents the headline activity outcome for the engineered variant compared with the starting enzyme described in the abstract.
The same mutant was reported to show a 20.4°C increase in melting temperature.
The authors also reported a broader operational pH range extending toward more alkaline conditions for the mutant.
They linked those combined changes to substantially enhanced potential for industrial applications.
The description of broader alkaline tolerance is important because industrial bioconversion often benefits from enzymes that remain functional outside narrow laboratory conditions, although the abstract does not assign a specific new pH interval with numerical limits.
The temperature and pH profiles of Bh NIT and M5-GGAEM are listed among the supporting figures, indicating that the comparison between the starting enzyme and an engineered variant was documented beyond the abstract summary.
Specific hydrolytic and hydration activities of purified Bh NIT and mutants are listed as Table S6 in the supporting description.
Temperature and pH profiles, electrostatic potential analysis and substrate-binding energy calculations are likewise listed in the supporting description as Figures S13 through S16.
Why pocket shape and access matter
Molecular dynamics simulations and structural analysis were used to interpret the improved behavior.
According to the abstract, the enhanced performance stems from effective reshaping of the substrate-binding pocket and access channel, along with a strengthened interaction network.
Pocket volume, tunnel radius and docking energy analysis of nitrilases from different sources are listed as Table S2 in the supporting materials.
Molecular dynamics simulations and non-covalent interaction analysis are listed as Figures S17 through S19.
A distribution map of Bh NIT mutation sites on the structure, plus structural comparison of wild-type Bh NIT and engineered variants, are listed as Figures S20 and S21.
Structural and sequence alignment of ProteinMPNN-designed variants with the AlphaFold3-predicted structure of the target enzyme is listed as Figure S22.
The supporting description also lists ProteinMPNN-generated sequence analysis and structural filtering as Figures S8 through S11.
Evolutionary conservation analysis of candidate mutation sites is listed as Figure S12.
Electrostatic potential analysis and substrate-binding energy calculations are listed as Figures S14 through S16.
Changes in pKa values of the catalytic triad are listed as Table S7.
Primers used for combinatorial mutagenesis are listed as Table S5, while computationally predicted beneficial mutations are listed as Table S4.
Sensitivity analysis of key threshold values is listed as Table S3.
Together, those entries show that sequence generation, structural filtering, stability evaluation, evolutionary review and dynamics simulation each contributed documentation to the prioritization process.
Transferability and stated limits
The team also described preliminary computational application to additional nitrilases.
That preliminary work suggested the potential transferability of the workflow within the nitrilase family, according to the abstract.
The wording remains cautious: the application to other nitrilases is described as preliminary and computational, and the suggested transferability is framed as potential rather than demonstrated industrial deployment.
The authors state that the work provides an experimentally validated case study for computationally guided nitrilase engineering.
They also state that it offers a practical workflow for prioritizing mutations that may improve activity and stability.
The final qualification in the abstract is that broader generalization will require further validation across more diverse enzyme systems.
That qualification matters for readers tracking AI-guided protein design, because a workflow validated on Bh NIT and examined computationally on related nitrilases is not presented as proven across unrelated enzyme families.
Why industry follows nitrilase engineering
A separate nitrilase study published days earlier helps explain why activity, stability and operating range attract industrial attention.
The study Immobilization of Alginate Microbeads With Whole‐Cells Expressing Nitrilase for Continuous Nicotinic Acid Production in Chip Millireactor - Bassut - Biotechnology and Bioengineering - Wiley Online Library was first published on September 30, 2026.
Its authors are Jonathan Bassut, Emina Mehić, Luana Cuvillier, Carine Vergne-Vaxelaire, Jean-Louis Petit, Maïté Michaud, Zoé Anxionnaz-Minvielle, Anne Zaparucha and Guillaume Nonglaton.
The affiliations listed include CEA, LETI, Univ. Grenoble Alpes in Grenoble, France, and Génomique Métabolique, Genoscope, Institut François Jacob, CEA, CNRS, Université Evry Paris-Saclay in Evry-Courcouronnes, France, with additional CEA, LITEN participation.
That study states that biocatalytic production of nicotinic acid using nitrilases has emerged as a promising alternative to conventional chemical synthesis, offering improved selectivity and environmental sustainability.
The system it developed was a continuous-flow millireactor employing Escherichia coli cells expressing a nitrilase from Paraburkholderia phymatum, referred to as Nit Phym, for continuous production of nicotinic acid.
Nicotinic acid is also identified in that study as niacin or vitamin B3.
The same study notes that nicotinic acid is a precursor of nicotinamide adenine dinucleotide and nicotinamide adenine dinucleotide phosphate cofactors, identified as NADH/NAD+ and NADPH/NADP+.
It further states that the compound plays a role in biochemical processes including energy production processes such as production of adenosine triphosphate, DNA repair and synthesis, tissue oxidation and lipid management.
The study adds that the majority of global nicotinic acid production is used as a nutritional additive, primarily in livestock and pet feed, and that it is also employed as a dietary supplement for humans.
What continuous testing showed
The French-led millireactor work compared two immobilization strategies for whole cells.
One strategy was covalent attachment to reactor walls.
The other strategy was encapsulation in alginate microbeads.
Performance of those immobilization techniques was evaluated in millireactors with two geometries: pillar and zigzag configurations.
Encapsulation of Nit Phym-expressing E. coli cells in alginate microbeads was achieved through a centrifuge-assisted microencapsulation method.
That method yielded beads with diameters of up to 350 µm.
Continuous-flow reactions demonstrated that the microencapsulated cells maintained activity over at least 7 days.
Those reactions achieved stable conversion rates of approximately 30% in both pillar and zigzag millireactors.
Compared with direct covalent immobilization, microencapsulation significantly reduced cell leaching and contamination of the reaction product.
The authors of that millireactor study highlighted the potential of whole-cell microencapsulation for enhancing stability and efficiency of biocatalytic processes in continuous-flow millireactors.
For readers of the DeepPCD report, the parallel is direct: one research line improves the enzyme molecule itself, while another improves how nitrilase-containing cells are retained and operated in flow.
Publication details readers can verify
The DeepPCD article carries the digital identifier 10.1021/acssynbio.6c00038, as reflected in its ACS record and in the related catalog entry that lists the journal as ACS Synthetic Biology and the publication date as 2026-10-03.
The supporting information is stated to be available free of charge at the ACS address associated with that identifier.
The catalog entry identifies the authors as Xiao-Ting Zhou, Qian Guo, Yao-Fang Ding, Qi Shen, Q. Shen and others, consistent with the author list shown on the ACS page.
The ACS page identifies the work as accepted on September 23, 2026 and published online on October 03, 2026.
The millireactor comparison study is identified by its Wiley record as first published on September 30, 2026, with correspondence to Guillaume Nonglaton at [email protected].
Its author details include ORCID identifiers for Jonathan Bassut, Emina Mehić, Luana Cuvillier, Carine Vergne-Vaxelaire, Jean-Louis Petit, Zoé Anxionnaz-Minvielle, Anne Zaparucha and Guillaume Nonglaton.
Taken together, the two records place computationally guided nitrilase redesign and continuous nitrilase-based production in the same late-September to early-October 2026 publication window.
What to watch next
The DeepPCD authors do not claim that every future nitrilase will respond the same way.
Their reported outcome is specific: seven candidate residues changed in Bh NIT, a 1.3 × 10⁵-fold gain in hydrolytic efficiency, a 20.4°C rise in melting temperature and operation extended toward alkaline pH.
Their mechanistic explanation is also specific: reshaped substrate-binding pocket and access channel plus a strengthened interaction network.
Their forward-looking statement is limited to preliminary computational application to additional nitrilases and suggested potential transferability within the nitrilase family.
Broader use across more diverse enzyme systems is explicitly left to further validation.
That framing makes the next verifiable steps straightforward to track: experimental testing of DeepPCD-prioritized mutations in other nitrilases, publication of full activity and stability datasets, and independent replication of the Bh NIT mutant behavior under industrial process conditions.