Machine Learning-Guided Identification of PET Hydrolases from Natural Diversity

Norton-Baker, Brenna [National Renewable Energy Laboratory (NREL), Golden, CO (United States); BOTTLE Consortium, Golden, CO (United States); USDOE Agile BioFoundry, Emeryville, CA (United States)]; Komp, Evan [National Renewable Energy Laboratory (NREL), Golden, CO (United States); USDOE Agile BioFoundry, Emeryville, CA (United States)] (ORCID:0000000183868814); Gado, Japheth E. [National Renewable Energy Laboratory (NREL), Golden, CO (United States); BOTTLE Consortium, Golden, CO (United States)] (ORCID:0000000200242531); Denton, Mackenzie C. R. [National Renewable Energy Laboratory (NREL), Golden, CO (United States); BOTTLE Consortium, Golden, CO (United States)]; Mathews, Irimpan I. [SLAC National Accelerator Laboratory (SLAC), Menlo Park, CA (United States). Stanford Synchrotron Radiation Lightsource (SSRL)] (ORCID:0000000162543519); Murphy, Natasha P. [National Renewable Energy Laboratory (NREL), Golden, CO (United States); BOTTLE Consortium, Golden, CO (United States)]; Erickson, Erika [National Renewable Energy Laboratory (NREL), Golden, CO (United States); BOTTLE Consortium, Golden, CO (United States)] (ORCID:0000000178069348); Storment, Olateju O. [National Renewable Energy Laboratory (NREL), Golden, CO (United States); BOTTLE Consortium, Golden, CO (United States)]; Sarangi, Ritimukta [SLAC National Accelerator Laboratory (SLAC), Menlo Park, CA (United States). Stanford Synchrotron Radiation Lightsource (SSRL)] (ORCID:0000000227642279); Gauthier, Nicholas P. [Harvard Medical School, Boston, MA (United States); Dana-Farber Cancer Institute, Boston, MA (United States)]; McGeehan, John E. [National Renewable Energy Laboratory (NREL), Golden, CO (United States); BOTTLE Consortium, Golden, CO (United States)] (ORCID:0000000267501462); Beckham, Gregg T. [National Renewable Energy Laboratory (NREL), Golden, CO (United States); BOTTLE Consortium, Golden, CO (United States); USDOE Agile BioFoundry, Emeryville, CA (United States)] (ORCID:000000023480212X)

doi:10.1021/acscatal.5c03460

DOE OSTI · 2945099

Machine Learning-Guided Identification of PET Hydrolases from Natural Diversity

View original source ↗Read source document ↗Export metadata ↓

Abstract

The enzymatic depolymerization of poly(ethylene terephthalate) (PET) is emerging as a leading chemical recycling technology for waste polyester. As part of this endeavor, new candidate enzymes identified from natural diversity can serve as useful starting points for enzyme evolution and engineering. In this study, we improved upon HMM searches by applying an iterative machine learning strategy to identify 400 putative PET-degrading enzymes (PET hydrolases) from naturally occurring homologs. Using high-throughput (HTP) experimental techniques, we successfully expressed and purified >200 enzyme candidates and assayed them for PET hydrolysis activity as a function of pH, temperature, and substrate crystallinity. From this library, we discovered 91 previously unknown PET hydrolases, 35 of which retain activity at pH 4.5 on crystalline material, which are conditions relevant to developing more efficient commercial processes. Notably, four enzymes showed equal to or higher activity than LCC-ICCG, a benchmark PET hydrolase, at this challenging condition in our screening assay, and 11 of which have pH optima <7. Using these data, we identified regions of PETases statistically correlated to activity at lower pH. We additionally investigated the effect of condition-specific activity data on trained machine learning predictors and found a precision (putative hit rate) improvement of up to 30% compared to a Hidden Markov Model alone. Our findings show that by pointing enzyme discovery toward conditions of interest with multiple rounds of experimental and machine learning, we can discover large sets of active enzymes and explore factors associated with activity at those conditions.

Explore related subjects

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CHEMISTRY 97 MATHEMATICS AND COMPUTING PET hydrolase biocatalysis high-throughput assay interfacial biocatalysis machine learning

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Norton-Baker, Brenna [National Renewable Energy Laboratory (NREL), Golden, CO (United States); BOTTLE Consortium, Golden, CO (United States); USDOE Agile BioFoundry, Emeryville, CA (United States)], Komp, Evan [National Renewable Energy Laboratory (NREL), Golden, CO (United States); USDOE Agile BioFoundry, Emeryville, CA (United States)] (ORCID:0000000183868814), Gado, Japheth E. [National Renewable Energy Laboratory (NREL), Golden, CO (United States); BOTTLE Consortium, Golden, CO (United States)] (ORCID:0000000200242531), Denton, Mackenzie C. R. [National Renewable Energy Laboratory (NREL), Golden, CO (United States); BOTTLE Consortium, Golden, CO (United States)], Mathews, Irimpan I. [SLAC National Accelerator Laboratory (SLAC), Menlo Park, CA (United States). Stanford Synchrotron Radiation Lightsource (SSRL)] (ORCID:0000000162543519), Murphy, Natasha P. [National Renewable Energy Laboratory (NREL), Golden, CO (United States); BOTTLE Consortium, Golden, CO (United States)], Erickson, Erika [National Renewable Energy Laboratory (NREL), Golden, CO (United States); BOTTLE Consortium, Golden, CO (United States)] (ORCID:0000000178069348), Storment, Olateju O. [National Renewable Energy Laboratory (NREL), Golden, CO (United States); BOTTLE Consortium, Golden, CO (United States)], Sarangi, Ritimukta [SLAC National Accelerator Laboratory (SLAC), Menlo Park, CA (United States). Stanford Synchrotron Radiation Lightsource (SSRL)] (ORCID:0000000227642279), Gauthier, Nicholas P. [Harvard Medical School, Boston, MA (United States); Dana-Farber Cancer Institute, Boston, MA (United States)], McGeehan, John E. [National Renewable Energy Laboratory (NREL), Golden, CO (United States); BOTTLE Consortium, Golden, CO (United States)] (ORCID:0000000267501462), Beckham, Gregg T. [National Renewable Energy Laboratory (NREL), Golden, CO (United States); BOTTLE Consortium, Golden, CO (United States); USDOE Agile BioFoundry, Emeryville, CA (United States)] (ORCID:000000023480212X). 2025-09-03. Machine Learning-Guided Identification of PET Hydrolases from Natural Diversity. https://doi.org/10.1021/acscatal.5c03460

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related reports

DOE OSTI2026

Frontiers in divalent $f$-block chemistry

Divalent f-block chemistry has undergone rapid expansion in recent years, driven by advances in the stabilization of low-valent lanthanide and actinide complexes. Although f-elements were historically accessed in the +3 or higher oxidation states, the isolation of +2 species has revealed unusual spectroscopic signatures, distinct bonding motifs, and multiple accessible electronic configurations that influence their chemical and physical properties. These developments have generated opportunities in areas including quantum information science, molecular magnetism, catalysis, and luminescence. Computational chemistry has played a pivotal role in interpreting the electronic structure and reactivity of these systems. However, accurately modeling divalent f-block complexes remains challenging because of strong electronic correlation, multiconfigurational character, and the presence of close-in-energy competing electronic states. As experimental capabilities and theoretical methodologies continue to advance, a comprehensive assessment of divalent chemistry is both timely and needed. In this Review, we integrate experimental and theoretical perspectives to provide a systematic analysis of divalent lanthanide and actinide complexes across the f-block series. We critically assess the role of ligands in determining competing ground-state configurations (f n d 1 vs. f n+1 ) resulting in their distinct spectroscopic, magnetic, and bonding properties. Finally, we discuss emerging strategies for predictive modeling and rational design of low-valent f-block molecular systems, with potential applications ranging from dinitrogen reduction to quantum technologies.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2026

Unveiling the electronic structure and chemical bonding of the deprotonated cisplatin anion [(NH 3 )(NH 2 )PtCl 2 ] − via low-temperature photoelectron spectroscopy and theoretical calculations

The dehydrogenated cisplatin anion, [(NH 3 )(NH 2 )PtCl 2 ] − , was investigated via low-temperature photoelectron spectroscopy and theoretical calculations. Seven and four spectral peaks are respectively resolved at 193 and 266 nm, yielding rich electronic structure information for both the anion and neutral. From the threshold and maximum of the lowest electron binding energy band, the experimental adiabatic (ADE) and vertical detachment energies (VDE) are determined to be 3.3 ± 0.1 and 3.525 ± 0.025 eV, respectively. Theoretical calculations indicate the dominant isomer adopting a cis-geometry, in which the platinum center is coplanar with two chlorine and two nitrogen ligands. The calculated VDE of 3.54 eV based on this structure agrees well with the experimental value. Charge analyses reveal that the excess electron in the anion is primarily localized on the Pt and Cl atoms. A suite of theoretical analysis tools was employed to elucidate the bonding characteristics and interaction strength between Pt and its ligands.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2026

Adsorption, charge transfer and a coverage-driven transition of alkali metals on rutile TiO 2 (110)

The interaction of alkali metals with metal oxide surfaces is central to tuning surface reactivity in heterogeneous catalysis and photocatalysis. Here we present a comprehensive DFT+U study of the adsorption of alkali metals (Li, Na, K, Rb, Cs) on the (110) surface of rutile TiO 2 . At low coverage (θ = 1/8), all alkali metals bind preferentially to bridging oxygen sites with adsorption energies in the range −4.06 to −3.33 eV, transferring nearly one full electron (0.92–0.99 |e|) to the substrate and inducing Ti 4+ → Ti 3+ reduction. The excess charge localizes preferentially at subsurface Ti sites in the form of small polarons. Diffusion barriers indicate facile motion along bridging-oxygen rows, whereas inter-row hopping is strongly hindered. Coverage effects were examined systematically for potassium: adsorption energy and charge transfer per K atom decrease monotonically with increasing θ. Strikingly, a sharp energy discontinuity occurs between θ = 4/8 and θ = 5/8 (ΔE ≈ 1 eV per atom), which we identify as a coverage-driven structural transition arising from steric packing constraints and enhanced K–K electrostatic repulsion once every (1×1) surface cell is occupied. This structural transition perfectly correlates with a dramatic drop in the work function down to an ultra-low minimum of 0.84 eV at θ=5/8, followed by a metallization- driven recovery at higher coverages. Ab initio molecular dynamics simulations confirm zigzag K arrangements at moderate coverage (θ = 1/3), while at high coverage (θ = 2/3) short-range K–K correlations emerge without long-range order. These results provide atomistic insight into the structure–activity relationships underlying alkali promotion effects on oxide-supported catalysts.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗