Depolymerization of polyesters by a binuclear catalyst for plastic recycling

Plastics play an essential role in modern society; however, the relentless growth of their production is threatening both human health and ecosystems. As a result, there are intensive efforts in developing recycling technologies to repurpose waste plastics into the building blocks for valuable materials. Here we show a binuclear complex that can catalyse the degradation of poly(ethylene terephthalate) (PET)—the most widely used polyester globally—and a wide spectrum of other plastics including polylactic acid, polybutylene adipate terephthalate, polycaprolactone, polyurethane and Nylon 66. Inspired by hydrolases, the group of enzymes that catalyse bond cleavages with water, the present catalyst design features biomimetic Zn‒Zn sites that activate the plastic, stabilize the key intermediate and enable intramolecular hydrolysis. This synthetic catalyst delivers an activity of 36 mgPET d−1 gcatal−1 toward PET depolymerization at pH 8 and 40 °C and an activity of 577 gPET d−1 gcatal−1 at pH 13 and 90 °C for scalable PET recycling. We further demonstrate a closed-loop production of bottle-grade PET. This work presents a practical and viable solution for the sustainable management of plastics waste. Plastic pollution forms a major global challenge to the ecosystem. Here the authors show a binuclear catalyst that could degrade various polyesters in an effective and scalable way, providing a promising technological solution to the challenge.

Plastics play an essential role in modern society; however, the relentless growth of their production is threatening both human health and ecosystems. As a result, there are intensive efforts in developing recycling technologies to repurpose waste plastics into the building blocks for valuable materials. Here we show a binuclear complex that can catalyse the degradation of poly(ethylene terephthalate) (PET)-the most widely used polyester globally-and a wide spectrum of other plastics including polylactic acid, polybutylene adipate terephthalate, polycaprolactone, polyurethane and Nylon 66. Inspired by hydrolases, the group of enzymes that catalyse bond cleavages with water, the present catalyst design features biomimetic Zn-Zn sites that activate the plastic, stabilize the key intermediate and enable intramolecular hydrolysis. This synthetic catalyst delivers an activity of 36 mg PET d −1 g catal −1 toward PET depolymerization at pH 8 and 40 °C and an activity of 577 g PET d −1 g catal −1 at pH 13 and 90 °C for scalable PET recycling. We further demonstrate a closed-loop production of bottle-grade PET. This work presents a practical and viable solution for the sustainable management of plastics waste.
The exponential growth of plastic production since the 1950s is heavily linked to fossil feedstocks and a linear take-make-waste plastic economy that contribute to an ongoing white crisis [1][2][3][4][5] . The ubiquity of plastic waste imposes major impacts on the environment, particularly the marine ecosystem 4,6 . The escalating plastic pollution calls for remediation and effective recycling strategies. Poly(ethylene terephthalate) (PET) is the most abundant polyester plastic, accounting for nearly 12% of total solid waste 7,8 . Mechanical processing is a means to recycle PET waste, but it often leads to deterioration of the properties 9 . Chemical recycling allows for the recovery of the individual monomers of PET without compromising functionality. However, the current chemical depolymerization is usually operated under harsh conditions (>180 °C) and/or pressure (20-40 atm), caustic bases (for example, NaOH at 4-20 wt%) or acids (for example, H 2 SO 4 at >87 wt%), resulting in high energy cost and output of hazardous waste 7 . Moreover, a large fraction of post-consumer PET escapes waste management systems and goes into the environment. Hence, it is essential to develop more efficient and environment-friendly recycling and remediation routes to depolymerize plastics under mild or even more favourable conditions. Enzymatic depolymerization shows substantial promise in plastic recycling and environmental remediation under mild conditions. Several plastic-degrading enzymes have been discovered from the habitat of microbes that could consume plastics for metabolism [10][11][12][13][14][15] . For instance, PET can be hydrolysed by cutinases and PETases under ambient conditions 12 , a process that occurs only under harsh conditions by non-enzymatic ways 16 . The capability of PET-degrading enzymes to break down PET under mild conditions arises in part from the proximity effect [17][18][19][20][21] , where the substrate-binding cleft brings the carbonyl and hydroxyl groups into proximity, akin to an effective increase in the local concentration of the reactants. In fact, similar strategies have

PET depolymerization under mild conditions
The PET depolymerization kinetics was first evaluated under mild or environmental conditions (pH 8 or seawater), using an amorphous PET (am-PET, from Goodfellow) as the substrate. In the presence of the Zn 2 /C, the am-PET was completely depolymerized into monomers within 10 weeks at pH 8 and 40 °C (Fig. 3a), with a specific activity of 36 mg PET d −1 g catal −1 . For comparison, conventional hydrolysis catalysts, such as zinc acetate and zinc oxide, were completely inactive at the same condition (Fig. 3a). This comparison suggests the unique activity of the Zn 2 /C for breaking down the chemically inert PET under mild condition.
To further evaluate the performance of Zn 2 /C, we compared it with a commercially available PET-degrading enzyme, namely Humicola insolens cutinase (HiC). It should be noted that commercial HiC is not the best-performing enzyme but we used it as the benchmark catalyst simply because it is widely available across labs. The comparison was conducted at pH 8 and 60 °C, an optimal condition for the commercial HiC. The commercial HiC showed a faster reaction rate at the beginning of the hydrolysis but soon reached a plateau at a conversion of 40% (Fig. 3b). This plateau may result from different causes, such as product inhibition or enzyme denaturation [28][29][30] . Most PET-degrading enzymes, including commercial HiC, have limited capability to depolymerize high-crystalline PET (> 20%) that is commonly found in PET products 11,21,28 (Fig. 3b). In contrast, the hydrolysis kinetics of crystalline PET granule (38%) over Zn 2 /C was in parallel to that of am-PET at pH 8 and 60 °C (Fig. 3b), indicating that the crystallinity of PET feedstock has limited impact on the hydrolysis kinetics of Zn 2 /C.
Given that the average pH of the ocean is around 8.1, Zn 2 /C is of great potential for plastic remediation in the marine environment. As a proof-of-concept demonstration, we investigated the depolymerization of crystalline PET granule (38%) in seawater from the Yellow Sea of China (pH 7.9). In the presence of Zn 2 /C, an ethylene glycol yield of 10% was observed in the seawater at 40 °C for 12 weeks (Fig. 3c).
The control experiments showed that the PET granule (38%) is highly resistant to degradation without the binuclear catalyst or using zinc been employed in many other hydrolases [22][23][24][25] , except that the binding clefts of PET-degrading enzymes are close to the surface of proteins, allowing access to polymeric substrates.
Translating the proximity effect of enzymatic catalysis to the structure design of synthetic plastic-degrading catalysts will help us better cope with the global plastic challenge. One strategy is to construct structural analogues or functional models for the active sites of relevant hydrolases. Mimicry of the active sites of cutinases or PETases is intimidating, as they are constituted by approximately 20 key residues with intricate and precise positioning 18 . We therefore directed our attention to metallohydrolases, which have been demonstrated to be more feasible for biomimetic design 22 . Binuclear metallohydrolases are a wide class of enzymes that utilize two metal centres to give maximal acceleration of various hydrolytic processes via the proximity effect. One of them, organophosphate-degrading enzyme from Agrobacterium radiobacter (OpdA) got on our radar due to its substrate promiscuity 23 . OpdA is a homobinuclear metallohydrolase whose active site consists of two metal ions (for example, di-Zn 2+ , di-Co 2+ and so on). OpdA accelerates the hydrolysis of phosphate via similar principles used by the PET-degrading enzymes despite the superficially disparate substrates. During the reaction, phosphate and hydroxyl bind to adjacent metal centres such that the reactants are brought in close proximity, leading to increased local concentration of nucleophilic hydroxide (Fig. 1a). Meanwhile, the binding increases the electrophilicity of phosphate and stabilizes the intermediate afterwards [23][24][25] . So far, there is no record indicating that OpdA or other binuclear metallohydrolases have activities toward PET degradation. This is probably because their active sites are not accessible to polymers. We hypothesize that the binuclear active site of OpdA can potentially catalyse PET degradation if the surrounding scaffold is removed (Fig. 1b).
Here we synthesize a Robson-type binuclear zinc catalyst for plastic degradation. Its structure resembles the active sites of OpdA. Remarkably, the binuclear zinc catalyst can depolymerize PET at pH 8 as well as in natural seawater. Mechanistic investigations indicate that the efficient PET depolymerization over this binuclear zinc catalyst originates from an intramolecular reaction pathway, exactly matching the characteristics of binuclear hydrolases. This synthetic     using the same Robson-type ligand. The specific activities of the binuclear catalysts follow the order of di-Zn(II) > di-Cu(II) > di-Ni(II) > di-Co(II) > di-Fe(II) (Fig. 3d), at pH 8 and 60 °C. This order is in line with the Lewis acid strength of divalent metal ions. The greater Lewis acidity of Zn(II) could bind the carbonyl group in PET and activate it for nucleophilic attack, as revealed by the following mechanism study (Fig. 4).

Intramolecular reaction pathway
To understand the unusual activity of the catalyst, we followed the catalyst evolution throughout the PET hydrolysis at pH 8 and found that it took an intramolecular pathway to achieve the rate enhancement ( Fig. 4a and Extended Data Fig. 2), exactly as OpdA and many other hydrolases do. We first noticed that the hydrolysis kinetics over the Zn 2 L(NO 3 ) 2 (1) experienced an induction period during which the depolymerization proceeded at a very slow rate (Figs. 3b and 4b). In many cases, the induction period could be an indicator of pre-catalyst activation. Mass spectrometry (MS) was therefore performed to investigate the possible transformation of 1 during this period. It was revealed that 1 evolved to hydroxyl-or water-coordinated binuclear zinc species after 45 h (Fig. 4b). Considering that the latter was likely to be the active catalyst, we dissolved 1 in PET-free NaOH solution (0.1 M) and obtained a hydroxyl-coordinated binuclear Zn complex (Zn 2 L(OH) 2 , 2; Extended Data Fig. 1b). Compared with 1, the initial lag phase was significantly suppressed when 2 was subjected to PET hydrolysis (Extended Data Fig. 3a). These results imply that 1 is a pre-catalyst and in situ generated binuclear zinc hydroxide species most probably participate in the catalytic cycle. We then investigated the adsorption of carbonyl group using diffuse reflection infrared Fourier transform spectroscopy (DRIFTS). With the presence of the binuclear zinc sites, the carbonyl stretch

Fig. 4 | PET depolymerization mechanism over the binuclear catalyst. a,
The key steps accounting for PET hydrolysis over the binuclear catalyst, including the co-adsorption of substrates in the adjacent zinc sites (3) and the stabilization of a six-membered intermediate (4). The blue loops represent the ligand backbones, and the grey lines represent the polymer chains. b, Induction period enlarged from the hydrolysis kinetics in Fig. 3b and the corresponding MS spectra at 15 and 45 h (inset). c, DRIFTS spectra of ethyl benzoate (EB), an analogue to PET but with low boiling point, over the binuclear zinc catalyst (Zn 2 ). The blue curve represents the physical mixture of Zn 2 and EB (Zn 2 /EB) and the red curve shows the adsorption of EB on Zn 2 (Zn 2 -EB). The inset depicts the DRIFT evolution from Zn 2 /EB to Zn 2 -EB during the removal of free and loosely bound EB from the Zn 2 surface. d, Potential energy profile of EGD decomposition on Zn 2 L(OH) 2 compound. The inset shows the DFT-optimized geometries for reactants and intermediates.
Article https://doi.org/10.1038/s41893-023-01118-4 presented a red shift from 1,719 cm −1 to 1,704 cm −1 (Fig. 4c). This suggests a bond-weakening effect on the carbonyl group upon coordination with zinc, whereby the electrophilicity of the substrate is enhanced [32][33][34] . EXAFS evidenced that the coordination number of the zinc indeed increased during the reaction (Extended Data Fig. 4), which came with a geometry switch between in-plane and out-of-plane modes (Extended Data Fig. 5). Density functional theory (DFT) calculations, performed using ethylene glycol dibenzoate (EGD) as a model compound, also support that the binuclear catalyst can stably bind to the hydroxyl and carbonyl groups in the EGD compound via two neighbouring zinc sites (3,4; Fig. 4a,d). The intermediate state (3) ties together the two reactants in a single molecule and transforms the hydrolysis in an intramolecular fashion. At this state, the ester carbonyl is primed for nucleophilic attack by the terminally bound hydroxyl in the immediate vicinity. The nucleophilic attack leads to a cyclic intermediate (4; Fig. 4a). The μ 2 -O bridged binuclear zinc sites are critical for the stabilization of 4 by forming a six-membered ring (Zn-O-C-O-Zn-μ 2 -O). According to DFT, a configurational change of EGD was required to form the six-membered intermediate (4), and this process was the most endothermic step with an energy cost of only 0.64 eV. Following 4, the reaction proceeds via ester bond cleavage. The active centre is temporally coordinated by the newly formed carboxylate (5; Fig. 4a,d) and then restored after a ligand replacement by free hydroxide. We obtained the single crystal of a carboxylate-coordinated binuclear zinc polymer [Zn 2 L(BDC)] n (BDC: 1,4-benzenedicarboxylate) when free hydroxide was depleted in the reaction solution (Extended Data Fig. 1c), providing circumstantial evidence of the last step. The calculated potential energy profiles indicate that the whole process is energetically feasible under the reaction conditions (Fig. 4d). In particular, DFT calculations also suggest that water molecules play essential roles in the catalytic process by stabilizing the under-coordinated Zn atom as well as strengthening Zn-O(EGD) binding, as detailed in Supplementary Information. We conjectured that the coordination of water to the metal site could occur via proton jump from the surrounding water hydrogen-bonded to the hydroxyl group on the zinc, as implied by the SCXRD of Zn 2 L(OH) 2 (2) (Extended Data Fig. 1d) 35,36 .

PET recycling
In addition to accelerating PET degradation under mild or environmental conditions, the Zn 2 /C also provides a solution to scalable PET recycling. We first evaluated the catalytic activities of Zn 2 /C over a wide range of operational conditions (30-90 °C, pH [8][9][10][11][12][13][14]. The Zn 2 /C functioned normally under all conditions and its specific activity increased significantly with the increases in temperature and pH (Extended Data Fig. 3b). Further stability test indicated that the binuclear catalyst remained stable below 340 °C (Extended Data Fig. 3c). We then performed the PET recycling at pH 13 and 90 °C to have a trade-off between productivity and energy-efficiency. The Zn 2 /C delivered a specific activity of 577 ± 35 g PET h −1 g catal −1 (Methods, Fig. 5a and Extended Data Fig. 3d), outperforming the other recycling processes, even though it was performed under relatively lower temperature and/or weaker alkaline condition (Supplementary Table 1). As compared in Supplementary Table 1, the glycolysis (for example, the Volcat process) is typically performed at temperatures above 160 °C in the presence of organic bases such as 1,5,7-triazabicyclo  Table 1). Catalyst reusability is another important measure for industrial practice. The Zn 2 /C exhibited a yield decrease of about 8% after 15 cycles of reuse, with a turnover number of >25,000 realized (Fig. 5b). The percentage of zinc loss per cycle was investigated using inductively coupled plasma optical emission spectroscopy (ICP-OES) (Fig. 5b). There was 4.6% loss of zinc in the first cycle, with the number dropping below 1% in the subsequent cycles. The structure of the spent Zn 2 /C remained the same as its original form according to electron microscopy and X-ray absorption spectroscopy (Extended Data Fig. 6).
As a further step towards real-world applications, we submitted the binuclear catalyst to different PET feedstocks. The post-consumer PET waste is of high crystallinity and complex mixtures. The specific activities of the Zn 2 /C toward different crystalline PET waste reached 65-82% of the activity for am-PET (2.9 mg h −1 mg catal −1 at pH 13 and 60 °C) (Methods, Fig. 5c and Extended Data Fig. 7). The impurities and additives in these post-consumer PET products, that is, dyed bottle, cloth fibre and carpet, did not obviously compromise the catalyst performance, as their activities were on par with the additive-free granule. In addition to the crystallinity issue, plastic waste is usually composed of mixed polymers. This complexity in the waste streams bears another technological challenge 9,37 . The Zn 2 /C presented very consistent performance towards the physical mixtures of PET with a variety of plastics (Fig. 5c), including polystyrene (PS), polyvinylchloride (PVC), polylactic acid (PLA), polycaprolactone (PCL), polypropylene (PP) and polyethylene (PE). These results suggest that the Zn 2 /C is competent in coping with real-life PET waste.
To demonstrate a closed-loop PET recycling (Methods and Extended Data Fig. 8a,b), post-consumer PET waste was processed in a 5 l reactor at pH 13 and 90 °C, with a PET/catalyst ratio of 500 g PET g catal −1 and a solids content of 16.7 wt%. The purity of the recovered tereph thalic acid was above 99% as confirmed by 1 H nuclear magnetic resonance (NMR) and high-performance liquid chromatography (Extended Data Fig. 8c,d). The recovered monomer (terephthalic acid) was then used as raw material to make recycled PET (rPET). The charac teristics of the rPET were similar to those of PET made from virgin pure terephthalic acid (Extended Data Fig. 8e), meeting the standard of bottle-grade PET. A tentative technoeconomic assessment indicated that the above process could generate a net profit of US$26.0 million on 0.1 megatons of PET waste per year (Extended Data Fig. 9 and Supplementary Information).
The synthetic binuclear catalyst features an open-access active site, which can depolymerize a wide spectrum of substrates. This is particularly important for environmental remediation, as there are different types of plastic waste to remove from soil or water. We demonstrated that the Zn 2 /C was catalytically active for 13 different kinds of polyester and Nylon 66 ( ). Aside from polyesters, considerable degradation of Nylon 66 was also observed over this catalyst at pH 13 and 60 °C, with a specific activity of 0.11 mg monomer h −1 mg catal −1 . The capacity of Zn 2 /C to digest recalcitrant polyamide (Nylon 66) warrants further attention, as few enzymes can do this task. It should be mentioned that the substrate universality does not compromise centralized PET recycling, because the collection and sorting of PET waste is already involved in the current recycling industry.

Discussion
In summary, we designed a binuclear zinc catalyst for polyester depolymerization. The biomimetic Zn-Zn sites endow the hydrolysis with an intramolecular character by binding plastic substrate and nucleophile together. This proximity effect can bring enormous rate enhancement under mild conditions (36 mg PET d −1 g catal −1 at pH 8 and 40 °C). This character can also generate energy savings to PET recycling under much less-demanding conditions (pH 13 and 90 °C) compared with conventional chemical recycling (Supplementary Table 1), delivering a specific activity of as high as 577 g PET h −1 g catal −1 . We also showcased closed-loop manufacturing of recycled PET (rPET) of virgin quality. We further demonstrated the compatibility of this binuclear catalyst with complex mixtures in plastic waste and its wide substrate scope. Considering that the structure of the reported binuclear zinc complex can be further tailored in many ways, we anticipate that this work will inspire more studies on efficient catalysts to degrade waste plastics.

Synthesis of catalysts
Binuclear complexes [Zn 2 C 25 H 28 N 6 O 9 , Zn 2 L(NO 3 ) 2 ] were prepared according to previously reported procedures with minor modifications 26,27 . The obtained compounds were characterized by elemental analysis, infrared spectroscopy, powder X-ray diffraction (PXRD) and SCXRD. Single  For Zn 2 /C, high-surface area carbon support (Ketjenblack EC-300J, 20 mg) was dispersed in 20 ml of methanol/water mixture (1:1 v/v). The dispersion was ultrasonicated for 10 min. In the meantime, 4 mg of Zn 2 L(NO 3 ) 2 was dissolved in 10 ml of methanol/water mixture (1:1 v/v). Subsequently, the solution of Zn 2 L(NO 3 ) 2 was added into the carbon dispersion and the mixture was ultrasonicated for another 10 min. Finally, the solid product was collected by filtration and dried in a vacuum oven. The content of the metal Zn is 3.1 wt% based on ICP-OES and elemental analysis.

PET hydrolysis under mild conditions
PET hydrolysis was performed in a 20 ml vial placed in an aluminum heating block on a stirring hotplate. The vial was charged with a magnetic stirring bar, 10 mg of PET and zinc-containing catalysts (0.8 mg in terms of zinc). NaOH solution (10 ml, pH 8) or natural seawater (Yellow Sea, China, pH 7.9) was then added to the mixture under stirring. The stirring speed was kept at 400 r.p.m. throughout the reaction, whereby external mass transfer limitations were eliminated. The reaction solution was heated to designated temperatures to initiate the hydrolysis, and aliquots were taken and analysed by 1 H NMR. The hydrolysis of PET yielded the same molar amount of disodium terephthalate and ethylene glycol. Due to the low solubility of disodium terephthalate at pH 8, the 1 H NMR of ethylene glycol was used to calculate the conversion of PET using maleic acid as the internal standard. The pH of the reaction solution was kept constant by adding NaOH. For PET hydrolysis over HiC (0.48 g, Novozym 51032, 6 wt%), the hydrolysis was conducted in the same way except that potassium phosphate buffer (pH 8, 1.0 M) was used instead of NaOH solution.

PET hydrolysis under industrially relevant conditions
PET hydrolysis was optimized at pH 13 and 90 °C in a 250 ml three-necked flask. The flask was charged with a mechanical stirrer, 50 g of PET and 5-50 mg Zn 2 L(NO 3 ) 2 supported on carbon. NaOH solution (50 ml, pH 13) was then added to the mixture under stirring. The stirring speed was kept at 300 r.p.m. throughout the reaction. The reaction solution was heated to 90 °C. The kinetics of the PET depolymerization was determined using base consumption (Extended Data Fig. 3d). The pH during the reaction was kept constant at 13 by adding NaOH.

Post-consumer PET waste recycling
Post-consumer PET waste was collected from dumpsters around campus and shredded into ~10 × 10 mm pieces. Then, we added these pieces into a 5 l flask with Zn 2 /C and adjusted the pH to 13 using NaOH. The reaction mixture was heated to 90 °C while periodically adding sodium hydroxide to maintain pH at around 13. After the reaction, Zn 2 /C was recycled by vacuum filtration using a Buchner funnel. The solid Zn 2 /C was collected on the top of a filter paper to form a black filter cake for reuse. Thereafter, acidification process was performed by adding sulfuric acid to convert soluble disodium terephthalate into terephthalic acid by precipitation. After washing with deionized water twice and drying in an oven overnight, pure terephthalic acid (PTA) was obtained. 1 H NMR and high-performance liquid chromatography (HPLC) were performed to test the purity of the product (Extended Data Fig. 8). The purity of the recycled PTA was above 99% according to the national standard GB/T 30921.1. The content of 4-hydroxybenzaldehyde impurity was 8 mg kg −1 , which is less than that of virgin PTA (11 mg kg −1 ). The recycled PTA was then used to synthesize PET.

Hydrolysis of polymeric substrates
Substrate screening was performed in a 20 ml vial placed in an aluminum heating block on a stirring hotplate. The vial was charged with a magnetic stirring bar, 0.32 g of different PET feedstocks, polyesters or polyamide and 2 mg Zn 2 L(NO 3 ) 2 supported on carbon. NaOH solution (10 ml, pH 13) was then added to the mixture under stirring. The stirring speed was kept at 400 r.p.m. throughout the reaction. The reaction solution was heated to 60 °C to initiate the hydrolysis, and aliquots were taken and analysed by 1 H NMR of the corresponding monomers generated during hydrolysis of different polymeric substrates. The pH was kept constant during the reaction by adding NaOH.

DFT calculations
Spin-polarized DFT calculations were performed using the projectoraugmented-wave method 38,39 and the Perdew−Burke−Ernzerhof exchange-correlation functional 40 as implemented in the Vienna Ab Initio Simulation Package (VASP) 38,41 . A kinetic energy cut-off of 400 eV was employed. Gaussian smearing with a width of 0.05 eV was used. The total energy was converged better than 10 −4 eV atom −1 , and the final force on each atom was less than 0.05 eV Å −1 . The first Brillouin zone was sampled on a Γ point. The van der Waals (vdW) corrections were calculated using the DFT-D2 method of Grimme to describe precise energies with dispersions. To capture the intermediates, the cluster model was constructed within a vacuum cube of 20 × 20 × 20 Å. In the calculation of potential energy profile references, the DFT energies of gaseous H 2 O and hydroxyl (OH − ) were calibrated to liquid water and solvated hydroxyl ion (OH -) in aqueous solution at room temperature according to the standard reference 42 . More details on DFT calculations are provided in Supplementary Information.

Characterization
SCXRD data were collected at 173 K on a SuperNova charge-coupled device (CCD) X-ray diffractometer, with Cu-Kα radiation (λ = 1.54184 Å). PXRD patterns were recorded on a Rigaku D/MAX-2500 diffractometer using a filtered Cu-Kα radiation source (λ = 1.54056 Å). High-angle angular dark field-scanning transmission electron microscopy (HAADF-STEM) images were acquired by a JEM-ARM200F transmission electron microscope operated at 200 kV and equipped with a probe spherical aberration corrector. Matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) was carried out with an AXIMA-Performance mass spectrometer (Shimadzu). Differen tial scanning calorimetry (DSC) measurements were performed with a Q200 calorimeter (TA Instrument). DRIFTS was obtained via a Fourier transform infrared spectrometer (Thermo Nicolet iS50) equipped with a diffuse reflection accessory (Harrick). 1 H NMR was recorded using AVANCE III (Bruker). The XAFS spectra at Zn K-edge were acquired at the 4B9A station of the Beijing Synchrotron Radiation Facility (BSRF), operated at 2.5 GeV with a maximum current of 250 mA. Article https://doi.org/10.1038/s41893-023-01118-4

Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability
All relevant data that support the findings of this study are presented in the article and Supplementary   A description of all covariates tested A description of any assumptions or corrections, such as tests of normality and adjustment for multiple comparisons A full description of the statistical parameters including central tendency (e.g. means) or other basic estimates (e.g. regression coefficient) AND variation (e.g. standard deviation) or associated estimates of uncertainty (e.g. confidence intervals) For null hypothesis testing, the test statistic (e.g. F, t, r) with confidence intervals, effect sizes, degrees of freedom and P value noted

Data analysis
Provide a description of all commercial, open source and custom code used to analyse the data in this study, specifying the version used OR state that no software was used.
For manuscripts utilizing custom algorithms or software that are central to the research but not yet described in published literature, software must be made available to editors and reviewers. We strongly encourage code deposition in a community repository (e.g. GitHub). See the Nature Portfolio guidelines for submitting code & software for further information.

Data
Policy information about availability of data All manuscripts must include a data availability statement. This statement should provide the following information, where applicable: -Accession codes, unique identifiers, or web links for publicly available datasets -A description of any restrictions on data availability -For clinical datasets or third party data, please ensure that the statement adheres to our policy All relevant data that support the findings of this study are presented in the article and Supplementary Information. Source data are provided with this paper.

nature portfolio | reporting summary
March 2021

Human research participants
Policy information about studies involving human research participants and Sex and Gender in Research.

Reporting on sex and gender
Use the terms sex (biological attribute) and gender (shaped by social and cultural circumstances) carefully in order to avoid confusing both terms. Indicate if findings apply to only one sex or gender; describe whether sex and gender were considered in study design whether sex and/or gender was determined based on self-reporting or assigned and methods used. Provide in the source data disaggregated sex and gender data where this information has been collected, and consent has been obtained for sharing of individual-level data; provide overall numbers in this Reporting Summary. Please state if this information has not been collected. Report sex-and gender-based analyses where performed, justify reasons for lack of sex-and gender-based analysis.

Population characteristics
Describe the covariate-relevant population characteristics of the human research participants (e.g. age, genotypic information, past and current diagnosis and treatment categories). If you filled out the behavioural & social sciences study design questions and have nothing to add here, write "See above."

Recruitment
Describe how participants were recruited. Outline any potential self-selection bias or other biases that may be present and how these are likely to impact results.

Ethics oversight
Identify the organization(s) that approved the study protocol.
Note that full information on the approval of the study protocol must also be provided in the manuscript.

Field-specific reporting
Please select the one below that is the best fit for your research. If you are not sure, read the appropriate sections before making your selection.

Life sciences
Behavioural & social sciences Ecological, evolutionary & environmental sciences