Highly stable and tunable peptoid/hemin enzymatic mimetics with natural peroxidase-like activities

Developing tunable and stable peroxidase mimetics with high catalytic efficiency provides a promising opportunity to improve and expand enzymatic catalysis in lignin depolymerization. A class of peptoid-based peroxidase mimetics with tunable catalytic activity and high stability is developed by constructing peptoids and hemins into self-assembled crystalline nanomaterials. By varying peptoid side chain chemistry to tailor the microenvironment of active sites, these self-assembled peptoid/hemin nanomaterials (Pep/hemin) exhibit highly modulable catalytic activities toward two lignin model substrates 2,2-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) and 3,3’,5,5’-tetramethylbenzidine. Among them, a Pep/hemin complex containing the pyridyl side chain showed the best catalytic efficiency (Vmax/Km = 5.81 × 10−3 s−1). These Pep/hemin catalysts are highly stable; kinetics studies suggest that they follow a peroxidase-like mechanism. Moreover, they exhibit a high efficacy on depolymerization of a biorefinery lignin. Because Pep/hemin catalysts are highly robust and tunable, we expect that they offer tremendous opportunities for lignin valorization to high value products.

P eroxidases are one of the largest groups of oxidoreductases that catalyze the oxidation of many substrates mostly in the presence of H 2 O 2 , offering a wide variety of applications such as environmental protection, food processing, and clinical diagnostics 1 . Peroxidases have shown promising potentials for the valorization of renewable feedstock such as lignin, the largest renewable aromatics on earth 2 . For example, oxidative depolymerization of lignin into oligomers in the presence of peroxidase and H 2 O 2 is a critical step in producing kerosene-type aviation fuels 3 . In these peroxidases, ferric center in hemin serves as an active site and the microenvironment around heme is surrounded by the polypeptide chains 4 . The cleavage of O-O bond of H 2 O 2 occurs around the ferric center to form Fe 4+ =O and the hemin iron is stabilized in the catalytic cycle. In addition, the amino acid residues in the polypeptide chains surrounding the hemin active site, such as histidine (His), tryptophan (Trp), arginine (Arg), and glutamic acid (Glu), play a key role in regulating the peroxidase microenvironment 4 . Natural peroxidases are composed of proteins; thus, one cannot overlook the deficiencies of natural peroxidases such as the low stability under elevated temperatures, narrow optimal pH range, and susceptibility to denaturing. To circumvent these barriers, researchers have developed many hemin-containing peroxidase mimetics with enhanced stability for a broader range of applications 5,6 . These peroxidase mimetics include protein 7 , DNAzymes 8,9 , graphene 10 , graphene quantum dots 11 , metal-organic framework 12 , supramolecular polymers 13 , and inorganic nanoparticles 14 . Undoubtedly, these hemin-based peroxidase mimetics have significantly broadened the application conditions compared to natural peroxidases. Despite these advances, constructing highly efficient and robust peroxidase mimetics with natural enzyme-like flexibility in tuning active sites and microenvironments remains a grand challenge. It is therefore highly desirable to explore alternative materials that can mimic and tune the microenvironment of peroxidases while maintaining high stability under various conditions.
In this context, we developed a class of self-assembled peptoid/ hemin (Pep/hemin) nanomaterials with tunable active sites and microenvironments that mimic peroxidases for lignin depolymerization, by taking advantage of the high tunability of peptoids (or poly-N-substituted glycines) and the uniqueness of their selfassembled crystalline nanomaterials in aligning active sites [15][16][17][18] . Compared to peptides, peptoids can be easily synthesized to achieve a greater side chain diversity while exhibiting much higher chemical and thermal stabilities [15][16][17][18][19][20][21][22] . Our recent work has shown that tuning amphiphilic peptoids can lead to the formation of hierarchically structured crystalline nanomaterials 17,18,[23][24][25][26][27][28] , including membrane-mimetic 2D nanosheets 15,23,29 and nanotubes 16,25,28 . These peptoid-based nanomaterials are highly stable in various pH conditions and at elevated temperatures. New functionalities and applications of peptoid nanomaterials can be easily realized by adjusting the side chain chemistry and by incorporating and aligning different functional groups 16,18,26,28 . Due to their high tunability and stability, peptoid nanomaterials are promising for creating optimal active sites and microenvironments. All these unique properties suggest that peptoidbased crystalline nanomaterials offer great opportunities for developing peroxidase mimetics.
Herein, a series of self-assembling peptoids with [2-(4-imidazolyl)ethylamine]glycine (Nhis), [2-(4-pyridyl)ethylamine]glycine (Npyr), or N-[2-(1H-indol-3-yl)ethyl]glycine (Ntrp) coordination sites were designed and co-assembled with hemin to form crystalline nanomaterials. The catalytic activity of these peroxidase mimetics were evaluated using two representative model compounds: 2,2-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS) and 3,3',5,5'-tetramethylbenzidine (TMB) that are commonly used to evaluate the oxidative catalytic capability of peroxidases. We found that peptoid-based mimetics containing Npyr groups showed the highest catalytic capability. By further varying the side chain chemistry of Npyr-containing peptoids, we demonstrated the tuning of hemin-Npyr microenvironments of these peroxidase mimetics. Our results showed those with N-(2carboxyethyl)glycine (Nce) groups and optimized distance between Npyr and the hydrophobic domain exhibited the best catalytic efficiency with a V max /K m value of 5.81 × 10 −3 s −1 at 0.079 mg/ml catalyst concentration, which is among the highest activities that have been previously reported for hemincontaining peroxidase mimetics [10][11][12]30,31 . By testing the oxidation activity of Pep/hemin at 60 and 90°C, we demonstrated that they remained active and increased reaction rates under elevated temperatures. Three most promising Pep/hemin mimetics were further tested for their high efficacy in the depolymerization of organosolv lignin obtained fromlodgepole pine. Our results showed that these peptoid-based enzyme mimetics can catalyze the oxidative depolymerization of organosolv lignin under much milder conditions with a shorter incubation time in contrast to peroxidase-based enzymic lignin depolymerization 2,32 .

Results
Design of ligand-containing peptoids. To mimic the catalytic site of natural peroxidases and take advantage of the coordination of hemin with self-assembling peptoids, a series of peptoids with various coordination sites (i.e., Npyr, Nhis, or Ntrp) are designed and synthesized: Nbrpm n Nce m -R (n = 6, m = 0-9, R = Npyr, Nhis, or Ntrp; Nbrpm = N-[(4-bromophenyl)methyl]glycine), as shown in Fig. 1a, b. We chose Npyr, Nhis, or Ntrp as terminal ligands because these functional groups are found to facilitate the stabilization of hemin in the active site of natural peroxidases 11 . Because the local environment of hemin active sites within natural peroxidases is important to facilitate lignin depolymerization, and amino acid residues, such as Asp and Phe surrounding hemin, facilitate the catalytic activity via carboxylate groups and hydrophobic interactions 11 , we further designed and synthesized a number of peptoid sequences with variations in coordination sites, the length of polar side-chain domain, and the side chain chemistry.
Assembly and characterizations of Pep-1/hemin nanotubes. To generate crystalline nanotubes with stabilized hemins as active sites, tube-forming peptoid Pep-1 containing Nbrpm 6 , Nce 6 , and Npyr ( Fig. 1a) was synthesized. Pep-1/hemin nanotube was prepared by mixing and incubating Pep-1 with equimolar hemin similar to our previously reported method 16 . Briefly, lyophilized Pep-1 was incubated with hemin in water and acetonitrile (v/v = 50:50) at 4°C for slow crystallization. Gellike materials containing a large amount of crystalline peptoid nanotubes formed after 2 to 3 days of incubation (see Methods for details). Atomic force microscopy (AFM) studies showed that the co-assembled Pep-1/hemin materials exhibited a welldefined tubular structure ( Fig. 1c) with an average tube height of 7.8 nm, which is similar to the data we previously reported for deformed nanotubes under dry condition 16 . Transmission electron microscopy (TEM) also revealed that the diameter of the nanotube is 29.6 ± 3.6 nm, and the wall thickness of the nanotube is 3.8 ± 0.6 nm ( Fig. 1d and Supplementary Fig. 1). Pep-1/hemin nanotubes observed in both AFM and TEM images are similar to our previous results 16 . To further examine the structure of the Pep-1/hemin assembly, we performed X-ray diffraction (XRD) measurement of self-assembled Pep-1/ hemin nanotubes. XRD data showed that Pep-1/hemin nanotubes are highly crystalline and exhibit a similar structure to our previously reported peptoid nanotubes (Fig. 1e). The peak 1.67 nm corresponds to the distance between two peptoid backbones and the peak at 5.7 Å corresponds to the ordered packing of aromatic side chains within the hydrophobic domain. The strong peak at 4.59 Å is from the space of the alignment of peptoid chains between peptoids. The peaks at 4.30 and 3.81 Å can be ascribed to the presence of π-π interaction [33][34][35] .
Having confirmed the assembly morphology and crystallinity of the Pep-1/hemin, we continued to use ultraviolet-visible (UVvis) absorption spectroscopy to determine the optimal assembly ratio between Pep-1 and hemin. An absorption peak at 503 nm, indicating the coordination of Pep-1 and hemin, increases as hemin is titrated into a solution of Pep-1 ( Supplementary  Fig. 2a) 36 . The rise of the peak plateaus at 1.0 mol of hemin for 1.0 mol of Pep-1 ( Supplementary Fig. 2b), suggesting a 1:1 ratio of Pep-1 and hemin for coordination.
Peroxidase-like activity in ABTS oxidation. To test the peroxidase-like activity for potential lignin depolymerization, we used self-assembled Pep-1/hemin for oxidation of ABTS, a commonly used model substrate to reveal peroxidase-like catalytic properties in the presence of H 2 O 2 . As shown in Fig. 2a (black), the concentration of oxidized ABTS increases quickly in the first 90 s upon reacting with self-assembled Pep-1/hemin nanotubes. The initial velocity of Pep-1/hemin nanotubes toward the oxidation of ABTS is 1.862 ± 0.069 μmol min −1 mg −1 (see "Methods" for detail).
To demonstrate the influence of nanotube metal centers on the oxidation of ABTS, we further prepared nanotubes by self-assembly of Pep-1 in the presence of Cu 2+ or Mn 2+ (see "Methods" for details), with the intention of using these metal cations to mimic the hemin active sites of natural peroxidase enzymes 37,38 . The ensuing tubular complexes, namely Pep-1/Cu and Pep-1/Mn, were prepared in a similar way to Pep-1/hemin (see "Methods" for experimental details). AFM images of Pep-1/Cu and Pep-1/Mn showed welldefined nanotube structures with average heights of 8.3 ± 0.4 nm and 7.9 ± 0.6 nm, respectively ( Supplementary Fig. 3). XRD patterns of Pep-1/Cu and Pep-1/Mn also revealed that both assembled nanotubes are highly crystalline ( Supplementary Fig. 4) and exhibit a similar tubular structure to those of Pep-1/hemin and Pep-1 nanotube. The nanotube structures of Pep-1/hemin, Pep-1/Cu, and Pep-1/Mn were further confirmed by scanning electron microscopy (SEM) imaging. Energy-dispersive X-ray spectroscopy (EDS) measurements confirmed the presence of corresponding metal cations bound on nanotubes ( Supplementary Fig. 5), indicating these metal cations are complexed into the nanotubes. As shown in Fig. 2a, under the same reaction conditions, ABTS absorbances with both Pep-1/Cu and Pep-1/Mn nanotubes showed significantly slower increases compared to that with Pep-1/hemin nanotubes. Specifically, the initial velocities of Pep-1/Cu (0.800 ± 0.065 μmol min −1 mg −1 ) and Pep-1/Mn (0.036 ± 0.004 μmol min −1 mg −1 ) complexes (Fig. 2b) were significantly smaller. These results show that the type of metal centers of self-assembled peptoid nanotubes are critical to ABTS oxidation. Among them, those with well-organized hemins showed the best activity. To further confirm the importance of high ordering of hemins in the ABTS oxidation, we prepared hemin-containing nanotubes by simply mixing the pre-formed Pep-1 nanotubes with equimolar hemin, we referred to these hemin-containing tubes as Pep-1 + hemin. The ABTS oxidation results showed that this Pep-1 + hemin sample only led to a poor performance with an initial velocity (V 0 ) of 0.145 ± 0.005 μmol min −1 mg −1 , further confirming the importance of hemin ordering in achieving high peroxidase-like activity.
Tunability of peptoid nanomaterials in ABTS oxidation. To demonstrate the feasibility of peptoid nanotubes as tunable  Backbone length. In the natural peroxidase, the active center of hemin is usually located in a hydrophobic environment, which is critical for achieving high catalytic activity 4 . To demonstrate that the distance between the catalytic sites and the hydrophobic domains is important for peroxidase-like activity, we obtained similar nanotubes (Supplementary Figs. 19-22) using peptoids (Pep-5 to Pep-8) with varied number (n) of Nce groups (from 0 to 9). ABTS oxidation results showed that the initial velocities of these nanotubes reached the maximum when the number of Nce group is 5 (Pep-7) or 6 (Pep-1); nanotubes assembled from either a low n (Pep-5, n = 0; Pep-6, n = 3) or a high n (Pep-8, n = 9) showed a decreased initial velocity. These results suggest that having an optimal distance between the catalytic sites and the hydrophobic domain is important to achieve a high efficiency. We speculate that by changing the distance between the coordinating terminal ligand and hydrophobic domain of peptoids, the active sites composed of hemin and Npyr are located at the optimal distance from the hydrophobic cores of assembled nanotubes, thus offering the optimal hydrophobic environment for interactions with the ABTS substrate.
Side chain chemistry. The active site in the natural peroxidase is often stabilized by various functional groups in the surrounding amino acid residues, such as the -COOH group from Asp 1,41 . To demonstrate the importance of -COOH group in the peptoid/ hemin nanotube to mimic peroxidase, we further prepared similar nanotubes using Pep-9 that has six N-(2-hydroxyethyl) glycine (Noe) groups as the polar side chains instead of Nce groups while keeping the same Npyr coordinating terminal ligand and hydrophobic domain ( Supplementary Fig. 23). ABTS oxidation results showed that these Pep-9/hemin nanotubes exhibited a significant decrease in catalytic activity. Compared with Pep-1, we reasoned that Pep-9 with six Noe groups is unable to facilitate the proton transfer and thus leads to reduced efficiency in oxidation reaction 11,42 .
Assembly morphology. The morphology of self-assembled enzyme mimetic can be significant for influencing its catalytic activity. In fact, various morphologies of self-assembled materials have been designed to mimic the catalytic activity of natural enzymes [43][44][45] .
To demonstrate the importance of tubular structure in the oxidation of ABTS, we synthesized Pep-10 by replacing Nbrpm 6 in the hydrophobic domain with Nbrpe 6 (Nbrpe = N-[2-(4-bromophenyl)ethyl]glycine). Co-assembly of this peptoid with hemin led to the formation of a nanosheet structure (Pep-10/hemin, Supplementary Figs. 24 and 25), which is similar to our previously reported membrane-mimetic nanosheets 24 . ABTS oxidation results showed that Pep-10/hemin nanosheets showed a slight decrease in catalytic activity with an initial velocity of 1.542 ± 0.019 µmol min −1 mg −1 , suggesting the overall morphology of self-assembled crystalline nanomaterials can impact the catalytic activity. Further optimization of nanosheet catalysts is now underway.
High stability of biomimetic catalysts. One of the main motivations for the development of peroxidase mimetics is to enable peroxidase-like reactions to perform under harsh conditions in which natural enzymes cannot survive. Therefore, the activities of these tubular Pep/hemin nanomaterials on ABTS oxidation were determined at elevated temperatures and different solution pHs. Figure 3a showed time-dependent ABTS oxidation under room temperature by Pep-1/hemin, by Pep-1/hemin after 30-day room temperature incubation, or by Pep-1/hemin that had been treated under 90°C for 2 h. It is clear that neither high temperature treatment (i.e., 90°C for 2 h) nor 30-day incubation of Pep-1/ hemin reduces its activity. To further evaluate the operational thermostability of Pep-1/hemin, we next examined Pep-1/hemin activities at 60 and 90°C. In this experiment, the Pep-1/hemin nanotubes were first incubated at 60°C or 90°C for 2 h. Immediately after the incubation, preheated ABTS solutions (to 60°C or 90°C) were added and the initial reaction velocities were determined (Fig. 3b). It is apparent that higher temperature favors the ABTS oxidation rate induced by Pep-1/hemin. Pep-1/hemin showed higher activity at 60°C or 90°C than at 25°C. In contrast, natural lignin peroxidases have poor thermostability and are typically denatured at 60°C 46 . The ability of Pep-1/hemin to remain active and increase reaction rate under elevated temperatures is certainly a significant advantage over natural lignin peroxidases. The high stability of Pep-1/hemin was further confirmed for the ABTS oxidation of Pep-1/hemin under various solution pHs (from 3.0 to 10.0) (Fig. 3c, d). As demonstrated by the initial velocity measurements, the maximum activity of Pep-1/hemin is reached at an acidic pH of 4, which is similar to the optimal pH for natural peroxidases 11 . To further confirm these Pep-1/hemin nanotubes are highly stable, we examined the morphology of Pep-1/hemin after ABTS oxidation reactions under 60 and 90°C. TEM studies showed that self-assembled Pep-1/hemin remained the tubular morphology ( Supplementary Fig. 26). All these findings showed that Pep-1/hemin enzyme mimetics are highly robust and remain active in a variety of conditions, i.e., elevated temperatures and solution pH from 3.0 to 10.0.
Kinetic investigation of the mechanism of Pep-1/hemin reaction with ABTS and H 2 O 2 . To react with the substrate, peroxidases commonly require the presence of H 2 O 2 . In bisubstrate enzyme reactions, there are two main kinetic mechanisms: sequential mechanism (ternary complex formation) and   1)). The reciprocal form of the initial velocity equation is shown in Eq. (2) which can be modified to Eq. (3). θ is the reciprocal coefficient expressed in Dalziel form or Dalziel coefficient 47 , and θ 12 is associated with both substrates. A significant θ 12 value confirms the formation of a ternary complex among the enzyme with both substrates whereas a negligible θ 12 value suggests a double replacement ping-pong mechanism 48 .
To investigate the kinetic mechanism of Pep-1/hemin oxidation of ATBS in the presence of H 2 O 2 , the initial ABTS oxidation rates were determined at four constant H 2 O 2 dosages, 0.05, 0.1, 0.2, and 0.5 mM, each with a variable amount of ABTS concentrations (as arranged in Eq. (3)). As shown in Fig. 4a, the four Lineweaver-Burk plots obtained from each constant H 2 O 2 dosage with various amounts of ABTS appear to be parallel to each other, suggesting that the oxidation of ABTS induced by Pep-1/hemin follows a ping-pong mechanism. The secondary plot, which is obtained using the slope values from the four Lineweaver-Burk plots (θ 1 þ θ 12 ½H 2 O 2 ) against inversed H 2 O 2 changes ( Supplementary Fig. 27), showed a linear fitting with a very small slope (0.00148) correlating to a negligible θ 12 value, verifying the double replacement ping-pong mechanism.
To further evaluate the catalytic performance of the Pep/hemin complexes compared to some existing peroxidase mimetics 5,10-12,30,31 , the enzymic kinetic parameters, including Michaelis-Menten constant (K m ), maximum initial velocity (V max ), and V max /K m ratio, were obtained on another substrate, 3,3',5,5'-tetramethylbenzidine (TMB). Similarly, typical Michaelis-Menten curves were plotted by measuring the initial velocity at different concentrations of TMB (0.4, 0.8, and 1.6 mM) and H 2 O 2 (50, 100, and 200 mM). Both K m and V max values were obtained by fitting the data to the Lineweaver-Burk equation (Fig. 4b, c). For Pep-1/hemin, the K m value of Pep-1/hemin for TMB is 0.264 mM, and the V max value of Pep-1/hemin for TMB is 1.54 × 10 −6 M s −1 . Consequently, the V max /K m value, a common measure to evaluate the catalytic efficiency, is calculated at 5.81 × 10 −3 s −1 , which is one order of magnitude higher than the most active hemin-containing peroxidase mimetics reported so far (Table 1) 5,[10][11][12]30,31 . The results indicate that our best Pep/ hemin catalyst displays a highly efficient performance amongst the reported hemin-containing peroxidase mimetics at yet a similar mass concentration (0.079 mg/ml).
Reaction mechanism. Based on the results, we propose that the oxidation of ABTS substrates in the presence of Pep-1/hemin and H 2 O 2 follows a pathway as shown in Fig. 5. Since the reaction follows a ping-pong pathway, an optimized microenvironment for the active site of Pep/hemin catalyst is critical for ABTS oxidation. In a typical Pep-1/hemin catalyst (Fig. 5a), the hydrophobic domain of peptoids is packed and embedded in the center of the tubular wall, while the hydrophilic domain is extending to the surface of the wall. Conjugating a hemin-binding ligand at the terminus of the hydrophilic domain enables the immobilization of hemins onto the peptoid nanotube. As the coordination between hemin and terminal coordination ligand occurs, self-assembled Pep/hemin nanotubes can have natural peroxidase-like microenvironments. As discussed above, the reaction mechanism follows a ping-pong model, i.e., the two substrates, H 2 O 2 and ABTS, sequentially react with the metal center, where H 2 O 2 oxidizes the Fe III to Fe IV (step-i in Fig. 5b), followed by oxidation of the other substrate ABTS (step-ii in Fig. 5b). Consequently, when the chemical structure of a peptoid is systematically varied, the overall catalytic activity of the Pep/ hemin complex can be tuned. For example, the length of the polar  domain plays a key role in the catalytic activity of peptoid/hemin for ABTS oxidation (Fig. 5c, left panel). We speculate that shorter hydrophilic domain (Nce 3 ) constrains the coordination between the terminal ligand and hemin, thus resulting in weaker coordination between peptoid and hemin. Longer hydrophilic domain (Nce 9 ), on the other hand, may be too flexible for strong coordination to take place. In addition, hemin is distanced from the hydrophobic domain by the long peptoid chain, whereas, in a natural peroxidase, the active site is encased in a hydrophobic environment. Thus, an optimized chain length (Nce 6 ) is needed for improving catalytic activity. Furthermore, we chose to use Nce groups to mimic amino acid residues like Asp in the polypeptide chain surrounding the active site in a natural peroxidase (Fig. 5c, center panel) 49 . As our results suggest, Pep/hemin nanotube with Nce groups in the side chain close to the terminus showed higher activity than that with Noe groups. Finally, in terms of assembly morphology, we speculate that the difference in catalytic activity may be attributed to the unique curvature of nanotube surfaces, in which hemins are more exposed than those on the sheet surfaces for interacting with ABTS substrate or H 2 O 2 molecule (Fig. 5c, 56 . LiP and MnP are biomacromolecules with a molecular weight (MW) of over 38 kDa. Lignin is also a macromolecule that is assembled by multiple phenylpropane subunits. It is conceivable that a single-active-sitecontaining peroxidase has a significant challenge to effectively catalyze the bond cleavage in bulk lignin substrate 2,57 . Due to the unique structural features and high efficacy in ABTS oxidization, Pep/hemin catalysts are promising to overcome the barrier of tuning enzyme-substrate interactions. As illustrated in Fig. 6a, a single Pep/hemin nanotube carries multiple highly ordered active sites which can simultaneously oxidize multiple reactive sites in lignin to promote effective depolymerization. These active sites are exposed on the Pep/hemin surface which minimizes the steric hindrance and electron transfer issues that are inherited to peroxidases 2,58,59 . Therefore, we reason that peptoid/hemin complexes with high efficiency in ATBS oxidation are suitable for efficient depolymerization of lignin polymers. To test the feasibility of Pep/hemin complexes in depolymerization of lignin, three Pep/hemin complexes, Pep-1/hemin, Pep-2/hemin, and Pep-3/hemin, which have shown significant activities in ABTS oxidation (Table 2), were used to depolymerize a biorefinery lignin sample, ethanol organosolv lignin (EOL) obtained from organosolv pretreatment lodgepole pine (Pinus contorta) 60  depolymerized and dissolved in the aqueous solution (See discussions below). The residual EOL solids were collected and analyzed via gel permeation chromatography (GPC) (see "Methods" for experimental details). The depolymerization products in the reaction solution were extracted by ethyl acetate and subjected to gas chromatography-mass spectroscopy (GC-MS) analysis after silylation. As shown in Fig. 6b Fig. 6c, the MW of untreated lignin expands from1 3,000 to~600 Da with two peaks of MW at about 4000 and 1500 Da. H 2 O 2 treatment alone did not change the MW distribution. It is, however, apparent that the MW peak at 4000 disappeared after Pep/hemin treatments. The Pep-1/hemin treatment, which led to the highest lignin mass loss, produced residue lignin with a lower peak MW at~1400 Da confirming Pep-1/hemin indeed depolymerize the lignin. Both Pep-1/hemin and Pep-2/hemin treated lignin also show lower peak MW. The MW around 600 Da presented in both untreated and H 2 O 2 control also disappeared in all Pep/hemin treated samples. A further analysis of the full width at half maximum (FWHM) of GPC peaks was used to evaluate the relative dispersity of depolymerized EOL. As shown in Supplementary Fig. 28, among these three catalysts, Pep-1/hemin showed the lowest FWHM, which is consistent with ABTS oxidation results to confirm that Pep-1/hemin showed the highest catalytic activity. The significant depolymerization of EOL induced by Pep/hemin catalysts was further confirmed by the analysis of GC-MS results of ethyl acetate extractable products. As shown in Fig. 6d, Pep/ hemin catalysts depolymerized EOL into monomeric phenolic compounds, such as a-e. The disappearing of lower MW lignin, i.e. those with MW around 600 Da presented in both untreated and H 2 O 2 treated samples in Fig. 6c, in Pep/hemin treated EOL samples was likely due to their depolymerization into these phenolic compounds. As a comparison, we treated EOL samples with LiP in the presence of H 2 O 2 and performed GPC and GC-MS analyses. As shown in Supplementary Fig. 29a, LiP enzyme was not able to depolymerize EOL. GPC data confirmed that LiP-treated EOL sample showed similar results to those before treatment or only with H 2 O 2 . The GC-MS results ( Supplementary Fig. 29b) showed that very little phenolic compounds were observed when EOL was treated with LiP in the presence of H 2 O 2 . These GPC and GC-MS results confirmed that LiP enzyme is not able to depolymerize lignin as Pep/hemin catalysts do.
To further confirm the depolymerization of EOL into phenolic compounds 62,63 , the yield of phenolic compounds following the depolymerization was quantified via microtiter-plated Folin-Ciocalteu (F-C) assay based on previously reported methods [64][65][66] (see "Methods" for experimental details). The yields of phenolic compounds following the depolymerization of EOL by Pep-1/ hemin, Pep-2/hemin, and Pep-3/hemin in the presence of H 2 O 2 were calculated at 61.7%, 60.4% and 47.2%, respectively (Supplementary Table 1). The ability of Pep/hemin catalysts to depolymerize nearly two-thirds of a realistic lignin sample is extraordinary. However, for the sample treated with only H 2 O 2 , the yield was calculated at only 10.2% (Supplementary Table 1 (Fig. 6c, d).
As far as we know, this is the only example showing the use of highly stable and tunable biomimetic enzymes for direct lignin depolymerization 73,74 . Although enzymatic methods have been used for depolymerizing lignin via white-rot fungi or bacteria 75,76 , this process often takes a long time (weeks) or is of low yield (from 7% to~30%) 77,78 . On the other hand, chemical catalysis methods such as transition metals and acid/base are carried out under high pressure or at high temperatures 79 . Our results reveal that peptoid-based nanomaterial can depolymerize lignin to generate low-molecular-weight products at mild conditions and holds a promising potential for depolymerization of lignin to generate high-value products such as jet fuels. Figure 6e illustrates the possible mechanism of Pep/hemin that are contributing to its superior ability to depolymerize lignin polymer. First of all, the density of active sites on peptoid nanomaterial is higher than that on an enzyme due to the significant difference in MW between peptoid (~2 kDa) and ligninolytic enzyme such as lignin peroxidase (>38 kDa) 80 . A 10nm-long Pep/hemin nanotube can carry~910 reactive sites 16 (one per 2 kDa) while a peroxidase molecule with a size of 50 × 50 × 40 Å typically has one active site (0.05 per 2 kDa) 81 . In addition to higher density, the active sites of Pep/hemin are exposed on the nanomaterial surface, making them much more accessible compared to those of peroxidase enzymes which are encased inside the enzyme protein structures (Fig. 6a). The Pep/ hemin complexes are likely smaller than a lignin polymer in size, thus they can easily attack multiple phenylpropane linkages. A possible interaction between the hemin active site and lignin subunit and subsequent cleavage of αC-βC linkage is shown in Fig. 6e. The ability of Pep/hemin complexes to conduct multiple bond cleavage in lignin is probably one major reason leading to the efficient lignin depolymerization. Finally, the electron transfer in the catalytic cycle can be hindered by the long-range electron transfer between lignin peroxidase and hemin active site 82 . In summary, our results demonstrated that Pep/hemin crystalline nanomaterials can effectively catalyze the oxidative depolymerization of real lignin and address some deficiencies of ligninolytic enzymes. Encouraged by these results and based on the tunability of peptoid nanomaterials, we expect to further enhance the catalytic activity of these peptoid-based enzyme mimetics by incorporating multiple coordinating ligands, such as Npyr and Nhis, in a single peptoid chain to further mimic the microenvironments of natural peroxidase enzymes.

Discussion
By varying the structural features of peptoids including terminal ligands, side chain chemistries, and self-assembly behavior, we successfully develop a class of peptoid-based crystalline nanomaterials as peroxidase mimetics with tunable active sites and microenvironments. These biomimetic catalysts are highly stable and exhibit high efficiency in ABTS oxidation and lignin depolymerization. Catalytic oxidation results showed that the chain length of self-assembling peptoids played a key role in achieving high efficiency in the peroxidase-like activity, that is, the tubular Pep/hemin complexes assembled from peptoids with six Nce groups show the highest efficiency. Deviation from the optimal number of Nce groups reduces the catalytic activity of the formed nanotubes. We further demonstrated that self-assembled nanotubes containing Npyr terminal ligands exhibited the highest efficiency in both ABTS oxidation and lignin depolymerization while changing Npyr to Nhis or Ntrp led to reduced activity. Our results also revealed that these peptoid-based peroxidase mimetics are highly stable and remain active at elevated temperatures and various solution pHs (from 3.0 to 10.0). Kinetic studies showed that these enzyme mimetics exhibited a ping-pong (double displacement) mechanism in catalytic oxidation. Because peptoidbased crystalline nanomaterials are highly tailorable and stable, we expect the self-assembly of peptoids into hierarchically structured crystalline nanomaterials with the ordered alignment and organization catalytic sites will provide a fascinating strategy for the design and synthesis of robust enzyme mimetics for various applications including lignin depolymerization.

Methods
Materials. Hemin and dimethyl sulfoxide were purchased from Sigma Chemical Co. (St. Louis, MO). β-Alanine t-butyl ester hydrochloride was purchased from Chem-Impex International, Inc. This hydrochloride salt was deprotected by addition of sodium hydroxide aqueous solution, then extracted with CH 2 Cl 2 , filtered, and rotary evaporated for further use. N,Nʹ-diisopropylcarbodiimide, bromoacetic acid, and trifluoroacetic acid (TFA) were purchased from Chem-Impex International, Inc and used as received. 4-Bromobenzylamine was purchased from Oakwood Products, Inc. All other amine submonomers and other reagents are obtained from commercial sources and used without further purification. MilliQ water at 18.2 MΩ cm was used for all experiments.
Peptoids sequence synthesis and self-assembly. All peptoids were synthesized using a modified solid-phase sub-monomer synthesis method as described previously 16 . Peptoid syntheses were performed either on a Aapptec Apex 396 robotic synthesizer or manually in 6-ml plastic vials. After solid-phase synthesis, all peptoids were cleaved from resins by mixing with 95% TFA in water for over 30 min. After evaporating TFA, the obtained peptoid crudes were dissolved in water and acetonitrile (1:1, v/v) for HPLC purification. All peptoids were purified using reverse-phase HPLC on an XBridge™ Prep C18 OBD™ (10 μm, 19 mm × 100 mm), having a narrow gradient of acetonitrile in H 2 O with 0.1% TFA over 20 min with a 20 ml min −1 flow rate. Waters ACQUITY reverse-phase UPLC (using a corresponding gradient at 0.4 ml min −1 over 7 min with an ACQUITY-BEH C18, 1.7 μm, 2.1 mm × 50 mm column) connected with a Waters SQD2 mass spectrometry system was used to analyze the purified peptoids. The final peptoid product was lyophilized at least twice from the mixture of H 2 O/CH 3 CN solution.
The obtained lyophilized sample was finally divided into small portions (1.0 or 2.0 μmol) and stored at −80°C.
Lyophilized and HPLC-grade peptoids were dissolved in the mixture of H 2 O/ CH 3 CN (1:1, v/v) to make a 5.0 mM clear solution, which was then transferred to a 4°C refrigerator for slow evaporation. Suspensions or gel-like materials containing crystalline nanotubes or nanosheets were formed after 3 days. For hemin-peptoid co-assembly, lyophilized peptoid and equimolar hemin were dissolved in H 2 O/ CH 3 CN (1:1, v/v), and the mixture was left at 4°C following the same protocol described above for slow evaporation to obtain brown and gel-like materials. For co-assembly of peptoids and metal ions, the equimolar CuSO 4 or MnCl 2 aqueous solution was mixed with peptoid and left at 4°C for slow evaporation to obtain gellike materials.
Preparation of Pep-1 + hemin. Pre-formed Pep-1 nanotubes were incubated with equimolar hemins in a 500 µl aqueous solution. After about 6-h incubation, this solution was centrifuged and washed well with water (500 µl) three times. The obtained hemin-containing nanotubes (namely Pep-1 + hemin) were then used for catalysis measurements.
Pep/hemin catalyst stability test. The corresponding gel-like Pep/hemin materials (1 µmol) were diluted with 500 µl H 2 O to yield a 0.2 mM solution. Then the obtained aqueous solution was incubated at 90°C for 2 h, or at room temperature for 30 days, or heated at 60 or 90°C for overnight before being used for catalytic studies or TEM imaging.
Transmission electron microscopy (TEM). For TEM characterizations, TEM samples were made by having a 2.0 μl drop of the self-assembled sample diluted with 5.0 μl of deionized water. Once this solution was deposited onto carbon-coated copper grids for 10 min, it was dried with filter paper. Afterward, 5.0 μl of 2% phosphotungstic acid (PTA) in distilled water (pH 7.0) was dropped onto the TEM grid for another 2 min for negative staining. After removing the extra PTA solution using filter paper, the TEM grid was dried and ready for TEM imaging. All TEM data were collected on a 200-kV FEI Tecnai TEM microscope.
Atomic force microscopy (AFM). AFM images were collected on a Bruker MultiMode 8 by using either a ScanAsyst mode or a tapping mode. For AFM sample preparation, a 1.0 μl drop of the self-assembled peptoid solution was diluted with 40 μl H 2 O. Then it was deposited onto a freshly cleaved mica surface and incubated for 10 min. After removing the solution using filter paper, the mica sample was dried using a stream of N 2 gas and ready for AFM imaging.
Scanning electron microscopy (SEM). Water-washed and diluted peptoid nanotubes or nanosheets were drop-cast onto silicon substrates to obtain SEM samples. The morphology of all samples was examined by using an FEI Helios NanoLab 600i SEM. The chemical composition of all samples was analyzed by using EDS with an Oxford X-Max 80 EDS detector (Oxford Instruments Analytical Ltd).
X-ray powder diffraction (XRD). All XRD data were collected on a multiple wavelength anomalous diffraction and monochromatic macromolecular crystallography beamline, 8.3.1 (possessing a 5 T single pole superbend source with an energy range of 5-17 keV) at the Advanced Light Source located at Lawrence Berkeley National Laboratory. All XRD data were collected with a 3 × 3 CCD array (ADSC Q315r) detector at a wavelength of 1.1159 Å. For XRD sample preparations, samples of peptoid assemblies were loaded onto a Kapton mesh (MiTeGen) and dried for XRD measurements. All XRD Data were analyzed using custom Python scripts.
UV-Visible (UV-Vis) Spectroscopy. UV-Vis spectroscopy was either performed on a TECAN Safire2 spectrometer using a 96-well plate or an Ultrospec 2100 pro UV/Visible spectrophotometer. For TECAN spectrometer with a plate reader, to each well of the plate, 100 μl of the solution was pipetted, and a full spectrum from 200 nm to 800 nm was recorded with triplicate samples. For Ultrospec 2100 spectrometer, a quartz cuvette with a path length of 1.0 cm was used for all measurements.
Specific activity (initial velocity screening). The specific activity (µmol mg −1 min −1 ) was calculated according to the equation: where a catalyst is the activity expressed in units per milligram (U mg −1 ), and [m] is the weight of catalyst used in each assay (mg). The b catalyst is determined as follows: where b catalyst is the catalytic activity of peptoids expressed in unites; A is the absorbance after subtraction of the blank value; ΔA/Δt is the initial rate of change in the absorbance at 420 nm for ABTS and 652 nm for TMB expressed in min −1 ; ε is the molar absorption coefficient of the substrate (36,000 M −1 cm −1 at 420 nm for ABTS, and 39,000 M −1 cm −1 at 652 nm for TMB). In a typical experiment, the gel-like peptoid/hemin material (1 µmol) was diluted to 500 µl using H 2 O to yield a 0.2 mM solution. Then, 1000 µl 0.2 mM ABTS solution is mixed with 10 µl 30% H 2 O 2 in a quartz cell with a path length of 1 cm. A single shot of 100 µl 0.2 mM peptoid/hemin catalyst was injected into the ABTS/H 2 O 2 substrate mixture to initiate the reaction, at which time the absorbance was recorded every 5 s at 420 nm and 652 nm for ABTS and TMB, respectively. To determine the reaction rate of Pep/hemin under different temperatures (i.e. 60, and 90°C), 0.2 mM Pep-1/hemin catalysts were incubated at 60°C or 90°C for 2 h. Immediately after the incubation, Pep-1/hemin samples were added to preheated ABTS solutions (60°C and 90°C respectively) to determine the initial reaction velocity.
Lignin depolymerization and characterizations. An organosolv lignin (lodgepole pine) was collected during pretreatment of lodgepole pine by 65% ethanol and 1.1% sulfuric acid (on wood) at 170°C for 60 min 61 . The purified and freeze-dried EOL sample was provided by Prof. Xuejun Pan from University of Wisconsin-Madison and used for depolymerization experiments in this study. Lignin peroxidase (>0.1 U mg −1 ) was purchased from Creative Biomart Inc (Shirley, NY, United States) and stored at −20°C until use. In a typical lignin depolymerization experiment, 50 mg of the EOL was mixed with 1000 µl 0.1 mM Pep/hemin or 2 mg ml −1 LiP solution in a glass vial and incubated at 200 rpm for 30 min at 60°C. Thereafter, 50 µl of 30% H 2 O 2 was added to the reaction mixture and the vial was further incubated at 200 rpm for 30 min, 1 h, or 2 h at 60°C.
F-C experiments were performed to determine the yield of phenolic products from lignin depolymerization following established methods [64][65][66] . In a typical F-C experiment, residual lignin sample was washed with water for two times and the aqueous phase was combined yielding a total volume of 10 ml. Afterward, 20 μl of the aqueous phase was added to 100 μl F-C reagent, 1.5 ml water and 300 μl Na 2 CO 3 (20%). The mixture is incubated at 45°C for 1 h. After this, 100 μl of the mixture was pipetted to a 96-well plate and the absorbance at 765 nm was measured against a blank water sample. The amounts of phenolic compounds were quantified by correlating the sample absorbance with phenolic compounds concentration using a standard curve generated from isoeugenol (a standard compound) at different concentrations.
Gas chromatography-mass spectroscopy (GC-MS). GC-MS was performed on an Agilent 7890A/5975C spectrometer using a DB-5 capillary column and it was used to analyze silylated ethyl acetate fraction (see Lignin depolymerization and characterizations section above). GC oven temperature setting was as follows: 100°C for 3 min, 5°C min −1 ramp to 200°C, 200°C for 3 min, 20°C min −1 ramp to 320°C, and 320°C for 10 min. Solvent delay was set to 3 min. Positive EI mode was selected for mass spectrometer detector with temperature setting as follows: MS Source Setpoint was set at 230°C and MS Quad Setpoint was set at 150°C. Scanning ion range was set from m/z 50-1000. Compounds between 5 and 30 min were identified and chromatograms were plotted.
Gel permeation chromatography (GPC). The analysis was performed according to an established method 83 using a Perkin Elmer HPLC system with a PDA LC detector. Three Agilent columns (60, 300, and 1000 Å porosity; 5 μm bead size) were used to perform the size exclusive separation at room temperature. The mobile phase was tetrahydrofuran (THF), applied at a flow rate of 0.6 ml min −1 . Samples were prepared at a concentration of 30 mg ml −1 in THF. A 100 µl injection volume was used. The data was acquired and analyzed using the TotalChrom Navigator software package.
All the lignin samples were acetylated prior to the analysis. To prepare an acetylated lignin sample, the residual lignin sample was mixed with 1 ml of pyridine and 1 ml of acetic anhydride and stirred for 24 h. After acetylation, 5 ml MeOH was added to the mixture, and the solvent was removed under reduced pressure. The remaining solid was dissolved in 1 ml THF and was used for GPC analysis. Standard samples of polystyrene beads (Fischer Scientific) with known M w of 1300, 3350, 5200, 13,000, and 50,000 and veratrole with M w of 138 were also dissolved in THF at 30 mg ml −1 concentration for GPC analysis.