Introduction

To determine whether our engineered TfCut2 and its variants can effectively degrade PET–cotton blended textiles, we employed a primary enzymatic assay, pNPB, along with several multi-step degradation experiments.

Catalytic activity was first confirmed using the pNPB assay, which showed that the enzymes did indeed hydrolyze ester bonds under various pH and temperature conditions.

Afterwards, we conducted a series of rigorous experiments to test for degradation of PET, by measuring soluble degradation products through UV absorbance at 260 nm. High-throughput PET film degradation was performed to determine the optimal experimental conditions. We then formally degraded PET film at a larger scale for direct evidence of PET breakdown.

Pretreatments were employed to reduce the crystallinity of the textile samples before degradation. The effect of the pretreatment is verified with FT-IR and DSC.

We then test our TfCut2 variant’s ability to degrade on pure PET textile. Similar to PET film, UV absorbance was also used to assess degradation of pure PET textiles. Moreover, techniques like SEM and HPLC were used to confirm degradation at the microscopic and molecular levels.

Lastly, we tested degradation on cotton-PET blended textiles. Prior to degradation we applied commercial cellulase to degrade cotton fibers, in order to improve degradation efficiency. DSC was then used to determine cotton fibers degradation. Afterwards, we evaluate degradation with the same methods as used in pure PET textile.

pNPB Assays

Before applying our TfCut2 variants in PET degradation, we first aimed to verify their general hydrolytic activity and determine optimal working conditions using the model substrate pNPB. The experiments below trace this validation process from baseline wild-type characterization to comparative variant testing across buffer systems.

To confirm our TfCut2- 5ZOA (Wild-type) and its six variants (Variants 1 to 6) are functional after protein production and purification, enzyme activity was measured using the pNPB assay, which is a widely used method to quantify enzymatic hydrolytic activity. When the enzyme hydrolyzes pNPB to p-nitrophenol, we can measure absorbance at 405 nm using a microplate reader. In the beginning, we used Tris-HCl as a reaction buffer based on literature review (Chen et al., 2010; Dong et al., 2020). Further research directs us into using different buffers since we found that Tris-HCl buffer, although utilized in most studies, is not the most ideal buffer due to its inhibitory effects on TfCut2 (Schmidt et al., 2016). In the end, there are three types of buffers that we utilized: Tris-HCl, Calcium-induced Tris-HCl, and Calcium-induced HEPES (Then et al., 2015). The experimental conditions, including pH and temperature, also followed the papers used before.

Preliminary Tests

To assess the optimal reaction conditions for each enzyme, we first tested TfCut2 wild-type (PDB ID: 5ZOA) at pH 6–9 and temperatures ranging from 55°C to 70°C. This was to ensure that wild-type enzyme (PDB ID: 5ZOA) was functional under these conditions, allowing us to proceed with testing the six variants. The results showed that enzyme activity had an increasing trend at pH 6–8 under temperatures of 55°C to 65°C (Figure 1). We observed decreasing OD405 values at pH 9, which is consistent with the description in the paper that the substrate undergoes spontaneous hydrolysis at elevated temperatures and pH values of 9 or above (Furukawa et al., 2019). Overall, our enzyme was most active around pH 7–8, with an over 90% decrease at pH 9. Temperature also had a great impact, since we observed that enzymes were most active under 55°C and least active under 70°C. Hence, subsequent experiments will be conducted at lower temperature to achieve better enzymatic activity.

In addition to using 2D plots for visualizing temperature and pH effects, we employed Response Surface Methodology (RSM) graphs to determine the most optimal pH and temperature conditions. RSM is a collection of statistical and mathematical techniques to model and analyze how a response is influenced by several quantitative input variables. The primary objective of RSM is to optimize this response by identifying the optimal levels of the input factors (Mushtaq et al., 2015). 3D surface plots and 2D contour plots are commonly used to visualize the relationships and interactions between the independent variables and the response variable. The RSM predictions were based on experimental data collected on June 4 and 6, with five replicates at each condition (Figure 2).

The optimal condition was identified at pH 7.5 and temperature around 50°C. The temperature obtained from RSM would be referenced in further experiments. Hence, the conditions that will be tested in later experiments would still be focusing on pH 7-8 and temperatures below 55°C.

pNPB assay with Tris-HCl buffer

We conducted the assays using 200mM Tris-HCl buffer and pNPB (2mM final concentration) with the addition of TfCut2 (10 mM final concentration). The conditions we obtained from the preliminary tests were used in our engineered enzyme. Based on our RSM graph, the peak appeared to be around pH 8. To obtain more precise results, we narrowed down the pH range to 7, 7.5, 8, and 8.5 and broadened the temperature range from 35°C to 70°C in 5°C increments. We observed that only the relative activity of Variant 1 (G62A/P193H/S197D), Variant 2 (H77R/D204C/E253C) and Variant 3 (A65H/L90A/I213S) were higher than TfCut2-5ZOA (wild-type) at pH 7 across the temperature range of 35°C to 60°C (Figure 3A). Evidently, the optimal working conditions of these three enzymes were at pH 7 and at a temperature between 55°C to 60°C. Variant 4 (D12S) and Variant 5 (T234L) expressed a similar trend and showed higher hydrolyzing activity—up to three times greater than the wild type—at a pH of 7 and under 65°C (Figure 3B). However, Variant 6 (Full MutCompute) showed little to no enzyme activity under each condition.This finding aligns with the drylab design, as Variant 6 (Full MutCompute) was primarily a machine-learning-guided enzyme in which all original mutation sites were replaced with machine-predicted ones. As a result, its structure differs drastically from the other Variants, making it a far-riskier construct. For more detailed information, explore the drylab page.

A

B

We also used RSM to graph out the results of our Tris-HCl buffer pNPB assays (Figure 4). On the graphs we mapped out, pH and temperature values are used on the x and y-axes, respectively. The z-axis shows the OD405 readings, which represents the activities of our enzymes. We also marked out the optimal point, which was predicted based on the surface.

With the exception of V6 (Figure 4.G), RSM graphs of the variants are identical. In terms of the pH axis, we see a parabola-shaped trend that peaks at pH range 7.5~8. The trend observed among temperature is also presented here, as we see an increased OD reading with lower temperatures. The optimal conditions (marked with a red star) in every graph are located at a temperature range of 35°C-40°C and pH 7.5-8.

The Tris-HCl results informed us that Variant 1-3 are outperforming wild type under moderate temperature and slightly alkaline pH. These patterns indicate that our enzyme exhibits stability under conditions typical for polyester degradation. However, the performance of Variant 6 identified that machine-predicted mutants may compromise efficiency.

pNPB assay with HEPES buffer

In a study regarding effects of various buffers on polyester hydrolase activity, Schmidt et al. (2016) concluded that Tris competitively inhibits aminopeptidases, cholinesterases, and hydrolases such as TfCut2 when used as a buffer. The activity of TfCut2 decreased when placed in the Tris-HCl buffer, especially when the molarity of Tris was increased, meaning that Tris-HCl inhibits TfCut2 activity. Hence, we changed our experimental approach from using Tris-HCl buffer for pNPB assays to using HEPES buffer to prevent inaccurate results and to improve the activity of TfCut2. Moreover, Then et al. (2015) revealed that TfCut2 is not sufficiently thermostable at high temperatures required for efficient PET degradation. According to literature, the addition Ca²⁺ cations to the enzymes and enhances thermostability, thereby increasing the melting point of the enzymes up to 14.1 °C from the usual 71.2 °C (Then et al, 2015). Therefore, we added CaCl₂ to our HEPES buffer to improve enzyme activity and more efficient PET degradation at elevated temperatures. The pNPB assay was conducted using the same protocol as previously described but with our new buffer and calcium ion addition. When the buffer system changed to HEPES, we observed a change in our results.

A

B

Under most conditions, we see Variants 1-3 (G62A/P193H/S197D, H77R/D204C/E253C, and A65H/L90A/I213S respectively) performing almost as well as, if not better than, the 5ZOA wild type. A trend was observed among pH levels, with most variants exhibiting the highest level of activity relative to the wild type at pH 7.5. However, at elevated temperatures from around 60°C to 70°C, Variant 3 performed substantially better than other variants and relative to the wild type, especially at a pH of 7, suggesting that its efficacy will be the highest in degrading PET. The effect of temperature was not as pronounced across all data values, suggesting the addition of calcium ions may have allowed the enzyme activity to resume at a similar rate to lower temperatures even under elevated temperatures.

Across all tested conditions, Variants 4 (A65H/L90A/I213S), 5 (D12S), and 6 (Full MutCompute Variant) had lower levels of activity than the wild type, with Variant 5 performing better than Variants 4 and 6 at all data points. Notably, Variant 6 exhibited generally little to no activity relative to the wild type at all tested conditions, similar to its performance under Tris-HCl buffer, suggesting that the Variant 6 is not suitable for our following degradation process.

Similar to the RSM of the Tris-HCl buffer, the RSM of the HEPES buffer also reflected trends among pH and temperature. With all variants except Variant 6 (Figure 7.G), temperature and activity are negatively correlated: an increase of temperature will decrease the activity. Conversely to the Tris-HCl buffer trend, the Figure 7 trend among pH levels seems to suggest a peak at pH level 7-7.5, which is a bit lower than the pH 7.5-8 of the Tris buffer.

Conclusion

In conclusion, we conducted pNPB assays to assess the activity of our engineered enzymes. Our results show that the majority of engineered variants, excluding Variant 6 (Full MutCompute), demonstrate comparable activity to wild type 5ZOA, indicating successful retention of enzyme function in ester bond cleavage. While these conditions cannot be directly applied to PET degradation, the data obtained and conditions tested serve as a reference for subsequent PET degradation experiments.

PET Film Degradation

Before using TfCut2 and its mutant to degrade textile, we must ensure they can digest PET film whose crystallinity is lower than the PET in textile and more easily to be degraded in theory. Thus, we used PET film to optimize the enzyme degrade condition and ensure its hydrolysis activity toward PET.

The results of our project are divided into four phases: preliminary PET film degradation, high throughput PET film degradation, expanding experimental conditions, and formal PET film degradation. During preliminary PET film degradation, because we never digested PET in our laboratory before, we used the condition from published research to test our enzyme and the method to evaluate enzyme activity. In this phase, we establish UV absorbance, which represents the PET degrading product terephthalic acid at 260 nm as a method by testing PET film degradation at pH 8 and 60 °C for 48 h. Secondly, during our high throughput PET film degradation, used reduced volumes and standardized PET film sizes (0.8 × 0.25 cm) across 96 samples, testing in 100 mM HEPES buffer under conditions matched to the pNPB assay, measuring UV absorbance across pH 7.0–8.5 and 55–70 °C, with no-enzyme blanks included as negative controls. Third, shifting to 500 mM HEPES to replicate literature conditions, expanding temperature testing to 35–70 °C, and visualizing results through surface plots while maintaining the same control setup. Fourth, formal PET film degradation. We increased the reaction scale and confirmed enzymatic PET degradation directly by UV absorbance and HPLC and calculated the PET film weight loss after 48 h at 51.8 °C.

Across these phases, Variant 3 (A65H, L90A, I213S) and 4 (D12S) consistently showed significantly higher activity, wild-type TfCut2 5ZOA performed moderately, and the remaining Variants either showed weak or negligible activity. Activity was strongest at 51.8 °C, while the surface plots provided a clear visualization of how pH and temperature influenced enzyme performance. Direct degradation assays confirmed that UV absorbance measurements reflected true PET hydrolysis.

Testing the Feasibility of PET Degradation

Unlike pNPB assay, the products of PET degradation such as terephthalic acid (TPA), MHET, BHET, and ethylene glycol (EG) are not able to be detected by the visible spectrophotometer. Although high performance liquid chromatography (HPLC) can precisely quantify the concentration of PET degradation products, it is impractical to use HPLC to optimalize the pH value and temperature because of its slow analysis time, high operating and maintenance cost. However, our ELISA reader can detect UV absorbance, and since PET degradation products can absorb UV on 254 nm, we can use UV absorbance to do the high throughput screening for the optimal reaction conditions for TfCut2 to degrade PET.

To validate our concept, we first performed PET degradation using the previously published reaction conditions (Wei et al., 2020) for TfCut2 and measured whether there was a significant change in the UV absorbance. We degraded one 0.25 x 0.5 cm PET film in 100 μL 200 mM Tris-HCl (pH 8) with 5 μg/mL enzyme at 60℃ 800 rpm for 48 hours. The results showed that UV absorbance at 260 nm can measure enzyme activity toward PET (Figure 1). Although UV absorbance at 240 nm can also be used for detecting PET degraded products, our UV microplate has a strong background below 250 nm. Thus, we used UV absorbance at 260 nm to do the further experiments.

High-Throughput Screening for Optimalizing PET Film Degradation

In order to optimize enzyme reaction conditions, we scaled down the enzymatic reaction and measured UV absorbance for simultaneous high-throughput screening to obtain the best reaction parameters.

In the first part of the experiment, we tested PET hydrolysis in 100 mM HEPES with 10 mM CaCl₂ to align with the conditions used in our pNPB assay. The addition of Ca²⁺ was informed by the previous study (Then et al., 2015), who showed that Ca²⁺/Mg²⁺ binding enhances polyester hydrolase stability and activity. Activity was measured across pH 7.0–8.5 and 55–70 °C. Results again showed Variant 3 as the most active enzyme, followed by the wild type, while other variants remained weak.

In the previous study, Then et al. used 500 mM HEPES as a reaction buffer for TfCut2 to degrade PET (Then et al., 2016). In order to obtain a better result, we changed the reaction buffer to 500 mM HEPES with 10 mM CaCl₂. PET films were incubated for 48 h across multiple temperatures (35°C–70°C) and pH values (7.0–8.5). All TfCut2 and its variants except Variant 6 (Full MutCompute) demonstrate activity in degrading PET films. Variant 3 (A65H/L90A/I213S) exhibits the highest activity at 50°C pH 8, followed by Variant 4 at the same temperature but pH 7.5. Variant 5 (T234L) also exhibits elevated activity compared to the wild type, while Variant 2 (H77R/D204C/E253C) is relatively close to the wild type.

Response Surface Methodology on High-Throughput Screening PET Film Degradation

In the previous section, we did the high-throughput screening and obtained the surface graph for TfCut2 wild-type (5ZOA) and its mutants. However, the optimal reaction conditions may lie between the conditions we tested. To obtain precise optimal reaction parameters, we used Response Surface Methodology to determine the best reaction conditions for TfCut2, along with its predicted UV absorbance (Table 1), and the response surface graphs for each enzyme (Figure 4).

For more information on RSM please check our drylab RSM page

In order to obtain the best reaction conditions for enzymes, we adopted three strategies using response surface methodology: considering all TfCut2 and its mutants regardless of their activity, excluding the mutant without activity, and only considering the mutants whose activity is better than wild-type. Next, we merged the datasets of the individual enzymes into three groups according to three strategies. Then determined their respective response surfaces and optimal reaction pH and temperature (Table 2). Due to the limitation of our instrument, we decided to use 51.8℃ pH 7.5 as the optimal reaction condition for the further experiments.

PET-degrading activity of six variants

After confirming assay conditions across buffers, pH values, and temperatures, we selected the most promising conditions for direct product quantification. In this section, we scaled up the reaction: used 8 pieces of 3 x 0.5 cm PET film (surface area around 12 cm2) in a 1.8 mL 500 mM HEPES with 10 mM CaCl₂ (pH 7.5) reaction buffer with 27.8 μg/mL enzyme. After incubating TfCut2 and its mutants in a thermoshaker at 51.8°C, 800 rpm for 48 hours, we utilized the ELISA reader to measure UV absorbance at 260 nm. The results were consistent with the absorbance data: Variant 3 (A65H/L90A/I213S) released the most TPA and produced the greatest weight loss, followed by Variant 4 (D12S)and the TfCut2-5ZOA (wild type) (Figure 5A).

To make the data results more intuitive, we roughly converted UV absorbance data into a mass of TPA produced during PET degradation. We established a standard curve of TPA (Figure 5B) and converted the UV absorbance at 260 nm to the concentration of TPA: Then we multiply the concentration of TPA by total reaction volume 1.8 mL to calculate the total mass of TPA in reaction buffer: However, the total mass of TPA is not equivalent to the PET film weight loss. One PET molecule can be degraded to one TPA molecule and one ethylene glycol (EG) molecule. To calculate the PET weight loss, total mass of TPA must multiply the ratio of PET monomer molecular weight (Mw of PET monomer) and TPA molecular weight (Mw of TPA). Through the calculations above, we were able to determine the PET weight loss (Figure 5C). However, this approximation is a rough estimate. PET degradation actually produces intermediate compounds MHET and BHET, which will absorb UV absorbance at 260 nm. This rough calculation can only combine all of these intermediate compounds as the final TPA.

As mentioned above, because UV absorbance counts all chemicals which can absorb UV light, it overestimates the concentration of TPA around 15 to 40% (Zhong-Johnson et al., 2021). To obtain more accurate data, we choose one set of samples from three PET film degradation and utilize high performance liquid chromatography (HPLC) to analyze the PET degraded product in the reaction buffer.

In order to quantify the intermediate product of PET degradation, we made a 4-((2-hydroxyethoxy)carbonyl)benzoic acid (MHET) standard curve and measured the concentration of MHET in the reaction buffer. Then we used the following equation to calculate these MHET weights equal to how much PET degraded. Besides UV absorbance and HPLC, we did gravimetric analysis by measuring the weight of PET films before and after TfCut2 digestion. After PET film degradation, we washed PET film by 1% SDS for 30 minutes and anhydrous ethanol for 5 minutes at 50℃. And rinse with deionized water and dry in the 50℃ oven for at least 48 hours. Then we measured the weight of PET films and the results show in table 3.

Each of the three methods for estimating PET weight loss presents distinct limitations. As reported by the previous study (Zhong-Johnson et al., 2021), the PET weight loss calculated from UV absorbance was greater than that derived from HPLC, in some instances even surpassing the mass determined by gravimetric analysis. In contrast, for the high A260 groups, the HPLC estimated mass exceeded the mass estimated from UV absorbance. This discrepancy can be attributed to the A260 values of these groups exceeding the upper detection limit of the ELISA reader, thereby precluding an accurate quantification of PET degradation products. With HPLC, the absence of a standard for the intermediate product, Bis(2-hydroxyethyl) terephthalate (BHET), prevents its quantification, potentially leading to an underestimation of the extent of PET degradation. Meanwhile, in gravimetric analysis, measuring the change in PET mass poses a significant challenge. Inadequate washing of the PET film can leave residual buffers or enzymes, which introduces artifacts into the measurement. This often results in a measured mass greater than that prior to degradation. Therefore, obtaining a BHET standard for HPLC-based quantification would likely yield a more precise estimation of PET weight loss.

To prove the TfCut2 and its mutants can degrade PET films, we used scanning electron microscopy (SEM) to observe the PET film surfaces (Figure 6). SEM images of the PET film surface also corroborate the enzyme degradation of PET. In the SEM images, mutants except Variant 6 can penetrate the PET film surface and create pore structures during degradation (Figure 6E-N).

In conclusion, the collective data from UV absorbance, HPLC, and gravimetric analysis demonstrate that TfCut2-Variant 3 (A65H/L90A/I213S) triple mutant and TfCut2-Variant 4 (D12S) single mutant exhibit enhanced PET degradation capabilities relative to the wild type TfCut2-5ZOA.

Pretreatment

PET textile often possesses exceedingly high tensile strength and thermal stability—both attributed to their higher-than-normal crystallinity, where the polymers in the crystalline regions are then densely packed, more rigid, less permeable, and melt at higher temperatures, thus resisting the effectiveness of enzymatic attack. These crystalline regions limit the available polymer surface for enzymes to penetrate the dense polymer matrix (Dewi et al., 2025). Therefore, to overcome these limitations, we incorporated a thermal-alkaline pretreatment method to decrease the crystallinity of PET and enhance the subsequent enzyme degradation of PET textiles. We applied the pretreatment method described by previous research (Boondaeng et al., 2023), which involves applying 15% NaOH to textiles at high temperatures. Mechanistically, the high temperature damages the PET matrix, allowing NaOH to depolymerize the ester bonds in PET, thereby forming terephthalate in its sodium form. In our study, we tested pretreatment under different conditions: 15% NaOH with autoclaving at 127°C for 15 and 30 minutes, and 15% NaOH using drybath for 4, 8, and 20 hours. This treatment effectively hydrolyzes the ester bonds at the surface while also modifying the polymer's structure to increase its accessibility and susceptibility to enzymatic hydrolysis (Dewi et al., 2025).

Alkaline Pretreatment of PET Using Autoclaving

After pretreatment with 15% NaOH and autoclaving for 15 minutes, we observed visual changes of the 100% PET textile and cotton-PET blends (55:45 and 65:35) (Figure 1). The treated 100% PET textile (Figure 1A, left tube) became fibrous and powdery, compared to the untreated sample (Figure 1A, right tube), which retained a smooth and tightly woven appearance. The untreated 55:45 cotton-PET blends (Figure 1B, left tube) were smooth and dense, whereas the treated textiles (Figure 1B, right tube) showed loosened fibers. Untreated 65:35 cotton-PET blend (Figure 1C, right) appeared compact, while treated textiles (Figure 1C, left) exhibited rugged fibers and a dull appearance.

Scanning electron microscopy (SEM) imaging further revealed the morphological changes in the fibers at magnification of 500x and 10,000x (Figure 2). The 100% PET textile (Figure 2A and 2B) showed fibers with rough and disrupted surfaces (Figure 2A) and the presence of surface ruptures in higher magnification (Figure 2B). The 65:35 cotton-PET blends exhibited uneven and loosened structure (Figure 2C), and roughened surfaces with grooves and peeling (Figure 2D). These observations confirmed the NaOH with autoclaving, effectively disrupting the fiber surface and structure of the PET fibers.

In addition, Fourier Transform Infrared Spectroscopy (FT-IR) analysis was used to determine the chemical modifications of pretreated textiles. Control spectra established the diagnostic peaks for cotton (–OH stretching at~3300 cm⁻¹ and C–O–C glycosidic bonds at 1159–995 cm⁻¹) and PET (ester carbonyl at 1712 cm⁻¹ and aromatic C–H at 724 cm⁻¹) (Figure 3A). The absorbance peak of ester carbonyl (–COO stretch) at 1712 cm⁻¹ were identified, and the reduction rate was calculated relative to the untreated control (Table 1).

After pretreatment with 15% NaOH and autoclaving for 15 minutes, the intensity of the PET ester carbonyl peak at 1712 cm⁻¹ decreased by 13% in pure PET textile (Figure 3B), confirming partial hydrolysis of ester bonds. In blended textiles, the 55:45 cotton-PET textile showed the highest reduction of 15% (Figure 3C), indicating that blends with less PET are more chemically susceptible, whereas the 65:35 cotton-PET blend displayed only a 4.7% reduction (Figure 3D), reflecting the greater resistance of PET-rich textiles. Overall, FT-IR results showed reduction of ester carbonyl after pretreatment, indicating partial disruption of ester linkages in PET matrix.

Alkaline Pretreatment of PET Using Dry Bath

Due to the safety risk and high energy consumption of autoclaving, we sought alternative methods to reduce PET crystallinity. Building on the evidence of physical and chemical disruption of PET fibers by 15% NaOH with autoclaving, we tested milder and more accessible pretreatments using 15% NaOH at 80°C for 4, 8, and 20 hours on the dry bath, as an alternative to harsher autoclaving conditions (Figure 4). Two types of 100% PET textiles (Textile A and B) and PET fibers were pretreated under these conditions to compare the crystallinity reduction achieved by each condition. After pretreatment, the samples were sent for SEM and DSC analysis to assess the morphology and crystallinity of PET samples.

SEM imaging of Textile A showed the changes in its morphology under different pretreatment conditions (Figure 5). Pretreatment with 15% NaOH at 80°C resulted in disruption of PET fibers progressively over time. After 4 hours of incubation, the fibers showed roughened surface (Figure 5A) and prolonged treatment for 8 hr exhibited thinning of the fibers and roughened edges (Figure 5B). Extended incubation for 20 hr further caused severe erosion and breakdown of the fiber (Figure 5C), showing effective PET structural disruption over time under 15% NaOH at 80°C pretreatment.

Differential scanning calorimetry (DSC) analysis was performed to further quantify the crystallinity of the PET textiles and fibers among all different pretreatment conditions. DSC results provided values for enthalpy of fusion ΔHm enthalpy of crystallization ΔHc as illustrated in Figure 5 for Textile A.. This two variable can be compared to the enthalpy of fusion theoretical value for 100% crystalline PET ΔHf, which is equal to 140 J/g, to calculate the sample’s crystallinity using the formula:

We summarized the degree of crystallinity of the samples across all tested conditions (Table 2). The untreated samples showed the crystallinities of 47.9%, 44.1% and 36.5% for Textile A, Textile B and PET fiber, respectively. Textile A and B both show a significant reduction in crystallinity under 15% NaOH at 127 °C for 15 and 30 minutes compared to untreated results. Consistent with the SEM results, pretreatment with 15% NaOH at 80°C showed time-dependent reduction in crystallinity with the 4 hr incubation decreased about half of the original crystallinity in all samples (22.1% for Textile A, 22.7% for Textile B and 11.7% for PET fiber), and dropped to as low as 1.5% for Textile A, 1.4% for Textile B and 5.6% for PET fiber when incubated 20 hr at 80°C. These results further indicate disruption in crystalline domains after pretreatment. Surprisingly, autoclaving for 15 minutes showed a lower crystallinity reduction compared to pretreatment at 80°C, indicating greater crystallinity reduction in moderate temperature for longer periods than short exposure to autoclaving temperature.

In summary, these results validated that our alkaline-thermal pretreatment strategies successfully altered fiber morphology, disrupted the PET matrix and reduced the crystallinity in PET textiles, establishing pretreatment as a critical first step to enhance the enzyme accessibility and enable efficient textile degradation. Considering only moderate crystallinity reduction is required for enzymatic degradation and ensuring the safety and practical conditions, we selected 15% NaOH at 80°C as the pretreatment conditions for subsequent degradation experiments. Although an 80°C dry bath for more than 8 hours achieves better decrystallization, the textile would be too fragile to wash by tap water since the textile became powder and escaped from the strainer. Therefore, we used 15% NaOH at 80°C for 4 hours as the pretreatment for the further experiments.

Pure PET textile degradation

After verifying the enzyme degradation on PET film and confirming the pretreatment conditions, we performed the enzyme degradation experiment on 100% PET textiles before moving on to cotton-PET blend textiles. 100% PET textiles have simpler components compared to blended textiles, making them ideal for testing enzyme performance on textile substrates. Success in high crystallinity pure PET textiles degradation indicated the viability of our process and its potential to degrade more complex cotton-PET blends.

Preliminary Tests on Pure PET Fiber Degradation

In the first round of pure PET textile degradation, we chose PET fiber because of its low crystallinity compared to textile A and B. Pretreatment was done by incubating the fibers in 15% NaOH at 80℃ dry bath for 8 hours. Since the PET fiber was too fragile to wash with tap water, we used 1 M HEPES buffer (pH 3.5 to 4) to neutralize the residual strong alkali for 3 volumes of NaOH. Around 30-80 mg of pretreated PET fibers were incubated in a reaction buffer containing 500 mM HEPES (pH 7.5) and 10 mM CaCl₂ with 27.8 μg/mL enzyme at 51.8℃ for 48 hours. However, we failed to detect the UV absorbance at 260 nm because all of the reading values from pretreated fiber were above the maximum limit of detection, including the negative control without enzyme. Even though we diluted the reaction buffer to 1/128, there is no significant difference among the groups (Figure 1). That may be because the PET degraded products still remain on the fiber, suppressing the enzymes or masking the signal caused by the enzymes.

Pure PET Textile Degradation of TfCut2 and Its Mutants

In order to investigate the characteristics of enzymatic degradation on 100% PET textiles, we utilized pure PET textiles from three different origins for degradation experiments: 100% PET fiber (Figure 2A), 100% PET textile A in weft-rib weave (Figure 2B), and 100% PET textile B in warp-knit mesh (Figure 2C).

To address the issue of excessive degradation products from pretreatment interfering with subsequent UV absorbance measurements, we shortened the pretreatment time from 8 hours to 4 hours. Instead of using HEPES buffer to neutralize the pretreated textile, we rinsed the pretreated textile with tap water for five minutes, followed by a rinse with deionized water. Finally, they were dried in an oven at 50°C for at least 48 hours before proceeding with enzymatic degradation.

We used approximately 30-80 mg of textiles in a 1.8 mL 500 mM HEPES with 10 mM CaCl₂ (pH 7.5) reaction buffer with 27.8 μg/mL enzyme. After incubating TfCut2 and its mutants in a thermoshaker at 51.8°C, 800 rpm for 48 hours, we utilized the ELISA reader to measure UV absorbance at 260 nm. Pretreated PET fibers released significantly more TPA than untreated samples. Variant 3 exhibited the highest activity, with absorbance at 260 nm nearly three-fold greater than the wild type. The wild-type 5ZOA and Variants 1, 2, 4, and 5 also showed increased TPA release for the pretreated samples, whereas Variant 6 displayed negligible activity. These results demonstrate that pretreatment significantly improves degradation efficiency and identify Variant 3 as the most effective enzyme for degrading pure PET textile.

Furthermore, the degradation efficiency of the enzyme was also observed to be dependent on the textile type. With the exception of Variant 6, all mutants and the wild type TfCut2 effectively degraded the pretreated PET fiber, and to a lesser extent, the pretreated textile B. The higher degradability of the pure PET fiber can be attributed to its loose structure and lower crystallinity. Notably, despite having comparable crystallinities, textile B showed higher susceptibility to degradation than textile A, implying that the textile's weaving pattern could be another contributing factor.

Next, we choose pretreated PET fiber, which is the most susceptible to our enzymes, to do the further HPLC analysis and SEM imaging. The HPLC results were consistent with the UV absorbance data, showing that Variant 3 (A65H/L90A/I213S) exhibited the highest degradation efficiency, degrading nearly 0.1 mg of PET (Table 1). This finding was further supported by SEM imaging. Control samples without enzyme treatment exhibited smooth fiber surfaces (Figure A and B). Fibers incubated with TfCut2 5ZOA wild type (Figure C and D), Variant 2 (H77R/D204C/E253C) (Figure G and H), Variant 4 (D12S) (Figure K and L), Variant 5 (T234L)(Figure M and N) and Variant 6 (Variant O and P) showed no noticeable surface alterations, whereas fibers treated with Variant 1(G62A/P193H/S197D) displayed uneven surface (Figure E and F). In contrast, Variant 3 (A65H/L90A/I213S) caused pronounced surface disruption, with visible grooves observed at higher magnification (Figure I and J), which were consistent with previous analyses, confirming the superior effectiveness of Variant 3(A65H/L90A/I213S) in PET degradation.

In conclusion, our results confirm that alkaline pretreatment significantly enhances enzymatic degradation in pure PET textiles, TfCut2-Variant 3 A65H/L90A/I213S triple mutant. outperformed TfCut2-5ZOA wild type by nearly two-fold, demonstrating clear potential for PET breakdown. These findings prove that our engineered enzymes are capable of degrading pure PET substrates and the foundation for more complex Cotton/PET blend textile degradation.

Cotton PET blend degradation

Our previous work established that TfCut2 mutants degrade PET, including alkali-pretreated, high-crystallinity textiles, more efficiently than the wild type. In this phase, we investigate their efficacy on the PET component of blend textiles. We will blend textiles with TfCut2 and its mutants and examine their ability to work in concert with cellulase. The ultimate goal is to develop a method for completely decomposing blended textiles, enabling their recycling and reuse within the textile industry to support a circular economy.

To assess the feasibility of our cotton-PET blend textile enzymatic degradation system, we prepared samples from PET Cotton blends (surface area ~ 62.5 cm^2) at three different PET contents: 35% cotton 65% PET blend textile (35:65 cotton-PET Blends) in plain weave (Figure 1A), 55% cotton 45% PET blend textile (55:45 cotton-PET Blends) in tricot warp knit (Figure 1B), and 65% cotton 35% PET blend textile (65:35 cotton-PET Blends) in single Jersey knit (Figure 1C).

In order to improve the efficiency of alkali pretreatment and enzymatic degradation, the textiles were cut into 0.1 x 0.5 cm pieces. Alkali pretreatment was performed according to the conditions described in the previous section. The purpose of this treatment was not only to reduce the crystallinity of the PET polymer but also to make the cotton component more susceptible to degradation by cellulase.

Cotton Degradation of Blend Textiles

After pretreatment, blend textiles were treated with or without commercial cellulase (Cellulase from Trichoderma reesei). Then we evaluated the cotton degradation via DNS assay (Table 1). DNS Analysis is used to provide quantitative data on the degradation rate within different cotton-PET textile blends, where it aids us in analyzing the amount of glucose released after enzymatic degradation. The glucose concentration determined from OD540 was converted into glucose mass and then multiplied by the initial common mass to calculate the overall degradation rate for each textile blend. The following equations were used to determine each measurement:

The results indicated that the 55:45 cotton-PET blend shows the highest degradation rate of 94.9% among the three textile ratios tested, where the 65:35 blend also showed strong enzymatic hydrolysis (81.7%), while the 35:65 blend resulted in a lower degradation rate (61.0%). These results successfully quantified reducing sugar release from different cotton-PET blend ratios and further demonstrate that moderate cotton contents provide an optimal condition for enzymatic attack.

PET Degradation of Blend Textiles

In this section, we conducted and evaluated PET degradation on cotton PET blend textiles with or without cotton degradation. After alkaline pretreatment and cotton degradation, we washed textiles with tap water for 5 minutes and rinsed with deionized water then dried in a 50℃ oven for at least 48 hours. Well-prepared blend textiles were aliquoted evenly into eight 2 mL microtubes, each containing 1.8 mL of 500 mM HEPES buffer (pH 7.5) with 10 mM CaCl₂ and 27.8 μg/mL enzymes. In other words, each reaction contained a total of 50 μg enzyme in a reaction volume of 1.8ml. After incubation on a thermoshaker at 51.8℃ 800 rpm for 48 hours, PET degraded products were measured by UV absorbance at 260 nm via ELISA reader.

As shown in Figure 2, TfCut2-Variant 3 A65H/L90A/I213S triple mutant (Variant 3) and TfCut2-Variant 4 D12S single mutant (Variant 4) exhibited the highest overall activity in degrading PET directly from the blend textiles, far exceeding the benchmark wild type enzyme TfCut2-5ZOA and other Variants. In contrast, other mutants showed little to no activity.

Interestingly, PET degradation activity was substantially lower after cotton removal compared to the direct treatment of the blended textiles. The most unexpected finding in this study was that the PET degradation efficiency was decreased, rather than increased, after the prior removal of the cotton component. We propose a hypothesis to explain this phenomenon based on the experimental protocol: the issue may stem from an incompatibility between the buffer systems used for the sequential enzymatic reaction. Another hypothesis is that the cotton fibers within the blend's structure may play a role in increasing the accessibility of PET fibers. The removal of cotton might lead to a collapse or tightening of the remaining PET structure, thereby reducing degradation efficiency.

Furthermore, our study revealed a key finding: the degradation efficiency of PET in blended textile does not have a direct relationship with its chemical composition—that is, the ratio of cotton to PET. For instance, although the 55:45 cotton PET blend yielded a relatively high concentration of degradation products, textiles with either higher or lower PET content did not exhibit a predictable corresponding trend of increase or decrease in degradation.

This lack of correlation strongly suggests that the dominant factor influencing enzymatic degradation may not be the chemical composition, but rather the physical architecture of the textile. Our data clearly indicate that the textile manufacturing method has a significant impact on the catalytic efficiency of TfCut2. Specifically, the two samples with a woven structure—the 35:65 cotton PET blend and the 100% PET Textile A—showed markedly lower degradation rates compared to the other samples, which possess a knitted structure. We hypothesize that the tightly interlaced warp and weft yarns in woven fabrics may limit enzyme penetration and diffusion, thus reducing the accessible surface area of the PET fibers. In contrast, the typically looser and more flexible loop structure of knitted fabrics may provide greater access to reaction sites for the enzyme, thereby facilitating a more effective degradation process.

To further validate the initial findings from the UV absorbance measurement, we selected the top-performing condition—the 55:45 cotton PET blend textile under direct enzymatic treatment—for subsequent experiments. The degradation products were then quantified more precisely using High-Performance Liquid Chromatography (HPLC) (Table 2) and supported by SEM imaging (Figure 4).

The HPLC analysis confirmed that Variant 3(A65H/L90A/I213S) was the most effective enzyme, liberating the highest concentration of degradation products (TPA and MHET) from the 55:45 cotton PET blend. In contrast, Variant 4 (D12S), which also appeared highly active in the initial UV screen, demonstrated only a marginal improvement in performance over the wild type enzyme in this more rigorous assay. These results underscore that while UV absorbance is a useful tool for rapid preliminary screening, HPLC analysis provides a more accurate differentiation of the true catalytic efficiencies among enzyme Variants.

In conclusion, we validated the use of TfCut2 Variants for the enzymatic degradation of PET from cotton-polyester blended textiles, identifying Variant 3(A65H/L90A/I213S) as the most effective catalyst. Our key finding is that the textile's physical architecture, specifically its knitted or woven structure, is a more critical determinant of degradation efficiency than the chemical ratio of cotton to PET. Unexpectedly, the prior removal of cotton inhibited the most active enzymes, likely due to experimental artifacts such as buffer incompatibility or structural collapse of the remaining PET fibers. These results collectively underscore the necessity of a holistic approach that considers both enzyme performance and the physical characteristics of textile waste to develop robust and efficient biorecycling strategies.

Discussion

Key Findings:

1. The optimal condition for pNPB assays are within pH ranges 7~8, and at temperature 35℃.
2. The conditions for Tris buffer and HEPES buffer are slightly different in terms of pH ranges, where the optimal condition for Tris buffer is between 7.5~8, the pH range for HEPES buffer is between 7~7.5.
3. All variants are proven to have activity, but Variant 6 exhibit sufficiently lower activities.

1. The optimal condition for TfCut2 and its mutants is 51.8℃ pH 7.5.
2. PET film degradation shows different optimal conditions from pNPB assay.
3. Variants 3, 4 show better PET film degradation activity than 5ZOA.
4. Variant 3 can degrade PET in a lower concentration of buffer (100 mM HEPES) while other mutants and wild type cannot.

1. Alkaline pretreatment is necessary for high crystallinity PET degradation.
2. Variants 3 show better PET textile degradation activity than 5ZOA.
3. TfCut2 and its mutants show better degradation activity in knitted structure (textile B) rather than woven structure (textile A).

1. PET degradation efficiency was decreased in cotton degraded blend textiles.
2. TfCut2 and its mutants show better degradation activity in knitted structure (55:45 cotton PET and 65:35 cotton PET) rather than woven structure (35:65 cotton PET).
3. Variants 3 show better PET degradation activity in blend textile than 5ZOA.

This study successfully characterized and engineered the wild-type TfCut2 enzyme (5ZOA) and a series of its variants, systematically evaluating their potential to degrade various forms of poly(ethylene terephthalate) (PET). By integrating rational design, literature-based mutations, and computational predictions, we aimed to enhance the enzyme's catalytic activity, stability, and substrate interaction.

Our findings reveal a complex interplay between mutation strategy and actual enzymatic performance. Variant 3 (A65H/L90A/I213S) consistently emerged as the most effective enzyme across all tested substrates, from PET film to complex blend textiles. Its design, which combined mutations inspired by the highly efficient LCC enzyme (L90A/I213S) with a novel mutation to increase PET docking sites (A65H), proved exceptionally successful. The superior performance of Variant 3 validates the hypothesis that reducing product inhibition by MHET through the creation of alternative substrate binding positions is a highly effective strategy for engineering PET-degrading enzymes.

Furthermore, the MutCompute-driven Variant 4 (D12S) also demonstrated better degradation activity than the wild type, showing that data-driven predictions can successfully identify beneficial single-point mutations. Conversely, Variant 1 (G62A/P193H/S197D), designed to incorporate a catalytic triad, did not yield a significant improvement over the wild type, indicating that theoretical catalytic enhancements do not always translate to effective degradation of a solid polymer substrate. Meanwhile, Variant 6, which incorporated all non-conserved stabilizing mutations, exhibited a drastic loss of activity, suggesting that excessive structural stabilization may restrict the conformational flexibility required for catalysis—a critical insight for future enzyme engineering.

A key finding of this study is the profound influence of textile physical structure on enzymatic degradation efficiency. All effective variants showed a clear preference for the knitted structure over the woven structure. This highlights that macroscopic physical accessibility is as critical as biochemical activity. Furthermore, we also observed that PET degradation was suppressed during the degradation of cotton, revealing the challenges of processing blended textiles.

Looking ahead, this study opens several avenues for future investigation. The discrepancy in optimal conditions between the pNPB model substrate and the actual PET substrate warrants further exploration; we hypothesize that the PBS buffer used during purification may interfere with calcium ions, and future work should investigate using an alternative such as borate buffer. The preference for knitted structures and the inhibitory mechanism of cotton on PET degradation require deeper investigation, possibly by testing different cellulases in concert with our variants.

To advance these enzymes toward practical application, long-term degradation experiments are necessary to assess their thermostability and improve total PET conversion, which may require specialized hardware and computational support. Finally, to overcome the challenges of quantifying reaction kinetics on a solid substrate, we propose the development of a TPA dual reporter system for more precise quantification.

In summary, this research not only identifies a highly efficient TfCut2 mutant (Variant 3) but also reveals critical factors influencing enzymatic efficiency, from substrate accessibility at the molecular level to textile structure at the macroscopic level. These findings lay a crucial foundation for developing robust enzymatic cocktails for the recycling of complex textiles, bringing the vision of a circular economy for the textile industry one step closer to reality.

Future Works:

1. Our optimal degradation condition differs from prior studies, potentially due to PBS used during protein purification interacting with calcium ions; therefore, future tests should explore borate buffers to enhance enzyme stability and activity.
2. Because TfCut2 and its mutants performed better on knitted structures than woven ones, further work should test different cellulases and textile structures to verify this pattern
3. PET degradation declined after cotton breakdown; future studies should determine whether this results from buffer issues, fiber collapse, or enzyme interactions.
4. Extended degradation experiments are needed to assess thermostability and improve PET hydrolysis with better temperature control and hardware.
5. A TPA dual-reporter system could enable real-time, quantitative tracking of PET degradation and kinetic modeling of TfCut2 variants.