Immobilized hydroxymethylfurfural oxidase enables robust biocatalytic production of 2,5-furandicarboxylic acid from crude 5-hydroxymethylfurfural

IMMOBILIZED HYDROXYMETHYLFURFURAL OXIDASE ENABLES ROBUST 1 BIOCATALYTIC PRODUCTION OF 2,5-FURANDICARBOXYLIC ACID FROM 2 CRUDE 5-HYDROXYMETHYLFURFURAL 3 Darly Concha †, Garazi Ortiz-Orruño †, Marina Guillén †, Oscar Romero †, Kírian 4 Bonet-Ragel † * 5 † Bioprocess Engineering and Applied Biocatalysis Group, Department of Chemical, 6 Biological and Environmental Engineering, Universitat Autònoma de Barcelona, 08193 7 Bellaterra, Spain. 8 Corresponding Author 9 * Kírian Bonet-Ragel 10 Department of Chemical, Biological and Environmental Engineering, Universitat 11 Autònoma de Barcelona, 08193 Bellaterra, Catalonia, Spain. 12 Telephone: +34 93 581 4791 13 [email protected] 14 15 16 17 ABSTRACT 18 Bioplastics such as poly(ethylene 2,5-furandicarboxylate) (PEF), synthesized from bio-19 based 2,5-furandicarboxylic acid (FDCA), offer a promising alternative to conventional 20 petrochemical-derived plastics. The enzymatic conversion of 5-hydroxymethylfurfural 21 (HMF) to FDCA using engineered variants of hydroxymethylfurfural oxidase (HMFO) 22 represents an environmentally benign approach, as the enzyme catalyzes all three 23 oxidation steps required for FDCA synthesis. 24 This study presents an intensified biocatalytic system for the sustainable production of 25 FDCA, employing the engineered enzyme 8BxHMFO fused with a carbohydrate-binding 26 module (CBM3) and immobilized on microcrystalline cellulose (Perloza MT100) to 27 avoid oxygen interfacial inactivation. The system was first optimized using pure HMF, 28 achieving high conversion rates (>95%), a maximum FDCA yield of 7.2 g/L. Reactions 29 were performed using a crude extract rich in HMF (15% wt/wt). Despite the presence of 30 impurities, the immobilized enzyme maintained high conversion efficiency, with only a 31 30% reduction in FDCA yield compared to pure substrate. FDCA recovery was achieved 32 with 72% efficiency using a mild, ethanol-based extraction process. 33 These findings, along with an evaluation of sustainable indicators, validate the robustness 34 of the immobilized HMFO system under realistic conditions and demonstrate its potential 35 as a sustainable platform for FDCA production, contributing to the broader adoption of 36 PEF bioplastics. 37 KEYWORDS 38 Biocatalysis; 2,5-furandicarboxylic acid (FDCA); 5-hydroxymethylfurfural (HMF); 39 Enzyme immobilization; Hydroxymethylfurfural oxidase (HMFO); Cellulose-based 40 supports; Sustainable plastics. 41 HIGHLIGHTS 42  One-step immobilization on renewable cellulose yields stable and cost-efficient 43 biocatalyst. 44  Immobilized HMFO (CBM3-8BxHMFO) enables efficient FDCA production 45 from crude HMF extract. 46  Conversion of crude HMF achieved an excellent 70.6% FDCA yield and good 47 titter (5.7 g/L). 48  Sustainable and efficient mild ethanol-based downstream process recover to 49 FDCA. 50 51 GRAPHICAL ABSTRACT 52 53 54 SYNOPSIS 55 This work demonstrates a green biocatalytic route converting lignocellulosic HMF into 56 FDCA, advancing sustainable polymer production through enzyme immobilization and 57 renewable supports. 58 1. INTRODUCTION 59 The widespread use of plastic materials, while indispensable to modern life due to their 60 durability and versatility, has generated severe environmental consequences. Among 61 them, polyethylene terephthalate (PET), a fossil-derived polymer extensively used in 62 packaging and textiles, contributes substantially to plastic pollution in terrestrial and 63 marine ecosystems (Djapovic et al., 2021; Sanders et al., 2024; Singh et al., 2021). The 64 shortcomings of current recycling technologies, which remain energetically intensive and 65 environmentally detrimental, underscore the urgency of developing sustainable 66 alternatives (Maurya et al., 2020; Ügdüler et al., 2020). 67 In response to these sustainability challenges, poly(ethylene 2,5-furandicarboxylate) 68 (PEF) has emerged as a promising bio-based alternative to PET. PEF is synthesized from 69 renewable resources and utilizes 2,5-furandicarboxylic acid (FDCA) as a monomer, 70 replacing terephthalic acid in conventional polyesters. In addition to its sustainable origin, 71 PEF offers superior barrier properties and thermal stability, making it a competitive 72 bioplastic for industrial applications (Eerhart et al., 2012; Hwang et al., 2020; Kumar et 73 al., 2024). 74 FDCA can be produced via the selective oxidation of 5-hydroxymethylfurfural (HMF), a 75 platform chemical obtained from lignocellulosic biomass. Traditional chemical oxidation 76 routes, such as noble metal or transition metal oxide catalysis, require harsh reaction 77 conditions, elevated temperatures, and toxic solvents, leading to undesirable byproducts 78 and high processing costs (Pei et al., 2023; Sajid et al., 2018; Totaro et al., 2022). Recent 79 advances using heterogeneous catalysts, including Co-based oxides or microwave-80 assisted noble metal systems, have improved activity and energy efficiency but still rely 81 on chemical oxidants and pressurized conditions (Liu et al., 2020; Ji et al., 2018; Peng et 82 al., 2025). In contrast, enzymatic oxidation offers an environmentally benign alternative, 83 operating under mild conditions, employing molecular oxygen as a clean oxidant, and 84 delivering high selectivity toward FDCA (Meyer et al., 2022; Troiano et al., 2020). 85 Enzymatic oxidation of HMF frequently relies on multi-enzymatic cascades, which 86 complicates process optimization. Nevertheless, hydroxymethylfurfural oxidase 87 (HMFO), a flavoprotein oxidoreductase belonging to the glucose-methanol-choline 88 (GMC) enzyme superfamily, exhibits notable substrate promiscuity, catalyzing the 89 oxidation of a broad range of furanic compounds, including the primary alcohol oxidation 90 of HMF (Dijkman et al., 2015; Dijkman and Fraaije, 2014). These enzymes contain flavin 91 adenine dinucleotide (FAD) as a tightly bound prosthetic group, which is essential for 92 their redox activity (Dijkman and Fraaije, 2014; Viñambres et al., 2020; Wu et al., 2021). 93 Native HMFOs, such as those from Methylovorus sp. MP688, demonstrate high catalytic 94 efficiency in oxidizing both alcohol and aldehyde functional groups (Dijkman and 95 Fraaije, 2014; Viñambres et al., 2020). The enzyme catalyzes the full oxidation of HMF 96 to FDCA through a sequential three-step process that includes the formation of 97 intermediate compounds: 2,5-diformylfuran (DFF) and 5-formylfuran-2-carboxylic acid 98 (FFCA). 99 Advances in protein engineering have enabled the development of improved HMFO 100 variants, such as 8BxHMFO, capable of catalyzing complete oxidation of HMF into 101 FDCA, exhibiting enhanced thermal stability and reduced susceptibility to oxidative 102 deactivation position them as promising candidates for industrial-scale biocatalytic 103 applications (Martin et al., 2018; Sánchez-Ruiz et al., 2021). Nevertheless, their practical 104 implementation is still hampered by challenges such as poor stability at gas-liquid 105 interfaces, product inhibition, particularly associated with hydrogen peroxide 106 accumulation, and low expression yields in recombinant hosts (Dijkman and Fraaije, 107 2014; Høst et al., 2025; Sánchez-Ruiz et al., 2021). To overcome these limitations, recent 108 efforts have focused on rational enzyme design and fine-tuning of reaction parameters to 109 enhance overall catalytic efficiency and robustness (Wu et al., 2021). 110 However, HMFOs still faces challenges in operational stability due to their oxidative 111 mechanism. The use of molecular oxygen as an electron acceptor leads to local oxygen 112 depletion and the formation of reactive oxygen species (ROS), which may induce enzyme 113 deactivation. Furthermore, the gas-liquid interface in aerated bioreactors promotes 114 protein denaturation via interfacial adsorption and aggregation (Høst et al., 2025; Wang 115 et al., 2023). To overcome these limitations, enzyme immobilization has been widely 116 adopted as a strategy to improve catalytic stability, enable enzyme reuse, and facilitate 117 integration into continuous processes (Maghraby et al., 2023; Ndochinwa et al., 2024). 118 Enzyme immobilization is a well-established technique that improves catalytic stability, 119 enhances resistance to denaturation, and enables application in multi-enzyme cascade 120 systems (Robescu and Bavaro, 2025). Additionally, immobilization facilitates enzyme 121 recovery and reuse, offering a cost-effective and sustainable solution (Maghraby et al., 122 2023). 123 In particular, immobilization via carbohydrate-binding modules (CBMs) offers a 124 sustainable and efficient platform. This approach aligns with the principles of green 125 chemistry by enabling enzyme purification and immobilization in a single step using 126 renewable supports such as microcrystalline cellulose. CBM3 from Clostridium 127 thermocellum binds specifically to crystalline cellulose and allows site-directed 128 immobilization (Benito et al., 2022; Oliveira et al., 2015). It also allows one-step 129 purification and immobilization through specific carbohydrate interactions, providing a 130 sustainable and cost-efficient alternative to traditional immobilization based on chemical 131 functionalization (Benito et al., 2022; Roberts et al., 2021). 132 Despite these developments underscore the increasing relevance of HMFOs in industrial 133 biotechnology, playing a central role in advancing sustainable polymer production 134 (Herlina et al., 2023; Sánchez-Ruiz et al., 2021; Viñambres et al., 2020; Wu et al., 2020). 135 Most biocatalytic studies for FDCA production have employed pure HMF as the 136 substrate, which does not reflect the complexity of real biomass-derived hydrolysates. 137 Crude HMF streams often contain impurities such as humins, organic acids, and residual 138 sugars that may inhibit enzyme activity or interfere with downstream processing 139 (Birmingham et al., 2021; Wang et al., 2024). Therefore, it is essential to validate the 140 performance of immobilized biocatalysts under these more industrial/biorefinery relevant 141 conditions to ensure their feasibility for industrial implementation. 142 In this study, we report the expression of the engineered 8BxHMFO enzyme fused to 143 CBM3, its immobilization on a microcrystalline cellulose supports, and its application in 144 the biocatalytic production of FDCA. To enhance enzyme stability, the immobilized 145 biocatalyst was further crosslinked with glutaraldehyde. The catalytic performance was 146 evaluated using both pure HMF and crude HMF. This work applied an intensification 147 strategy, in which multiple HMF concentrations were tested. The results contribute to the 148 integration of biocatalytic systems into circular biorefinery schemes for bio-based 149 polymer manufacturing, highlighting the system's adaptability to realistic feedstocks and 150 alignment with sustainability goals. 151 2. METHODS 152 2.1.Materials 153 5-Hydroxymethylfurfural (HMF), 2,5-diformylfuran (DFF), 5-formyl-2-furancarboxylic 154 acid (FFCA), and 2,5-furandicarboxylic acid (FDCA) were purchased from TCI 155 Chemicals (Japan). 25 – 35% 5-Hydroxymethylfurfural (Crude HMF) was obtained from 156 AVA Biochem (Switzerland). Flavin adenine dinucleotide (FAD) was obtained from 157 Sigma-Aldrich (St. Louis, MO, USA). Microcrystalline cellulose Perloza MT100 was 158 provided by Perloza s.r.o. (Czech Republic). Avicel® PH-200 microcrystalline cellulose 159 sample was kindly donated by DuPont™ N&B (New York, NY, USA). 160 2.2.Enzymes 161 An octuple mutant variant of HMF oxidase (8BxHMFO; I73V, H74Y, G356K, V367R, 162 T414K, A419Y, A435E, and W466F) from Methylovorus sp. was genetically fused to a 163 type 3 cellulose-binding module (CBM3) from Clostridium thermocellum. The synthetic 164 gene was obtained from GenScript (USA) and cloned into the pVEF expression vector. 165 The fusion enzyme was expressed in an auxotrophic M15-derived Escherichia coli strain, 166 cultivated in high cell density cultures in fed-batch reactor using an antibiotic-free defined 167 medium (Vidal et al., 2008). Cell disruption was performed at 1.6 kbar using high-168 pressure disruptor (Constant Systems cell disruptor), and the enzyme was recovered in 169 the soluble fraction of the lysate, yielding a specific activity of 14.8 ± 0.7 U·gDCW⁻¹, a 170 volumetric activity of 1209.5 ± 59.3 U·L⁻¹, and an enzyme concentration of 2762.3 ± 171 126.5 mg·L⁻¹. This corresponded a specific activity of 0.44 ± 0.03 U·mg⁻¹ . Catalase was 172 purchased from Sigma-Aldrich (St. Louis, MO, USA) and used as auxiliary enzyme. 173 2.3.HMFO Activity assay 174 The oxidase activity of CBM3-8BxHMFO was measured by monitoring oxygen 175 consumption using a Pyro Science robust oxygen probe coupled to a FireSting-PRO 176 oxygen meter. Reactions were performed in 2.5 mL of 50 mM Tris-HCl buffer (pH 8.0) 177 at 30°C with 300 rpm agitation. Enzymatic reactions were initiated by adding the enzyme 178 (soluble and immobilized) and 60 mM HMF. One unit of enzymatic activity was defined 179 as the amount of enzyme required to consume 1 μmol of molecular oxygen per minute 180 under these conditions. 181 can constitute 50-80% of total manufacturing expenses (Barbosa et al., 2015; Boodhoo et 323 al., 2022; Rathore and Kapoor, 2015). 324 325 Figure 1: A) Immobilization kinetics of CBM3-8BxHMFO in Perloza MT100 Support. 326 B) SDS-PAGE: lanes 1 and 6 MWM (kDa); (CBM3-8BxHMFO, MW: 75.1 kDa) lanes 327 2 and 4: blank and immobilized enzyme onto Avicel® PH-200; lanes 4 and 5: blank and 328 immobilized enzyme onto Perloza MT100. 329 Table S1summarizes key performance indicators of one step purification immobilization 330 process. A high immobilization yield and recovery activity of 72% and 16% was obtained 331 for Avicel® PH-20, and 98% and 10% for Perloza MT100, respectively. Both supports 332 exhibited a high protein loading capacity of 50 mg per gram of support. These outcomes 333 are consistent with previous reports employing various microcrystalline cellulose 334 matrices such as Avicel® PH-101, regenerated amorphous cellulose (RAC), and bacterial 335 nanocellulose which have demonstrated similar or slightly lower adsorption capacities 336 (Benito et al., 2022; Li et al., 2025; Qin et al., 2019; Ramón-Luing et al., 2006). The high 337 immobilization efficiency and rapid binding kinetics observed in this study support the 338 effectiveness of CBM3-based affinity immobilization as a robust, one-step purification 339 strategy. The low-cost process and renewable nature of the support highlight its industrial 340 relevance, particularly in the context of sustainable biocatalyst design. Such an approach 341 directly aligns with the principles of circular economy, leveraging low-cost agro-342 industrial residues and enhancing the environmental sustainability of bioprocesses 343 (Benito et al., 2022; Oliveira et al., 2015). 344 First attempt of FDCA production using immobilized CBM3-8BxHMFO showed a 345 progressive loss of enzyme activity (Figure S3). This phenomenon was attributed to the 346 desorption of the enzyme from the cellulose matrix, which released the enzyme into the 347 solution where it was likely denatured (Figure S4 and Table S2). This outcome is 348 consistent with recent studies demonstrating that CBM-cellulose interactions, particularly 349 those involving CBM3, are dynamic and reversible, with possible elution in the presence 350 of glucose or low-polarity solvents (Liu et al., 2022; Oliveira et al., 2015). 351 To improve retention and enhance operational stability, glutaraldehyde was employed as 352 cross-linking agent. This bifunctional reagent can form covalent bonds between enzyme 353 molecules. Its effectiveness in stabilizing immobilized enzymes by reinforcing their 354 structural conformation is well supported in the literature (Mateo et al., 2007; Sheldon 355 and van Pelt, 2013) 356 Figure 2A shows that cross-linking with glutaraldehyde 0.25% resulted in a moderate 357 reduction in recovered enzymatic activity to 20%. This decrease suggests partial 358 restriction of enzyme flexibility, possibly impairing substrate access to the active site. 359 Notably, in the absence of glutaraldehyde, only 30% of the initial activity was recovered, 360 which can be attributed to the high protein loading on the support estimated at nearly 50 361 mg protein per gram of support. Such high surface occupancy may promote a hinder 362 substrate diffusion, leading to limited catalytic performance (Arana-Peña et al., 2020). At 363 higher concentrations (1%), catalytic activity was almost complete reduced, indicating 364 that excessive cross-linking severely hinders enzymatic activity. 365 SDS-PAGE analysis of the cross-linked derivatives (Figure 2B) confirmed enhanced 366 enzyme retention on the support, as demonstrated by the progressive disappearance of the 367 soluble enzyme band with increasing glutaraldehyde concentration. This pattern reflects 368 stronger covalent bonding between the enzyme and among enzyme molecules, improving 369 immobilization stability. However, this enhanced retention came at the cost of catalytic 370 performance. At 0.25% glutaraldehyde, recovered enzymatic activity dropped to 20%, 371 suggesting a trade-off between retention and catalytic functionality. Notably, high 372 concentrations of glutaraldehyde (e.g., 1%) resulted in near-complete inactivation, 373 consistent with previous reports describing enzyme deactivation due to excessive cross-374 linking (Liu et al., 2024; Modenez et al., 2018; Perzon et al., 2017; Xu et al., 2018). While 375 low glutaraldehyde concentrations have been shown to preserve catalytic function and 376 reduce leaching (Liu et al., 2024; Modenez et al., 2018), the 0.25% cross-linked derivative 377 was selected for subsequent experiments as the most balanced condition, providing 378 improved retention while maintaining acceptable activity. 379 380 Figure 2: A) Recovery activity of the different concentrations of glutaraldehyde used as 381 crosslinker. B) SDS-PAGE gel of immobilization fractions with different glutaraldehyde 382 concentration, where lane 1 is the molecular weight marker in kDa; (CBM3-8BxHMFO, 383 MW = 75.1 kDa). Lanes 2 to 6 show the immobilized derivative with 0, 0.1, 0.25, 0.5, 1 384 % glutaraldehyde respectively. 385 3.2.CHARACTERIZATION OF IMMOBILIZED DERIVATIVES: THERMAL 386 STABILITY AND MICROSCOPIC DISTRIBUTION 387 To evaluate the impact of immobilization and cross-linking on the structural stability of 388 CBM3-8BxHMFO, the melting temperature (Tm) was measured for the soluble enzyme, 389 its immobilized form, and the glutaraldehyde cross-linked derivatives (Table S3). The 390 immobilized enzyme exhibited a Tm of 53.2°C, which represents a significant increase 391 from the Tm of the soluble counterpart (49.7°C). This enhancement in thermal stability is 392 likely attributable to the conformational restrictions imposed by the covalent and non-393 covalent interactions with the cellulose matrix, which can limit the flexibility of the 394 polypeptide chain and reduce the likelihood of thermal unfolding (Ariaeenejad and 395 Motamedi, 2025; Rodrigues et al., 2021; Shen et al., 2024). Such stabilization has been 396 widely observed for immobilized enzymes and is particularly relevant for biocatalytic 397 applications requiring elevated operational temperatures. The observed increase in Tm 398 further supports the role of immobilization not only in facilitating enzyme reuse but also 399 in enhancing resistance to thermal denaturation, a desirable trait for long-term processing 400 under industrial conditions. 401 Conversely, increasing the concentration of glutaraldehyde beyond the optimal range led 402 to a slight but consistent decrease in Tm values. This observation, consistent with 403 previously reports, suggests that excessive cross-linking can introduce localized structural 404 stresses or distortions, disrupting the optimal enzyme conformations required for thermal 405 resilience (Barbosa et al., 2014; López-Gallego et al., 2005). In contrast, variations in 406 enzyme loading, even at high immobilization densities (51.9 mg·g-1support), did not lead to 407 measurable effects on thermal stability within the tested range. This indicates that the 408 conformational stabilization conferred by the support matrix predominates over potential 409 crowding effects (Ariaeenejad and Motamedi, 2025), at least within the conditions 410 evaluated. 411 To gain detailed spatial insights into the localization and distribution of the immobilized 412 enzyme within the microcrystalline cellulose matrix, confocal fluorescence microscopy 413 was carried out. Leveraging the intrinsic fluorescence of the FAD cofactor in 8BxHMFO, 414 visualization was achieved without the need for external labelling, preserving the native 415 structure of the biocatalyst. Multiplanar imaging revealed a uniform and homogeneous 416 distribution of the enzyme throughout the porous structure of the microcrystalline 417 cellulose particles, as observed in different focal planes (Figure 3). This spatial uniformity 418 confirms effective diffusion and attachment of the enzyme across the full depth of the 419 support, which is a desirable feature for heterogeneous catalysis. Such even spatial 420 distribution facilitates consistent substrate accessibility across the matrix and minimizes 421 diffusional gradients, which can otherwise limit reaction efficiency in packed-bed or 422 batch systems (Diamanti and López-Gallego, 2024; Santiago-Arcos et al., 2023). 423 Moreover, this structural uniformity is expected to enhance the operational stability of 424 the system by preventing localized enzyme overloading and associated mass transfer 425 limitations (Bolivar and López-Gallego, 2020). 426 427 Figure 3: Confocal fluorescence images of three different focal planes of the CBM3-428 8BxHMFO in Perloza MT100 Support. 429 3.3.INTENSIFICATION OF ENZYMATIC FDCA PRODUCTION USING PURE 430 HMF 431 To assess the catalytic performance of the immobilized CBM3-8BxHMFO system, batch 432 reactions were carried out using varying concentrations of pure HMF as substrate. The 433 reactions were conducted in a 20 mL basket reactor, which maintained a constant air 434 supply 12 mL·min-1 (0.6 vvm), temperature 30°C, and pH 8, scheme of the reaction 435 system is shown in Figure S5. This evaluation aimed to determine the optimal conditions 436 for conversion to FDCA, while simultaneously quantifying the accumulation of reaction 437 intermediates and assessing system productivity. 438 The enzymatic oxidation of HMF to FDCA was studied under a range of substrate 439 concentrations. Figure 4 focuses on the enzymatic reaction of CBM3-8BxHMFO 440 immobilized on Perloza with pure HMF. It includes the reaction profiles with (A) 6 mM, 441 (B) 12 mM, (C) 24 mM, (D) 50 mM, and (E) 125 mM HMF, as well as (F) the key process 442 metrics (conversion, yield, productivity, and FDCA titer) obtained from these reactions. 443 HMF was rapidly consumed under all tested conditions, and no significant accumulation 444 of intermediates was observed, except in the reaction conducted at 125 mM HMF, where 445 a notable accumulation occurred. In this case, DFF accumulated to approximately 50 mM 446 and persisted throughout the process, indicating inefficient oxidation. Additionally, 447 FFCA, a known potent inhibitor of 8BxHMFO (Sánchez-Ruiz et al., 2021) also 448 accumulated significantly, reaching concentrations of about 37 mM after 72 hours. This 449 accumulation of FFCA correlated directly with the lower FDCA yields observed in this 450 reaction compared to those with lower HMF concentrations, which showed negligible 451 intermediate accumulation and achieved near complete FDCA yields. It is also important 452 to note that reaction time increased proportionally with substrate concentration, further 453 influencing productivity at higher HMF levels. 454 455 Figure 4. Enzymatic reaction of CBM3-8BxHMFO immobilized on Perloza with pure 456 HMF at 30 °C, pH 8, 500 rpm, 0.6 vvm air bubbling, and catalase (97.7 U mL⁻¹) at 457 different substrate concentrations. Reactions with (A) 6 mM, (B) 12 mM, (C) 24 mM, (D) 458 50 mM, and (E) 125 mM HMF. (F) Key process metrics of the enzymatic reactions 459 (conversion, yield, productivity, and FDCA titer) obtained with pure HMF at 6, 12, 24, 460 50, and 125 mM. 461 From the reactions carried out at different HMF concentrations, key process metrics such 462 as conversion, yield, and space-time yield (STY) were calculated to identify optimal 463 conditions. The results, summarized in Figure 4 F, show that high conversion rates were 464 consistently achieved across all tested concentrations. FDCA yield increased 465 progressively with substrate concentration up to 50 mM, beyond which no significant 466 improvement was observed. In contrast, productivity peaked at 12 mM and showed a 467 slight decline at 24 and 50 mM, with a marked decrease observed for the reaction 468 performed at 125 mM HMF. This behavior is directly related to reaction time, as substrate 469 concentration increases, longer reaction times are required, affecting the productivity. 470 Finally, the total amount of FDCA produced (g·L-1) increased proportionally with the 471 substrate concentration, as expected from the direct relationship with yield, reaching a 472 maximum FDCA titer in the 50 mM reaction. 473 This behaviour is consistent with the accumulation of inhibitory intermediates such as 474 FFCA at high HMF concentrations, particularly in the case of the reaction performed at 475 125 mM HMF, which can adversely affect enzymatic turnover (Sánchez-Ruiz et al., 476 2021). Among the evaluated conditions, the reaction with 50 mM HMF provided the 477 highest FDCA titer (7.2 g·L-1), representing the best compromise between high yield and 478 productivity. This performance is notably higher than most enzymatic systems reported 479 to date, which typically reach ≤2 g·L-1 FDCA under substrate loadings of 1.5–10 mM 480 HMF using wild-type HMFO or its early variants (Dijkman and Fraaije, 2014; Sánchez-481 Ruiz et al., 2021; Viñambres et al., 2020). Even engineered enzymes such as 8BxHMFO 482 generally yielded <2 g·L-1 under similar conditions (Sánchez-Ruiz et al., 2021), while 483 more recent cascades such as BpLac/CglAlcOx achieved only ~0.8 g·L-1 FDCA at 5 mM 484 HMF (Yang et al., 2023). Only very recent advances with alternative enzymes, such as 485 the evolved PedH variant, have approached comparable levels, reaching ~6 g·L-1 from 40 486 mM HMF (Liu et al., 2025). 487 3.4.EVALUATION OF IMMOBILIZED BIOCATALYTS UNDER RELEVANT 488 INDUSTRIAL/BIOREFINERY CONDITIONS. 489 To evaluate the system’s applicability under real biorefinery, the reaction was performed 490 using a crude substrate extract containing 15% wt/wt of HMF. The reaction was carried 491 out at a 50 mM HMF concentration, a condition that corresponds to the highest titer using 492 pure HMF. As shown in Figure 5, substrate conversion was nearly complete. DFF levels 493 increased initially but were depleted after 73 hours, while FFCA remained at sub-494 inhibitory concentrations. The FDCA production profile showed a sustained increase 495 throughout the process, indicating that the immobilized enzyme remained functionally 496 active even in the presence of potential impurities present in the crude feedstock. 497 498 Figure 5. Enzymatic reaction of CBM3-8BxHMFO immobilized on Perloza with 51 mM 499 crude HMF at 30°C, pH8, 500rpm, 0.6vvm air bubbling and catalase (97.7U·mL-1). 500 A comparison between reactions using pure and crude HMF is presented in Table 1. The 501 immobilized CBM3-8BxHMFO efficiently oxidized pure HMF, achieving 99.7% 502 conversion with a 91.3% yield, corresponding to 7.2 g·L-1 FDCA and a space-time yield 503 (STY) of 89 mg·L-1·h-1 in 81 h. When crude HMF was used under the same conditions, 504 the conversion remained high (98.7%), but the yield dropped to 70.6%, resulting in 5.7 505 g·L-1 FDCA and a lower STY of 57 mg·L-1·h-1 over 100 h. 506 This decline is likely attributable to the presence of inhibitory compounds such as furfural 507 derivatives, phenolics, or residual sugars commonly found in non-purified lignocellulosic 508 hydrolysates (Wang et al., 2024). These components may interfere with enzymatic 509 turnover or destabilize the biocatalyst over prolonged reaction times. In the present study, 510 a characterization of the crude extract was performed (Figure S6), and no compounds of 511 this nature could be identified, even though non-identified impurities or by-products were 512 detected which may act as enzyme inhibitors. Nonetheless, the enzyme’s ability to sustain 513 FDCA formation in such a complex matrix highlights its robustness and supports its 514 practical application in the context of biorefinery integration and circular economy 515 strategies. To our knowledge, no reports using enzymes (soluble or immobilized) have 516 previously described the production of FDCA from crude HMF. However, Birmingham 517 et al. (2021) demonstrated that even the partial oxidation of crude HMF to DFF required 518 extensive enzyme engineering (GOase variants) and biphasic systems to maintain high 519 conversion at 100 g·L-1. Similarly, Milić et al. (2023) reviewed patents (Rode et al., 2021) 520 reporting low FDCA titers (20–50 mM) when Pseudomonas M11 resting cells processed 521 crude lignocellulosic extracts containing residual HMF, while only 66% yield was 522 achieved with Klebsiella oxytoca MCC0144 using HMF extracted from fructose. These 523 examples highlight that unrefined feedstocks strongly reduce both yield and achievable 524 product titters. In comparison, the immobilized CBM3-8BxHMFO maintains high 525 Birmingham, W.R., Toftgaard Pedersen, A., Dias Gomes, M., Bøje Madsen, M., Breuer, 658 M., Woodley, J.M., Turner, N.J., 2021. Toward scalable biocatalytic conversion of 659 5-hydroxymethylfurfural by galactose oxidase using coordinated reaction and 660 enzyme engineering. Nat Commun 12. https://doi.org/10.1038/s41467-021-25034-3 661 Bolivar, J.M., López-Gallego, F., 2020. Characterization and evaluation of immobilized 662 enzymes for applications in flow reactors. Curr Opin Green Sustain Chem. 663 https://doi.org/10.1016/j.cogsc.2020.04.010 664 Boodhoo, K.V.K., Flickinger, M.C., Woodley, J.M., Emanuelsson, E.A.C., 2022. 665 Bioprocess intensification: A route to efficient and sustainable biocatalytic 666 transformations for the future. Chem Eng Process Process Intensif. 172, 108793. 667 https://doi.org/10.1016/j.cep.2022.108793 668 Diamanti, E., López-Gallego, F., 2024. Single-Particle and Single-Molecule 669 Characterization of Immobilized Enzymes: A Multiscale Path toward Optimizing 670 Heterogeneous Biocatalysts. Angew. Chem. Int. Ed.. 671 https://doi.org/10.1002/anie.202319248 672 Dijkman, W.P., Binda, C., Fraaije, M.W., Mattevi, A., 2015. Structure-based enzyme 673 tailoring of 5-hydroxymethylfurfural oxidase. ACS Catal. 5, 1833–1839. 674 https://doi.org/10.1021/acscatal.5b00031 675 Dijkman, W.P., Fraaije, M.W., 2014. Discovery and characterization of a 5-676 hydroxymethylfurfural oxidase from Methylovorus sp. strain MP688. Appl Environ 677 Microbiol 80, 1082–1090. https://doi.org/10.1128/AEM.03740-13 678 Djapovic, M., Milivojevic, D., Ilic-Tomic, T., Lješević, M., Nikolaivits, E., Topakas, E., 679 Maslak, V., Nikodinovic-Runic, J., 2021. Synthesis and characterization of 680 polyethylene terephthalate (PET) precursors and potential degradation products: 681 Toxicity study and application in discovery of novel PETases. Chemosphere 275. 682 https://doi.org/10.1016/j.chemosphere.2021.130005 683 Domínguez de María, P., 2024. General equations to estimate the CO₂ production of 684 (bio)catalytic reactions in early development stages. RSC Sustain. 2, 3817–3825. 685 https://doi.org/10.1039/D4SU00535J 686 Eerhart, A.J.J.E., Faaij, A.P.C., Patel, M.K., 2012. Replacing fossil based PET with 687 biobased PEF; Process analysis, energy and GHG balance. Energy Environ Sci 5, 688 6407–6422. https://doi.org/10.1039/c2ee02480b 689 Gennari, A., Simon, R., Sperotto, N.D. de M., Bizarro, C.V., Basso, L.A., Machado, P., 690 Benvenutti, E.V., Renard, G., Chies, J.M., Volpato, G., Volken de Souza, C.F., 2022. 691 Application of cellulosic materials as supports for single-step purification and 692 immobilization of a recombinant β-galactosidase via cellulose-binding domain. Int 693 J Biol Macromol 199, 307–317. https://doi.org/10.1016/j.ijbiomac.2022.01.006 694 Herlina, I., Krisnandi, Y.K., Ridwan, M., 2023. MOF-199 and Ni-BTC: Synthesis, 695 Physicochemical Properties, and Catalytic Activity in Oxidation of 5-696 Hydroxymethylfurfural. Bull Chem React Eng Catal 18, 724–735. 697 https://doi.org/10.9767/bcrec.20060 698 Høst, A.V., Meyer, M., Bongratz, P., Fraaije, M.W., Woodley, J.M., 2025. Exposure 699 Experiments to Test the Kinetic Stability of 5-Hydroxymethylfurfural Oxidase 700 (HMFO) in Different Reactor Environments. ChemCatChem. 701 https://doi.org/10.1002/cctc.202500038 702 Hwang, K.R., Jeon, W., Lee, S.Y., Kim, M.S., Park, Y.K., 2020. Sustainable bioplastics: 703 Recent progress in the production of bio-building blocks for the bio-based next-704 generation polymer PEF. Chem Eng J. https://doi.org/10.1016/j.cej.2020.124636 705 Ji, T., Liu, C., Lu, X., Zhu, J., 2018. Coupled Chemical and Thermal Drivers in 706 Microwaves toward Ultrafast HMF Oxidation to FDCA. ACS Sustainable Chem 707 Eng. 6, 11493–11501. https://doi.org/10.1021/acssuschemeng.8b01630 708 Kumar, R., Lalnundiki, V., Shelare, S.D., Abhishek, G.J., Sharma, S., Sharma, D., 709 Kumar, A., Abbas, M., 2024. An investigation of the environmental implications of 710 bioplastics: Recent advancements on the development of environmentally friendly 711 bioplastics solutions. Environ Res. https://doi.org/10.1016/j.envres.2023.117707 712 Li, J., Sun, J., Wang, W., Jiang, C., Hao, J., 2025. Purification, immobilization, evolution, 713 and characterization of D-allulose 3-epimerase from Antarctic Microbacterium. Int 714 J Biol Macromol 310. https://doi.org/10.1016/j.ijbiomac.2025.143294 715 Liu, K., Jiang, L., Wang, L., Zhang, Q., Yang, L., Wu, J., Yu, H., 2025. Sequential single-716 enzyme oxidation of 5-hydroxymethylfurfural to 2,5-furandicarboxylic acid by an 717 engineered lanthanide-dependent alcohol dehydrogenase. Green Chem 27, 5073–718 5090. https://doi.org/10.1039/d5gc00157a 719 Liu, Y., Huang, A., Wen, X., 2024. Enhanced stability and catalytic performance of 720 immobilized phospholipase D on chitosan-encapsulated magnetic nanoparticles 721 using oxidized dextran and glutaraldehyde as cross-linkers. Biochem Eng J 212. 722 https://doi.org/10.1016/j.bej.2024.109499 723 Liu, Y., Wang, P., Tian, J., Seidi, F., Guo, J., Zhu, W., Xiao, H., Song, J., 2022. 724 Carbohydrate-Binding Modules of Potential Resources: Occurrence in Nature, 725 Function, and Application in Fiber Recognition and Treatment. Polymers (Basel). 726 https://doi.org/10.3390/polym14091806 727 Liu, X., Zhang, M., Li, Z., 2020. CoOx–MC (MC = Mesoporous Carbon) for Highly 728 Efficient Oxidation of 5-Hydroxymethylfurfural (5-HMF) to 2,5-Furandicarboxylic 729 Acid (FDCA). ACS Sustainable Chemistry & Engineering 8, 4801–4808. 730 https://doi.org/10.1021/acssuschemeng.9b07443 731 López-Gallego, F., Betancor, L., Mateo, C., Hidalgo, A., Alonso-Morales, N., Dellamora-732 Ortiz, G., Guisán, J.M., Fernández-Lafuente, R., 2005. Enzyme stabilization by 733 glutaraldehyde crosslinking of adsorbed proteins on aminated supports. J Biotechnol 734 119, 70–75. https://doi.org/10.1016/j.jbiotec.2005.05.021 735 Maghraby, Y.R., El-Shabasy, R.M., Ibrahim, A.H., Azzazy, H.M.E.S., 2023. Enzyme 736 Immobilization Technologies and Industrial Applications. ACS Omega. 737 https://doi.org/10.1021/acsomega.2c07560gMartin, C., Ovalle Maqueo, A., Wijma, 738 H.J., Fraaije, M.W., 2018. Creating a more robust 5-hydroxymethylfurfural oxidase 739 by combining computational predictions with a novel effective library design. 740 Biotechnol Biofuels 11. https://doi.org/10.1186/s13068-018-1051-x 741 Mateo, C., Palomo, J.M., Fernandez-Lorente, G., Guisan, J.M., Fernandez-Lafuente, R., 742 2007. Improvement of enzyme activity, stability and selectivity via immobilization 743 techniques. Enzyme Microb Technol. 744 https://doi.org/10.1016/j.enzmictec.2007.01.018 745 Maurya, A., Bhattacharya, A., Khare, S.K., 2020. Enzymatic Remediation of 746 Polyethylene Terephthalate (PET)–Based Polymers for Effective Management of 747 Plastic Wastes: An Overview. Front Bioeng Biotechnol. 748 https://doi.org/10.3389/fbioe.2020.602325 749 Meyer, L.E., Hobisch, M., Kara, S., 2022. Process intensification in continuous flow 750 biocatalysis by up and downstream processing strategies. Curr Opin Biotechnol. 751 https://doi.org/10.1016/j.copbio.2022.102835 752 Milić, M., Domínguez de María, P., Kara, S., 2023. A patent survey on the 753 biotechnological production of 2,5-furandicarboxylic acid (FDCA): Current trends 754 and challenges. EFB Bioeconomy Journal 3, 100050. 755 https://doi.org/10.1016/j.bioeco.2023.100050 756 Modenez, I.A., Sastre, D.E., Moares, F.C., Marques Netto, C.G.C., 2018. Influence of 757 dlutaraldehyde cross-linking modes on the recyclability of immobilized lipase b 758 from candida antarctica for transesterification of soy bean oil. Molecules 23. 759 https://doi.org/10.3390/molecules23092230 760 Ndochinwa, O.G., Wang, Q.Y., Amadi, O.C., Nwagu, T.N., Nnamchi, C.I., Okeke, E.S., 761 Moneke, A.N., 2024. Current status and emerging frontiers in enzyme engineering: 762 An industrial perspective. Heliyon. https://doi.org/10.1016/j.heliyon.2024.e32673 763 Oliveira, C., Carvalho, V., Domingues, L., Gama, F.M., 2015. Recombinant CBM-fusion 764 technology — Applications overview. Biotechnol Adv 33, 358–369. 765 https://doi.org/10.1016/j.biotechadv.2015.02.006 766 Pei, F., Liu, L., Zhu, H., Guo, H., 2023. Recent Advances in Lignocellulose-Based 767 Monomers and Their Polymerization. Polymers (Basel). 768 https://doi.org/10.3390/polym15040829 769 Peng, X., Song, X., Cao, W., Tang, Z., Wang, T., Wen, Y., Wu, Z., Zhou, T., Lin, H., Lu, 770 X., Ji, T., Zhu, J., 2025. Microwave-Activated Peroxyl Radicals Accelerate 771 Hydroxymethyl Oxidation on the AuPd/C Catalyst for Mild Synthesis of FDCA. 772 ACS Sustainable Chem Eng 13, 1233–1244. 773 https://doi.org/10.1021/acssuschemeng.4c07309 774 Perzon, A., Dicko, C., Çobanoğlu, Ö., Yükselen, O., Eryilmaz, J., Dey, E.S., 2017. 775 Cellulase cross-linked enzyme aggregates (CLEA) activities can be modulated and 776 enhanced by precipitant selection. J Chem Technol Biotechnol 92, 1645–1649. 777 https://doi.org/10.1002/jctb.5160 778 Qin, Z., Lin, S., Qiu, Y., Chen, Q., Zhang, Y., Zhou, J., Zhao, L., 2019. One-step 779 immobilization-purification of enzymes by carbohydrate-binding module family 56 780 tag fusion. Food Chem 299. https://doi.org/10.1016/j.foodchem.2019.125037 781 Ramón-Luing, L.A., Cruz-Migoni, A., Ruíz-Medrano, R., Xoconostle-Cázares, B., 782 Ortega-Lopez, J., 2006. One-step purification and immobilization in cellulose of the 783 GroEL apical domain fused to a carbohydrate-binding module and its use in protein 784 refolding. Biotechnol Lett 28, 301–307. https://doi.org/10.1007/s10529-005-5714-x 785 Rathore, A.S., Kapoor, G., 2015. Application of process analytical technology for 786 downstream purification of biotherapeutics. J Chem Technol Biotechnol. 787 https://doi.org/10.1002/jctb.4447 788 Roberts, A.D., Payne, K.A.P., Cosgrove, S., Tilakaratna, V., Penafiel, I., Finnigan, W., 789 Turner, N.J., Scrutton, N.S., 2021. Enzyme immobilisation on wood-derived 790 cellulose scaffoldsviacarbohydrate-binding module fusion constructs. Green Chem 791 23, 4716–4732. https://doi.org/10.1039/d1gc01008e 792 Robescu, M.S., Bavaro, T., 2025. A Comprehensive Guide to Enzyme Immobilization: 793 All You Need to Know. Molecules. https://doi.org/10.3390/molecules30040939 794 Rode, C.V., DHARNE, M.S., PARATE, R.D., 2021. WO2021124354A1. 795 Rodrigues, R.C., Berenguer-Murcia, Á., Carballares, D., Morellon-Sterling, R., 796 Fernandez-Lafuente, R., 2021. Stabilization of enzymes via immobilization: 797 Multipoint covalent attachment and other stabilization strategies. Biotechnol Adv. 798 https://doi.org/10.1016/j.biotechadv.2021.107821 799 Rodriguez, S.R., Álvaro, G., Guillén, M., Romero, O., 2025. Multienzymatic Platform 800 for Coupling a CCU Strategy to Waste Valorization: CO2 from the Iron and Steel 801 Industry and Crude Glycerol from Biodiesel Production. ACS Sustain Chem Eng 802 13, 1440–1449. https://doi.org/10.1021/acssuschemeng.4c04908 803 Sajid, M., Zhao, X., Liu, D., 2018. Production of 2,5-furandicarboxylic acid (FDCA) 804 from 5-hydroxymethylfurfural (HMF): Recent progress focusing on the chemical-805 catalytic routes. Green Chem 20, 5427–5453. https://doi.org/10.1039/c8gc02680g 806 Sánchez-Ruiz, M.I., Martínez, A.T., Serrano, A., 2021. Optimizing operational 807 parameters for the enzymatic production of furandicarboxylic acid building block. 808 Microb Cell Fact 20, 1–13. https://doi.org/10.1186/s12934-021-01669-1 809 Sanders, J.H., Cunniffe, J., Carrejo, E., Burke, C., Reynolds, A.M., Dey, S.C., Islam, 810 M.N., Wagner, O., Argyropoulos, D., 2024. Biobased Polyethylene Furanoate: 811 Production Processes, Sustainability, and Techno-Economics. Adv Sustain Syst. 812 https://doi.org/10.1002/adsu.202400074 813 Santiago-Arcos, J., Velasco-Lozano, S., López-Gallego, F., 2023. Multienzyme 814 Coimmobilization on Triheterofunctional Supports. Biomacromolecules 24, 929–815 942. https://doi.org/10.1021/acs.biomac.2c01364 816 Sheldon, R.A., van Pelt, S., 2013. Enzyme immobilisation in biocatalysis: Why, what and 817 how. Chem Soc Rev 42, 6223–6235. https://doi.org/10.1039/c3cs60075k 818 Shen, X., Zhong, L., Li, L., Zou, B., Suo, H., Yan, L., 2024. Coupling chemical 819 modification and immobilization of cellulase improves thermal stability and low-820 melting mixture solvent tolerance for in situ saccharification of bagasse. J Mol Liq 821 398. https://doi.org/10.1016/j.molliq.2024.124253 822 Singh, A.K., Bedi, R., Kaith, B.S., 2021. Composite materials based on recycled 823 polyethylene terephthalate and their properties – A comprehensive review. Compos 824 B Eng 219, 108928. https://doi.org/10.1016/j.compositesb.2021.108928 825 Totaro, G., Sisti, L., Marchese, P., Colonna, M., Romano, A., Gioia, C., Vannini, M., 826 Celli, A., 2022. Current Advances in the Sustainable Conversion of 5-827 Hydroxymethylfurfural into 2,5-Furandicarboxylic Acid. ChemSusChem. 828 https://doi.org/10.1002/cssc.202200501 829 Troiano, D., Orsat, V., Dumont, M.J., 2020. Status of Biocatalysis in the Production of 830 2,5-Furandicarboxylic Acid. ACS Catal 10, 9145–9169. 831 https://doi.org/10.1021/acscatal.0c02378 832 Ügdüler, S., Van Geem, K.M., Denolf, R., Roosen, M., Mys, N., Ragaert, K., De Meester, 833 S., 2020. Towards closed-loop recycling of multilayer and coloured PET plastic 834 waste by alkaline hydrolysis. Green Chem 22, 5376–5394. 835 https://doi.org/10.1039/d0gc00894j 836 Vidal, L., Pinsach, J., Striedner, G., Caminal, G., Ferrer, P., 2008. Development of an 837 antibiotic-free plasmid selection system based on glycine auxotrophy for 838 recombinant protein overproduction in Escherichia coli. J Biotechnol 134, 127–136. 839 https://doi.org/10.1016/j.jbiotec.2008.01.011 840 Viñambres, M., Espada, M., Martínez, A.T., Serrano, A., 2020. Screening and evaluation 841 of new hydroxymethylfurfural oxidases for furandicarboxylic acid production. Appl 842 Environ Microbiol 86, 1–17. https://doi.org/10.1128/AEM.00842-20 843 Wang, J., Erdem, E., Woodley, J.M., 2023. Effect of Nitrogen, Air, and Oxygen on the 844 Kinetic Stability of NAD(P)H Oxidase Exposed to a Gas-Liquid Interface. Org 845 Process Res Dev 27, 1111–1121. https://doi.org/10.1021/acs.oprd.3c00095 846 Wang, Y., Zhang, Y., Cui, Q., Feng, Y., Xuan, J., 2024. Composition of Lignocellulose 847 Hydrolysate in Different Biorefinery Strategies: Nutrients and Inhibitors. Molecules. 848 https://doi.org/10.3390/molecules29102275 849 Wu, Q., Lu, D., Jin, S., Lu, J., Wang, F., Liu, L., Nie, K., 2021. Engineered stable 5-850 hydroxymethylfurfural oxidase (HMFO) from 8BxHMFO variant of methylovorus 851 sp. MP688 through B-factor analysis. Catalysts 11. 852 https://doi.org/10.3390/catal11121503 853 Wu, S., Liu, Q., Tan, H., Zhang, F., Yin, H., 2020. A Novel 2,5-Furandicarboxylic Acid 854 Biosynthesis Route from Biomass-Derived 5-Hydroxymethylfurfural Based on the 855 Consecutive Enzyme Reactions. Appl Biochem Biotechnol 191, 1470–1482. 856 https://doi.org/10.1007/s12010-020-03290-1 857 Xu, M.Q., Wang, S.S., Li, L.N., Gao, J., Zhang, Y.W., 2018. Combined cross-linked 858 enzyme aggregates as biocatalysts. Catalysts. https://doi.org/10.3390/catal8100460 859 Yang, C.-F., Huang, C.-R., 2016. Biotransformation of 5-hydroxy-methylfurfural into 860 2,5-furan-dicarboxylic acid by bacterial isolate using thermal acid algal hydrolysate. 861 Bioresour. Technol. 214, 311–318. https://doi.org/10.1016/j.biortech.2016.04.122 862 Yang, C.-F., Huang, C.-R., 2018. Isolation of 5-hydroxymethylfurfural biotransforming 863 bacteria to produce 2,5-furan dicarboxylic acid in algal acid hydrolysate. J. Biosci. 864 Bioeng. 125, 407–412. https://doi.org/10.1016/j.jbiosc.2017.11.005 865 Yang, F., Liu, J., Li, B., Li, H., Jiang, Z., 2023. Effective biosynthesis of 2,5-866 furandicarboxylic acid from 5-hydroxymethylfurfural via a bi-enzymatic 867 cascade system using bacterial laccase and fungal alcohol oxidase. Biotechnol 868 Biofuels Bioprod 16. https://doi.org/10.1186/s13068-023-02406-z 869